Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
On 12 Jun 2009, at 06:46, Graham Murray wrote: Norman Rieß writes: What do you want to do with your accesspoint. You will need a bridge to a wired network if you want your ap attached to that wired network. This is quite usual though... Without a bridge to a wired network, only the wlan systems are connected and can not connect to your wired systems. Would it not normally be better to route between the wireless and wired networks, with appropriate firewall rules in place, rather than bridging them? That is the intent of a project I'm working on, and I think it will work well. However most folks don't need the additional complexity of multiple networks. In that case just bridging to the existing subnet is sufficient.
[gentoo-user] OT: Choosing a graphics tablet
Hi list! I'm thinking about buying a graphics tablet to facilitate my work with Dia and Inkscape. The price tag should be below EUR 100. At the moment, I'm wondering if I should go with a Wacom or an Aiptek. Are they similarly good (bad?) supported on Linux? Any gotchas with xorg-server-1.5.3's new "don't configure me, I configure you" mentality? Thanks in advance! Florian Philipp signature.asc Description: OpenPGP digital signature
[gentoo-user] Trouble installing Plone
Hi list! I'm currently trying to install the Plone-CMS on my vServer. However, for some reason, Zope and Plone just don't get together. I've followed this howto: [1] I've also looked at other docs around the net as well as Plone's readme. Zope runs and I can access it. Plone is installed and I've used zprod-manager to add it to the zope-instance. I've also restarted Zope. The problem is: "plone site" just doesn't show up in the list of items I can add! I've checked the permissions, I can see the products which belong to Plone in Zope's product tree. I've tried plone-2.5.5 with zope-2.9.10 and plone-3.1.7 with zope-2.10.7. [1] http://en.gentoo-wiki.com/wiki/Plone I hope someone has an idea. Florian Philipp signature.asc Description: OpenPGP digital signature
Re: [gentoo-user] kernel config for eee w/Atom N270 CPU
Paul Hartman schrieb: > On Thu, Jun 11, 2009 at 3:07 PM, Kelly Hirai wrote: >> the N270 is a single core with hyperthreading, which will apear as 2 >> cpus (with the same core id) in dmesg. > > Ah, I forgot about hyperthreading masquerading as multiple CPUs. In > that case, Maxim can safely disable SMP if he wants to. I don't know > if the theoretical speed gains of disabling SMP outweigh the > theoretical speed gains of enabling hyperthreading. I think it'll > probably be about the same either way. > Well, I don't know about real workloads but once I did a little benchmark: One versus two instances of `dd < /dev/zero | md5sum`. Two instances had a 30% higher throughput than one. I haven't tried it with disabled SMP but I really can't imagine that the extra scheduling would cost nearly enough to compensate for this. From a technical point of view, I think HT makes more sense for an Atom than for a Nehalem: The Atom has only one pipeline, no out-of-order execution and probably a less effective branch prediction. HT might compensate this. However, I'm wondering if it wouldn't have been better to implement out-of-order execution instead of HT (like VIA Nano, for example). Maybe HT doesn't need as many transistors as out-of-order execution? In the end, unless there is some hard evidence against the use of HT, I'd say: They've spend their transistor budget on HT, now we should use what we've got. signature.asc Description: OpenPGP digital signature
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
Grant schrieb: Leave INTERFACES blank. As you keep the networks seperated, hostapd does not depend on any other devices. wlan0 is initialized by hostapd. So you are good to go. The accesspoint itself, so to say the wlan part does not have any IP adress, at it is merely a connectionpoint for normal wlan systems. The IP adress to your device however is defined by the other nics. In your case eth1. I don't have eth1 set up yet. For now I just want eth0 on the WAN and wlan0 on the LAN. eth0 dhcp's from my ISP, but I need to specify a local IP address for my LAN somewhere right? wlan0 in master mode does _not_ have an IP adress. So far eth0 is the only ip adress your device has. If you do not spezify a local ip adress on eth1, you will not have any local ip adress. For the shorewall business, you have to tell, what you want to do with shorewall exactely. I dare say you have a wlan zone as your AP and a loc zone with eth1. As i am using bridging i can not tell you if and how shorewall responds. But if you want to keep eth1 an wlan0 seperate, what so you need shorewall for? Since the AP system is also the router, I use shorewall for NAT, port closing, port forwarding, and packet shaping. shorewall gives an empty loc zone error if I don't have net.wlan0 started because wlan0 is the only loc interface. - Grant You can let shorewall depend on hostapd, so your shorewall starts after hostapd and your wlan0. Check the "depend()" section in shorewalls rc-script.
Re: [gentoo-user] Lost free space on /
Alexander Pilipovsky schrieb: > bn and KH, exuse me if I send not a good question, I have no many > experience yet :) > > That's where all of us started on day ;-) kh
Re: [gentoo-user] Lost free space on /
bn schrieb: > > Isn't "do A if you are in situation X , because of Z" the right pattern? > > m. > That's only if situation x is constant and absolutely known to the one replying. But then it might not be of any use for somebody else. Most likely there will not be a second person with the exact same situation x around. Also I learn a lot by following the way to the solution. Like give me the formula and not the answer and next time I can search for everything myself. kh
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
Matt Causey schrieb: On 12 Jun 2009, at 06:46, Graham Murray wrote: Norman Rieß writes: What do you want to do with your accesspoint. You will need a bridge to a wired network if you want your ap attached to that wired network. This is quite usual though... Without a bridge to a wired network, only the wlan systems are connected and can not connect to your wired systems. Would it not normally be better to route between the wireless and wired networks, with appropriate firewall rules in place, rather than bridging them? That is the intent of a project I'm working on, and I think it will work well. However most folks don't need the additional complexity of multiple networks. In that case just bridging to the existing subnet is sufficient. It really depends on the users needs. I said this was quite usual because with bridging produces the behaviour someone expects from an out of the box accesspoint. If someone wants to control the connections or create a dmz or whatever, routing would be the way, yes. In Grant's situation routing should be the better choice, as he seems to want to have a router with wlan, rather than a simple accesspoint. The wlan becomes the local network and the wired nic, the web. So this would again produce the behavious one expects from a out of the box router. If he later one create a wlan-router setup, i dare say he would bridge wlan and local wired and NAT/route that to the wired web nic. But that are my views... as i said, it depends on the users needs.
Re: [gentoo-user] Lost free space on /
On Friday 12 June 2009 12:48:24 KH wrote: > bn schrieb: > > Isn't "do A if you are in situation X , because of Z" the right pattern? > > > > m. > > That's only if situation x is constant and absolutely known to the one > replying. But then it might not be of any use for somebody else. Most > likely there will not be a second person with the exact same situation x > around. > Also I learn a lot by following the way to the solution. Like give me > the formula and not the answer and next time I can search for everything > myself. The original question was "how big should /var be?" and the correct answer to that question is "mu" (google it) If we had the output of "df -h" and "du -sh /var/*" plus a description of what the machine actually does, some general advice could be given to the OP. As it stands with the information given, the only useful answer is "you'll have to work that out yourself" which is what I said right back at the beginning. -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Lost free space on /
On Friday 12 June 2009, 12:58, Alan McKinnon wrote: > The original question was "how big should /var be?" and the correct > answer to that question is "mu" (google it) > > If we had the output of "df -h" and "du -sh /var/*" plus a description > of what the machine actually does, some general advice could be given > to the OP. As it stands with the information given, the only useful > answer is "you'll have to work that out yourself" which is what I said > right back at the beginning. The OP might be a newbie with little or no idea of how things work in a mailing list like this. Instead of the "you'll have to work that out yourself" answer, which is a correct answer strictly speaking, but not very useful to the OP, one could have asked "can you provide more information on this and that, so people can help you better? And please, in the future, remember that the more information you provide, the better answer you are likely to get." (which, btw, is what you did anyway in later replies). That would be a more useful answer imho, because it will educate (or try to educate) the OP a little more, which hopefully will result in him asking questions in a better way in the future. Just my 2c.
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
>>> Leave INTERFACES blank. As you keep the networks seperated, hostapd does >>> not >>> depend on any other devices. >>> wlan0 is initialized by hostapd. So you are good to go. >>> The accesspoint itself, so to say the wlan part does not have any IP >>> adress, >>> at it is merely a connectionpoint for normal wlan systems. The IP adress >>> to >>> your device however is defined by the other nics. In your case eth1. >>> >> >> I don't have eth1 set up yet. For now I just want eth0 on the WAN and >> wlan0 on the LAN. eth0 dhcp's from my ISP, but I need to specify a >> local IP address for my LAN somewhere right? >> >> > > wlan0 in master mode does _not_ have an IP adress. So far eth0 is the only > ip adress your device has. > If you do not spezify a local ip adress on eth1, you will not have any local > ip adress. I'm very confused. I've been running wlan0 in master mode for about 3 years with IP 192.168.0.1 and no eth1. Here was my entire /etc/conf.d/net: config_eth0=( "dhcp" ) mode_wlan0=( "master" ) essid_wlan0=( "networkname" ) channel_wlan0=( "11" ) config_wlan0=( "192.168.0.1 broadcast 192.168.0.255 netmask 255.255.255.0" ) All I'm trying to do is switch wireless drivers from madwifi-ng to the in-kernel ath5k. With madwifi-ng, I started net.wlan0, started hostapd, and started shorewall and everything worked perfectly. Now with ath5k, net.wlan0 won't start in master mode. This causes 2 problems: 1. I can't specify a local IP for wlan0 in /etc/conf.d/net like I've been doing for years. 2. shorewall checks whether or not net.wlan0 has started because wlan0 is the only device in zone loc, so shorewall won't start. So I'm required to have an eth1 because I'm switching from madwifi-ng to ath5k? That doesn't seem right. >>> For the shorewall business, you have to tell, what you want to do with >>> shorewall exactely. >>> I dare say you have a wlan zone as your AP and a loc zone with eth1. As i >>> am >>> using bridging i can not tell you if and how shorewall responds. >>> But if you want to keep eth1 an wlan0 seperate, what so you need >>> shorewall >>> for? >>> >> >> Since the AP system is also the router, I use shorewall for NAT, port >> closing, port forwarding, and packet shaping. shorewall gives an >> empty loc zone error if I don't have net.wlan0 started because wlan0 >> is the only loc interface. >> >> - Grant >> >> > > You can let shorewall depend on hostapd, so your shorewall starts after > hostapd and your wlan0. > Check the "depend()" section in shorewalls rc-script. I'm confused here too. shorewall seems to be checking whether or not net.wlan0 has started, not whether the wlan0 interface is up. Trying to start shorewall after hostapd has started results in the same error described above because net.wlan0 hasn't been started. - Grant
[gentoo-user] Re: Official Gentoo MythTV Install Guide: never installs mythtv
On 2009-06-12, Mark Knecht wrote: >>> Use the Gentoo specific info on the www.MythTV.org site and >>> you'll do fine. >> >> At the risk of seeming a bit dim... what Gentoo specific info? >> >> The page at http://www.mythtv.org/wiki/Linux_Distros#Gentoo has >> 4 links to Gentoo setup guides. ??Three of them don't exist, and >> the fourth is the previously mentioned "official" install guide >> that hasn't been updated in 4 years and doesn't actually show >> mythtv being installed. > > That's the wiki, not the documentation. OK, thanks. > I think that was all lost when the Gentoo WIki got wiped out. That's what I was afraid of. I thought I'd seen it more recently than that, but I guess not. > I remember what you are talking about but I don't remember > seeing it in the last year. Thanks again. -- Grant Edwards grante Yow! I feel ... JUGULAR ... at visi.com
[gentoo-user] Re: Official Gentoo MythTV Install Guide: never installs mythtv
On 2009-06-12, Mike Kazantsev wrote: > But old gentoo-wiki pages were recovered from google cache and > now reside at gentoo-wiki.info. > > http://www.gentoo-wiki.info/HOWTO_Setup_MythTV > http://www.gentoo-wiki.info/Category:MythTV Bingo! -- Grant Edwards grante Yow! Well, O.K. at I'll compromise with my visi.comprinciples because of EXISTENTIAL DESPAIR!
[gentoo-user] How to veiw absolute latest on partage without syncing
Is there a way to veiw the very latest packages on portage without syncing my OS?
Re: [gentoo-user] How to veiw absolute latest on partage without syncing
Harry Putnam schrieb: > Is there a way to veiw the very latest packages on portage without > syncing my OS? > > http://sources.gentoo.org/viewcvs.py/gentoo-x86/ signature.asc Description: OpenPGP digital signature
Re: [gentoo-user] How to veiw absolute latest on partage without syncing
On Fri, Jun 12, 2009 at 9:29 AM, Justin wrote: > Harry Putnam schrieb: >> Is there a way to veiw the very latest packages on portage without >> syncing my OS? >> >> > > http://sources.gentoo.org/viewcvs.py/gentoo-x86/ also http://packages.gentoo.org/ or http://gentoo-portage.com/Newest there is a similar site for all of the layman overlays, but I seem to have lost the bookmark and my Google searches are turning up nothing. Does anyone know it?
[gentoo-user] About procmail and getline
Trying to install procmail I hit a known bug: http://bugs.gentoo.org/show_bug.cgi?id=270551 And from emerge: [...] In file included from formail.c:25: formisc.h:20: error: conflicting types for 'getline' /usr/include/stdio.h:651: error: previous declaration of 'getline' was here make[1]: *** [formail.o] Error 1 make[1]: Leaving directory `/var/tmp/portage/mail-filter/procmail-3.22-r10/work/procmail-3.22/src' What I'm not understanding is why the current procmail is not flagged in some way... or why its still not fixed and is buggy to install with most recent OS and most recent tools. Its not, apparently a local phenomina, if its a known bug. Or are there people who have been able to install procmail with no problems using current tools? Even downloading the tar ball and building outside of emerge I hit the getline problem, so apparently something incompatible there. There is a patch offered but still one would think using standard emerge on a package that is outside the `~' daredevil stage and is not masked, it should `just work' [tm].
Re: [gentoo-user] Trouble installing Plone
Florian Philipp schrieb: > Hi list! > > I'm currently trying to install the Plone-CMS on my vServer. However, > for some reason, Zope and Plone just don't get together. > > I've followed this howto: [1] I've also looked at other docs around the > net as well as Plone's readme. > > Zope runs and I can access it. Plone is installed and I've used > zprod-manager to add it to the zope-instance. I've also restarted Zope. > The problem is: "plone site" just doesn't show up in the list of items I > can add! > > I've checked the permissions, I can see the products which belong to > Plone in Zope's product tree. I've tried plone-2.5.5 with zope-2.9.10 > and plone-3.1.7 with zope-2.10.7. > > [1] http://en.gentoo-wiki.com/wiki/Plone > > I hope someone has an idea. > I've found this bug report: https://bugs.launchpad.net/ubuntu/+source/plone3/+bug/363065 It states my problem with the solution of installing python-imaging (PIL). But I have dev-python/imaging installed! Is it possible that the problem occurs because Zope depends on dev-lang/python:2.4 but the system default is dev-lang/python:2.5? signature.asc Description: OpenPGP digital signature
Re: [gentoo-user] Trouble installing Plone
Florian Philipp schrieb:
> Florian Philipp schrieb:
>> Hi list!
>>
>> I'm currently trying to install the Plone-CMS on my vServer. However,
>> for some reason, Zope and Plone just don't get together.
>>
>> I've followed this howto: [1] I've also looked at other docs around the
>> net as well as Plone's readme.
>>
>> Zope runs and I can access it. Plone is installed and I've used
>> zprod-manager to add it to the zope-instance. I've also restarted Zope.
>> The problem is: "plone site" just doesn't show up in the list of items I
>> can add!
>>
>> I've checked the permissions, I can see the products which belong to
>> Plone in Zope's product tree. I've tried plone-2.5.5 with zope-2.9.10
>> and plone-3.1.7 with zope-2.10.7.
>>
>> [1] http://en.gentoo-wiki.com/wiki/Plone
>>
>> I hope someone has an idea.
>>
>
> I've found this bug report:
> https://bugs.launchpad.net/ubuntu/+source/plone3/+bug/363065
>
> It states my problem with the solution of installing python-imaging
> (PIL). But I have dev-python/imaging installed! Is it possible that the
> problem occurs because Zope depends on dev-lang/python:2.4 but the
> system default is dev-lang/python:2.5?
>
Hmm, this really seems to be the problem. Running zope like
/var/lib/zope/zope-${instance}/bin/zopectl fg gives an error about a
missing PIL (full error report attached).
If the dependency on python-2.4 causes the problem, how is this supposed
to work then? After all, portage depends on python-2.5!
2009-06-12 17:03:39 INFO ZServer HTTP server started at Fri Jun 12 17:03:39 2009
Hostname: 0.0.0.0
Port: 8080
2009-06-12 17:03:39 INFO Zope Set effective user to "zope-binarywings"
2009-06-12 17:03:41 ERROR PortalTransforms Problem importing module
image_to_png : No module named PIL.Image
2009-06-12 17:03:41 ERROR PortalTransforms Problem importing module
image_to_gif : No module named PIL.Image
2009-06-12 17:03:41 ERROR PortalTransforms Problem importing module
image_to_jpeg : No module named PIL.Image
2009-06-12 17:03:41 ERROR PortalTransforms Problem importing module
image_to_pcx : No module named PIL.Image
2009-06-12 17:03:41 ERROR PortalTransforms Problem importing module
image_to_ppm : No module named PIL.Image
2009-06-12 17:03:41 ERROR PortalTransforms Problem importing module
image_to_tiff : No module named PIL.Image
2009-06-12 17:03:41 ERROR PortalTransforms Problem importing module
image_to_bmp : No module named PIL.Image
2009-06-12 17:03:43 INFO Archetypes Products/Archetypes/Field.py[102]:?
Warning: no Python Imaging Libraries (PIL) found.Archetypes based ImageField's
don't scale if neccessary.
2009-06-12 17:03:43 ERROR Application Could not import Products.ATContentTypes
Traceback (most recent call last):
File
"/var/tmp/portage/net-zope/zope-2.10.7/image/usr/lib/zope-2.10.7/lib/python/OFS/Application.py",
line 709, in import_product
File "/var/lib/zope/zope-binarywings/Products/ATContentTypes/__init__.py",
line 64, in ?
import Products.ATContentTypes.content
File
"/var/lib/zope/zope-binarywings/Products/ATContentTypes/content/__init__.py",
line 26, in ?
import Products.ATContentTypes.content.link
File
"/var/lib/zope/zope-binarywings/Products/ATContentTypes/content/link.py", line
39, in ?
from Products.ATContentTypes.content.base import registerATCT
File
"/var/lib/zope/zope-binarywings/Products/ATContentTypes/content/base.py", line
63, in ?
from Products.CMFPlone.PloneFolder import ReplaceableWrapper
File "/var/lib/zope/zope-binarywings/Products/CMFPlone/__init__.py", line
215, in ?
from browser import ploneview
File "/var/lib/zope/zope-binarywings/Products/CMFPlone/browser/ploneview.py",
line 12, in ?
from Products.CMFPlone import utils
File "/var/lib/zope/zope-binarywings/Products/CMFPlone/utils.py", line 6, in ?
from PIL import Image
ImportError: No module named PIL
Traceback (most recent call last):
File "/usr/lib/zope-2.10.7/lib/python/Zope2/Startup/run.py", line 56, in ?
run()
File "/usr/lib/zope-2.10.7/lib/python/Zope2/Startup/run.py", line 21, in run
starter.prepare()
File "/usr/lib/zope-2.10.7/lib/python/Zope2/Startup/__init__.py", line 102,
in prepare
self.startZope()
File "/usr/lib/zope-2.10.7/lib/python/Zope2/Startup/__init__.py", line 278,
in startZope
Zope2.startup()
File "/usr/lib/zope-2.10.7/lib/python/Zope2/Startup/__init__.py", line 47, in
startup
# Don't allow any code to call start_zope() twice.
File
"/var/tmp/portage/net-zope/zope-2.10.7/image/usr/lib/zope-2.10.7/lib/python/Zope2/App/startup.py",
line 45, in startup
File
"/var/tmp/portage/net-zope/zope-2.10.7/image/usr/lib/zope-2.10.7/lib/python/OFS/Application.py",
line 686, in import_products
File
"/var/tmp/portage/net-zope/zope-2.10.7/image/usr/lib/zope-2.10.7/lib/python/OFS/Application.py",
line 709, in import_product
File "/var/lib/zope/zope-binarywings/Products/ATContentTypes/__init__.py",
line 64, in ?
import Products.ATContentTypes.content
F
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
Grant schrieb: Leave INTERFACES blank. As you keep the networks seperated, hostapd does not depend on any other devices. wlan0 is initialized by hostapd. So you are good to go. The accesspoint itself, so to say the wlan part does not have any IP adress, at it is merely a connectionpoint for normal wlan systems. The IP adress to your device however is defined by the other nics. In your case eth1. I don't have eth1 set up yet. For now I just want eth0 on the WAN and wlan0 on the LAN. eth0 dhcp's from my ISP, but I need to specify a local IP address for my LAN somewhere right? wlan0 in master mode does _not_ have an IP adress. So far eth0 is the only ip adress your device has. If you do not spezify a local ip adress on eth1, you will not have any local ip adress. I'm very confused. I've been running wlan0 in master mode for about 3 years with IP 192.168.0.1 and no eth1. Here was my entire /etc/conf.d/net: config_eth0=( "dhcp" ) mode_wlan0=( "master" ) essid_wlan0=( "networkname" ) channel_wlan0=( "11" ) config_wlan0=( "192.168.0.1 broadcast 192.168.0.255 netmask 255.255.255.0" ) All I'm trying to do is switch wireless drivers from madwifi-ng to the in-kernel ath5k. With madwifi-ng, I started net.wlan0, started hostapd, and started shorewall and everything worked perfectly. Now with ath5k, net.wlan0 won't start in master mode. This causes 2 problems: 1. I can't specify a local IP for wlan0 in /etc/conf.d/net like I've been doing for years. 2. shorewall checks whether or not net.wlan0 has started because wlan0 is the only device in zone loc, so shorewall won't start. So I'm required to have an eth1 because I'm switching from madwifi-ng to ath5k? That doesn't seem right. For the shorewall business, you have to tell, what you want to do with shorewall exactely. I dare say you have a wlan zone as your AP and a loc zone with eth1. As i am using bridging i can not tell you if and how shorewall responds. But if you want to keep eth1 an wlan0 seperate, what so you need shorewall for? Since the AP system is also the router, I use shorewall for NAT, port closing, port forwarding, and packet shaping. shorewall gives an empty loc zone error if I don't have net.wlan0 started because wlan0 is the only loc interface. - Grant You can let shorewall depend on hostapd, so your shorewall starts after hostapd and your wlan0. Check the "depend()" section in shorewalls rc-script. I'm confused here too. shorewall seems to be checking whether or not net.wlan0 has started, not whether the wlan0 interface is up. Trying to start shorewall after hostapd has started results in the same error described above because net.wlan0 hasn't been started. - Grant Well, madwifi-ng is a matured project with an insanely great featureset. ath5k ap mode till this day is not activated in the kernel. You have to activate it with a code patch, the gentoo rc-script can not cope with it yet. hostapd needs to be a new version and has to initialize the device itself. Of course you can not expect the same features and easy to use behaviour from such an experimental software. You seem to have a working setup, which suits your needs. Unless you have a serious reason i would not change a running and supported system.
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
> Leave INTERFACES blank. As you keep the networks seperated, hostapd > does > not > depend on any other devices. > wlan0 is initialized by hostapd. So you are good to go. > The accesspoint itself, so to say the wlan part does not have any IP > adress, > at it is merely a connectionpoint for normal wlan systems. The IP > adress > to > your device however is defined by the other nics. In your case eth1. > > I don't have eth1 set up yet. For now I just want eth0 on the WAN and wlan0 on the LAN. eth0 dhcp's from my ISP, but I need to specify a local IP address for my LAN somewhere right? >>> >>> wlan0 in master mode does _not_ have an IP adress. So far eth0 is the >>> only >>> ip adress your device has. >>> If you do not spezify a local ip adress on eth1, you will not have any >>> local >>> ip adress. >>> >> >> I'm very confused. I've been running wlan0 in master mode for about 3 >> years with IP 192.168.0.1 and no eth1. Here was my entire >> /etc/conf.d/net: >> >> config_eth0=( "dhcp" ) >> mode_wlan0=( "master" ) >> essid_wlan0=( "networkname" ) >> channel_wlan0=( "11" ) >> config_wlan0=( "192.168.0.1 broadcast 192.168.0.255 netmask 255.255.255.0" >> ) >> >> All I'm trying to do is switch wireless drivers from madwifi-ng to the >> in-kernel ath5k. With madwifi-ng, I started net.wlan0, started >> hostapd, and started shorewall and everything worked perfectly. Now >> with ath5k, net.wlan0 won't start in master mode. This causes 2 >> problems: >> >> 1. I can't specify a local IP for wlan0 in /etc/conf.d/net like I've >> been doing for years. >> 2. shorewall checks whether or not net.wlan0 has started because wlan0 >> is the only device in zone loc, so shorewall won't start. >> >> So I'm required to have an eth1 because I'm switching from madwifi-ng >> to ath5k? That doesn't seem right. >> >> > > For the shorewall business, you have to tell, what you want to do with > shorewall exactely. > I dare say you have a wlan zone as your AP and a loc zone with eth1. As > i > am > using bridging i can not tell you if and how shorewall responds. > But if you want to keep eth1 an wlan0 seperate, what so you need > shorewall > for? > > Since the AP system is also the router, I use shorewall for NAT, port closing, port forwarding, and packet shaping. shorewall gives an empty loc zone error if I don't have net.wlan0 started because wlan0 is the only loc interface. - Grant >>> >>> You can let shorewall depend on hostapd, so your shorewall starts after >>> hostapd and your wlan0. >>> Check the "depend()" section in shorewalls rc-script. >>> >> >> I'm confused here too. shorewall seems to be checking whether or not >> net.wlan0 has started, not whether the wlan0 interface is up. Trying >> to start shorewall after hostapd has started results in the same error >> described above because net.wlan0 hasn't been started. >> >> - Grant >> >> > > Well, madwifi-ng is a matured project with an insanely great featureset. > ath5k ap mode till this day is not activated in the kernel. You have to > activate it with a code patch, the gentoo rc-script can not cope with it > yet. hostapd needs to be a new version and has to initialize the device > itself. > Of course you can not expect the same features and easy to use behaviour > from such an experimental software. > > You seem to have a working setup, which suits your needs. Unless you have a > serious reason i would not change a running and supported system. OK, thank you Norman. The reason I'm trying to switch (this is my third serious attempt) is some kind of a bug that crashes the system when SMP is enabled and the madwifi driver is in master mode. I've been running without SMP, but I could really use the extra power. Do you know if there is better Gentoo support for this on the horizon? - Grant
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
On 12 Jun 2009, at 14:17, Grant wrote: ... wlan0 in master mode does _not_ have an IP adress. So far eth0 is the only ip adress your device has. If you do not spezify a local ip adress on eth1, you will not have any local ip adress. I'm very confused. I've been running wlan0 in master mode for about 3 years with IP 192.168.0.1 and no eth1. Here was my entire /etc/conf.d/net: config_eth0=( "dhcp" ) mode_wlan0=( "master" ) essid_wlan0=( "networkname" ) channel_wlan0=( "11" ) config_wlan0=( "192.168.0.1 broadcast 192.168.0.255 netmask 255.255.255.0" ) All I'm trying to do is switch wireless drivers from madwifi-ng to the in-kernel ath5k. With madwifi-ng, I started net.wlan0, started hostapd, and started shorewall and everything worked perfectly. Now with ath5k, net.wlan0 won't start in master mode. This causes 2 problems: 1. I can't specify a local IP for wlan0 in /etc/conf.d/net like I've been doing for years. 2. shorewall checks whether or not net.wlan0 has started because wlan0 is the only device in zone loc, so shorewall won't start. So I'm required to have an eth1 because I'm switching from madwifi-ng to ath5k? That doesn't seem right. For master mode AP to work, you should indeed be allocating it an IP address, just as you did before. My experience was with madwifi some years ago, when it was the only driver for Atheros chips (am I remembering correctly?) and this combination was absolutely the best for an access-point setup. At that time the only other 802.11g driver that did master mode was, I think, Prism54 and it was a little difficult to get hold of cards featuring that chipset (consequently I got into the side-business of selling them, and probably have 20 left here). madwifi was better because it featured "virtual APs" (VAPs) and allowed you to run separate WEP & unencrypted wireless networks on the same card (and run iptables rules on the interface allocated to each VAP). So I'm not sure why you're changing from madwifi to ath5k. But it _should_ be possible to assign an address to the wireless interface in master mode. And in the situation you describe - a router with only 2 interfaces, WAN & wLAN, then that's exactly what you want to do. The client machines on the wLAN will have IP addresses and they must be told the IP address of the gateway. The WAN IP address will be issued by your ISP, and of course the wLAN IP addresses must be in a private range. The gateway's LAN IP address must be on the same subnet as all the client PCs on the wLAN. Stroller.
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
On 12 Jun 2009, at 16:38, Grant wrote: ... OK, thank you Norman. The reason I'm trying to switch (this is my third serious attempt) is some kind of a bug that crashes the system when SMP is enabled and the madwifi driver is in master mode. I've been running without SMP, but I could really use the extra power. That's interesting. I had an old 4 x processor machine running as an access-point (madwifi or madwifi-ng) running in master mode for at least a year or two. It was unstable as heck, and I never attributed it to this. It would, however, stay up for days or weeks at a time. Maybe this bug has crept in more recently? I'm not sure that it will apply to my new system (on which I'd like to run an AP, as soon as I get round to it) as that is a single processor P4. Do you know if there is better Gentoo support for this on the horizon? I did find the dev uberlord immensely helpful when I was first doing this. He was the baselayout guy at the time, although I don't know if he still is or if you might be able to get hold of him. IMO the first thing to do is get the AP up & running without resort to the Gentoo init.d scripts. Try allocating an IP address to wlan0 just using `ifconfig` as root. If that works then you know the hardware & principles of operation are all ok. Stroller.
[gentoo-user] {OT} "make uninstall" with no rule
I'm just now learning how to compile and install manually. I installed makemkv-1.4.1 manually and now I found an ebuild so I'd like to install the latest version via the ebuild, but I get: # make uninstall make: *** No rule to make target `uninstall'. Stop. The makefile doesn't mention uninstall. Should I just install over the current installation via the ebuild? Is there any way to do this cleanly? - Grant
Re: [gentoo-user] Lost free space on /
Etaoin Shrdlu wrote: > On Friday 12 June 2009, 12:58, Alan McKinnon wrote: > > >> The original question was "how big should /var be?" and the correct >> answer to that question is "mu" (google it) >> >> If we had the output of "df -h" and "du -sh /var/*" plus a description >> of what the machine actually does, some general advice could be given >> to the OP. As it stands with the information given, the only useful >> answer is "you'll have to work that out yourself" which is what I said >> right back at the beginning. >> > > The OP might be a newbie with little or no idea of how things work in a > mailing list like this. Instead of the "you'll have to work that out > yourself" answer, which is a correct answer strictly speaking, but not > very useful to the OP, one could have asked "can you provide more > information on this and that, so people can help you better? And please, > in the future, remember that the more information you provide, the > better answer you are likely to get." (which, btw, is what you did > anyway in later replies). That would be a more useful answer imho, > because it will educate (or try to educate) the OP a little more, which > hopefully will result in him asking questions in a better way in the > future. > > Just my 2c. > > > Alan's point is, there is no way for us to know that. Example, I sometimes use http-replicator on my machine which is placed in /var. Therefore, that alone could need 2 to 3GBs. If you use ccache, then add some more. Also, doesn't portage use /vat to compile? If so, then that is some more space that would be needed. Does the person use OOo from source or binary? Is this a web server of some sort? Is it going to be used for a DVR type system? A even better question would be this, how much space does the OP have to begin with? I have two 80GB drives. The OP may have some huge 300GB drive. I dont' have any idea what the answer to most of those are so no one really can answer the question yet. Dale :-) :-)
Re: [gentoo-user] How to veiw absolute latest on partage without syncing
On 12 Jun 2009, at 15:40, Paul Hartman wrote: On Fri, Jun 12, 2009 at 9:29 AM, Justin wrote: Harry Putnam schrieb: Is there a way to veiw the very latest packages on portage without syncing my OS? http://sources.gentoo.org/viewcvs.py/gentoo-x86/ also http://packages.gentoo.org/ or http://gentoo-portage.com/Newest The problem with this is that it's difficult to determine which packages on one's own system have updated. One must check individually for each atom in world. Harry: I'm not sure if it's possible _without_ syncing, but you can `cp -a / usr/portage /usr/portage.orig`, sync, `emerge -pv world` and then move the original tree back if you want to. It's not really clear why you're asking, or why you're unable to sync. If the PC has no internet connection, for instance, security updates are unimportant. Stroller.
Re: [gentoo-user] {OT} "make uninstall" with no rule
On Fri, Jun 12, 2009 at 11:02 AM, Grant wrote: > I'm just now learning how to compile and install manually. I > installed makemkv-1.4.1 manually and now I found an ebuild so I'd like > to install the latest version via the ebuild, but I get: > > # make uninstall > make: *** No rule to make target `uninstall'. Stop. > > The makefile doesn't mention uninstall. Should I just install over > the current installation via the ebuild? Is there any way to do this > cleanly? Hopefully your manually-installed version was installed to /usr/local and the ebuild version will be installed to /usr ... once the ebuild is installed, check out the files list and then manually delete those same files from the /usr/local hierarchy. That's what I would do anyway :) If you did not install the original to /usr/local then yeah I think just having it overwrite the files should be okay in this case.
[gentoo-user] Re: {OT} "make uninstall" with no rule
Grant writes: > I'm just now learning how to compile and install manually. I > installed makemkv-1.4.1 manually and now I found an ebuild so I'd like > to install the latest version via the ebuild, but I get: > > # make uninstall > make: *** No rule to make target `uninstall'. Stop. > > The makefile doesn't mention uninstall. Should I just install over > the current installation via the ebuild? Is there any way to do this > cleanly? When I've wanted to remove manually built packages... I've rerun make install like this: make install >../_install.log 2>&1 Then from the install log you can see what has been installed and remove it by hand.
[gentoo-user] Re: How to veiw absolute latest on partage without syncing
Stroller writes: > It's not really clear why you're asking, or why you're unable to sync. > If the PC has no internet connection, for instance, security updates > are unimportant. Thanks for the tips... no it was something totally mundane. I wanted to see if anything had been done to mail-filter/procmail about the install bug involving getline. I can sync fine... just didn't want to do an update world just now as my sources are quite new, but still didn't want to get sources too far ahead of installed packages. The handy urls given in this thread allowed me to see all I wanted. Thanks for the URLS posters...
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
Stroller schrieb: On 12 Jun 2009, at 16:38, Grant wrote: ... OK, thank you Norman. The reason I'm trying to switch (this is my third serious attempt) is some kind of a bug that crashes the system when SMP is enabled and the madwifi driver is in master mode. I've been running without SMP, but I could really use the extra power. That's interesting. I had an old 4 x processor machine running as an access-point (madwifi or madwifi-ng) running in master mode for at least a year or two. It was unstable as heck, and I never attributed it to this. It would, however, stay up for days or weeks at a time. Maybe this bug has crept in more recently? I'm not sure that it will apply to my new system (on which I'd like to run an AP, as soon as I get round to it) as that is a single processor P4. Do you know if there is better Gentoo support for this on the horizon? I did find the dev uberlord immensely helpful when I was first doing this. He was the baselayout guy at the time, although I don't know if he still is or if you might be able to get hold of him. IMO the first thing to do is get the AP up & running without resort to the Gentoo init.d scripts. Try allocating an IP address to wlan0 just using `ifconfig` as root. If that works then you know the hardware & principles of operation are all ok. Stroller. I would recomment the same thing. Play around manualy. Find out what works and what does not. And if you found a manual way, you can start scriptworkarounds and automating things.
Re: [gentoo-user] Re: How to veiw absolute latest on partage without syncing
Harry Putnam wrote: > Stroller writes: > > >> It's not really clear why you're asking, or why you're unable to sync. >> If the PC has no internet connection, for instance, security updates >> are unimportant. >> > > Thanks for the tips... no it was something totally mundane. > > I wanted to see if anything had been done to mail-filter/procmail > about the install bug involving getline. > > I can sync fine... just didn't want to do an update world just now as > my sources are quite new, but still didn't want to get sources too far > ahead of installed packages. > > The handy urls given in this thread allowed me to see all I wanted. > Thanks for the URLS posters... > > > > You could always sync then do a emerge -uv procmail. then it would only upgrade procmail and any friends that need to be updated. That would mostly likely miss most of the other updates that you are wanting to skip for the moment. Dale :-) :-)
[gentoo-user] weirdness with the mounting of /
I'm having some weirdness with my '/' mount. Here's the line from /etc/fstab: douglas ~ # grep 'md3' /etc/fstab /dev/md3 / ext4 noatime,journal_checksum,defaults 0 1 Yet, the output of mount shows: douglas ~ # mount rootfs on / type rootfs (rw) /dev/root on / type ext4 (rw,noatime,barrier=1,data=ordered) As you can see, it doesn't show journal_checksum as being one of the mount options nor does it show noatime. Why? For that matter, why is it listed as 'rootfs' *and* '/dev/root' and not '/dev/md3' ? My grub config: kernel /boot/vmlinuz-2.6.30-gentoo-r1 root=/dev/md3 rootfstype=ext4 Thanks in advance for cluebatting me. -- Douglas J Hunley - d...@hunley.homeip.net http://douglasjhunley.com Twitter: @hunleyd
[gentoo-user] Re: How to veiw absolute latest on partage without syncing
Dale writes: > You could always sync then do a emerge -uv procmail. then it would only > upgrade procmail and any friends that need to be updated. That would > mostly likely miss most of the other updates that you are wanting to > skip for the moment. yeah... its a thought... but why emerge world then procmail again when I know it won't emerge unless something is done about the getline bug? I wanted to check in most recent portage to see if any changes were made to procmail.
Re: [gentoo-user] weirdness with the mounting of /
On Fri, Jun 12, 2009 at 11:28 AM, Doug Hunley wrote: > I'm having some weirdness with my '/' mount. Here's the line from /etc/fstab: > douglas ~ # grep 'md3' /etc/fstab > /dev/md3 / ext4 noatime,journal_checksum,defaults 0 1 > > Yet, the output of mount shows: > douglas ~ # mount > rootfs on / type rootfs (rw) > /dev/root on / type ext4 (rw,noatime,barrier=1,data=ordered) > > As you can see, it doesn't show journal_checksum as being one of the > mount options nor does it show noatime. Why? For that matter, why is > it listed as 'rootfs' *and* '/dev/root' and not '/dev/md3' ? > > My grub config: > kernel /boot/vmlinuz-2.6.30-gentoo-r1 root=/dev/md3 rootfstype=ext4 > > Thanks in advance for cluebatting me. I think your root is mounted before fstab comes into play... You may want to look into the "rootflags" option in your grub kernel commandline for passing the mount options for your root partition.
[gentoo-user] Re: weirdness with the mounting of /
On Fri, Jun 12, 2009 at 12:41, Paul Hartman wrote: > > I think your root is mounted before fstab comes into play... You may > want to look into the "rootflags" option in your grub kernel > commandline for passing the mount options for your root partition. > That jives with what I'm (slowly) finding on Google as well. It seems the kernel uses /dev/root when the rootfs is RO. Apparently an initrd would solve this. Grr... -- Douglas J Hunley - d...@hunley.homeip.net http://douglasjhunley.com Twitter: @hunleyd
Re: [gentoo-user] How to veiw absolute latest on partage without syncing
On Fri, Jun 12, 2009 at 9:07 AM, Stroller wrote: > > On 12 Jun 2009, at 15:40, Paul Hartman wrote: >> >> On Fri, Jun 12, 2009 at 9:29 AM, Justin wrote: >>> >>> Harry Putnam schrieb: Is there a way to veiw the very latest packages on portage without syncing my OS? >>> >>> http://sources.gentoo.org/viewcvs.py/gentoo-x86/ >> >> also http://packages.gentoo.org/ >> >> or http://gentoo-portage.com/Newest > > The problem with this is that it's difficult to determine which packages on > one's own system have updated. One must check individually for each atom in > world. > > > Harry: > > I'm not sure if it's possible _without_ syncing, but you can `cp -a > /usr/portage /usr/portage.orig`, sync, `emerge -pv world` and then move the > original tree back if you want to. > > It's not really clear why you're asking, or why you're unable to sync. If > the PC has no internet connection, for instance, security updates are > unimportant. > > Stroller. I've wanted a way to do something like this for a long time. One problem with the way portage works with ( I guess) rsync or whatever it uses is that when someone decides to remove a package from portage that I'm currently using syncing removes it from my system also. Unfortunately before I do the sync I have no idea it has been removed so I don't know that it's going to get taken off my system. Once it does I can go find a copy and put it in a personal overlay but that requires I do the work after the damage is done. It would be nice if there was a message ahead of time that told me certain packages were going to be removed, etc., before it was actually done, but I understand from previous conversations that syncing doesn't work that way. This has come up numerous times for me on older hardware where, for instance, maybe some on-board graphics chip only works with older ATI drivers, and that ATI driver only works with older kernels. By the time sync is done I've lost the code for what my system is running, and unfortunately there's no messages that this is happening when I'm doing the sync so maybe I only figure it out a few weeks later and then have to mess around building an overlay using the attic. - Mark
[gentoo-user] Re: weirdness with the mounting of /
On 06/12/2009 07:44 PM, Doug Hunley wrote: On Fri, Jun 12, 2009 at 12:41, Paul Hartman wrote: I think your root is mounted before fstab comes into play... You may want to look into the "rootflags" option in your grub kernel commandline for passing the mount options for your root partition. That jives with what I'm (slowly) finding on Google as well. It seems the kernel uses /dev/root when the rootfs is RO. Apparently an initrd would solve this. Grr... It is remounted rw after the fschk using fstab options. If you change the options, you'll see they will apply (why would noatime apply otherwise.) You don't need an initrd.
[gentoo-user] building packages remotely
Hi group, I've read references here and in other forums to building packages on a desktop PC and installing them on a note/netbook remotely as a way of relieving stress on the smaller machine. Can someone point me to the documentation or howto? I can't seem to come up with the proper google input that doesn't lead to garbage. Maxim
Re: [gentoo-user] Re: {OT} "make uninstall" with no rule
>> I'm just now learning how to compile and install manually. I >> installed makemkv-1.4.1 manually and now I found an ebuild so I'd like >> to install the latest version via the ebuild, but I get: >> >> # make uninstall >> make: *** No rule to make target `uninstall'. Stop. >> >> The makefile doesn't mention uninstall. Should I just install over >> the current installation via the ebuild? Is there any way to do this >> cleanly? > > When I've wanted to remove manually built packages... I've rerun > make install like this: > > make install >../_install.log 2>&1 > > Then from the install log you can see what has been installed and > remove it by hand. That worked brilliantly, thanks a lot. - Grant
Re: [gentoo-user] How to veiw absolute latest on partage without syncing
On Fri, 12 Jun 2009 09:45:27 -0700
Mark Knecht wrote:
> I've wanted a way to do something like this for a long time. One
> problem with the way portage works with ( I guess) rsync or whatever
> it uses is that when someone decides to remove a package from portage
> that I'm currently using syncing removes it from my system also.
> Unfortunately before I do the sync I have no idea it has been removed
> so I don't know that it's going to get taken off my system. Once it
> does I can go find a copy and put it in a personal overlay but that
> requires I do the work after the damage is done. It would be nice if
> there was a message ahead of time that told me certain packages were
> going to be removed, etc., before it was actually done, but I
> understand from previous conversations that syncing doesn't work that
> way.
But why not?
alias emerge-sync='rm -Rf /usr/portage.bak && mv /usr/portage{,.bak} \
&& emerge --sync"
and to make.conf goes:
PORTAGE_RSYNC_OPTS="--link-dest=/usr/portage.bak $PORTAGE_RSYNC_OPTS"
And there you go: hardlinked new tree w/ old one easily accessible.
Note that it won't take much more time or bandwith or space than syncing
on top of the older tree, since same check will see that the files
in .bak dir and remote tree are identical and will just create another
hardlink to the same file.
And you're free to dispose of any dir with "rm -Rf" when you see fit.
Simple three-line no-brainer script will help you keep 10, 100, 1000 or
however many trees you like, occupying just a few MB more than a single
tree.
Furthermore, if you want to keep a hundred-year history of this tree,
just say something like this:
cd /usr/portage
echo -e "local\ndistfiles\npackages" > .gitignore
git init
git add .
git commit -a -m "portage bump"
alias emerge-sync='cd /usr/portage && git add . \
&& git commit -a -m "portage bump" && emerge --sync'
and there you go, you'll never loose even a single bit of ebuild, no
matter how many times a day you keep syncing.
And .git storage will keep storage requiments of the whole thing to
minimum, keeping each change in the single place, compressing them,
etc...
Git-foo is too cryptic? There are few dozens of other VCS, of, for that
matter, ways to keep track of changes: snapshots, fs like venti/fossil,
rdup, even cp/tar.
Guess funtoo project is also worth mention in such context since it
uses git instead of rsync out-of-the-box.
> This has come up numerous times for me on older hardware where, for
> instance, maybe some on-board graphics chip only works with older ATI
> drivers, and that ATI driver only works with older kernels. By the
> time sync is done I've lost the code for what my system is running,
> and unfortunately there's no messages that this is happening when I'm
> doing the sync so maybe I only figure it out a few weeks later and
> then have to mess around building an overlay using the attic.
No dev can ever satisfy every requiment of everyone if they are too
lazy to lift a finger to type a line or two themselves.
--
Mike Kazantsev // fraggod.net
signature.asc
Description: PGP signature
Re: [gentoo-user] Can I exclude a package from --depclean's consideration?
Paul Hartman wrote: On Thu, Jun 11, 2009 at 11:45 AM, Mike Kazantsev wrote: So, my question: Is there a way to tell depclean to never remove *any* version of gentoo-sources? That's where portage-2.2 sets find another use. Just add following set to /usr/share/portage/config/sets.conf: [kernels] class = portage.sets.dbapi.OwnerSet world-candidate = False files = /usr/src And append "@kernels" line to /var/lib/portage/world_sets Now any installed (even with -1) kernel should be safe from ravenous depclean. Perfect! It does exactly what I wanted. I created sets.conf in /etc/portage/ as Boris pointed out. Thank you very much for your help! What would I add to /etc/portage/sets.conf to exclude gcc from depclean? thanks -david -- Powered by Gentoo GNU/Linux http://linuxcrazy.com
[gentoo-user] conflict in fstab w/ lvm?
Hi group, Following the LVM2 gentoo doc I have in fstab: ... /dev/vg/tmp /tmp ext2 noatime 0 2 ... But also(suggested by the eee forum): ... #shm/dev/shmtmpfs nodev,nosuid,noexec 0 0 tmpfs /tmptmpfs defaults,noatime,mode=1777 0 0 Is this legal? Mounting two things at the same place? Maxm
Re: [gentoo-user] conflict in fstab w/ lvm?
On Fri, Jun 12, 2009 at 16:45, Maxim Wexler wrote: > Hi group, > > Following the LVM2 gentoo doc I have in fstab: > > ... > /dev/vg/tmp /tmp ext2 noatime 0 2 > ... > > But also(suggested by the eee forum): > > ... > #shm /dev/shm tmpfs nodev,nosuid,noexec 0 0 > tmpfs /tmp tmpfs defaults,noatime,mode=1777 0 0 > > Is this legal? Mounting two things at the same place? > AFAIK, no. First off, what do you want to do? The EEE forum suggested mounting /tmp using tmpfs cause that keep temporary stuff on your RAM, not disk, this way you reduce disk access. Its your decision to use RAM for /tmp or disk (LVM logical volume). Obviously you can't have both (it doesn't even make sense). -- Daniel da Veiga
Re: [gentoo-user] Lost free space on /
On Friday 12 June 2009 18:05:29 Dale wrote: > Alan's point is, there is no way for us to know that. Example, I > sometimes use http-replicator on my machine which is placed in /var. > Therefore, that alone could need 2 to 3GBs. If you use ccache, then add > some more. Also, doesn't portage use /vat to compile? If so, then that > is some more space that would be needed. Does the person use OOo from > source or binary? Is this a web server of some sort? Is it going to > be used for a DVR type system? There's a few guidlelines one can give (but only a few). The variables tend to be large than the amounts with guidelines though. /var/tmp/portage should be at least 1G on a modern system, 6G+ if building mozilla stuff and OOo is something you intend to do. portage cache is about 200M ccache needs as much space as it was given in make.conf /var/mail or /var/spool/mail is completely dependant on number of user and how much mail they get. Often none for a desktop and huge amounts for a mail store. mysql and postgres need as much as the amount of data intended to be stored. log space is very big or not too much depending on what you do. So. The partition holding /var needs at least 1.5G on a modern gentoo system, probably more, sometimes LOTS more. Only the admin knows how much more and there is no silver bullet answer no matter how much users ask for one. Explaining why there is no simple bullet answer is pointless as google already knows where everyone else already answered that. Years of bitter hard experience ramming my head against n00b user expectations has taught me never to hint at an answer - anything I say gets taken as gospel truth and a bunch of folk are now going to read this and interpret it as saying "Alan says /var must be 1.5G". How much swap space? is another such question. Perhaps a better first answer than "Only you know that" is "Only you know that and you need to know how to calculate it. You find that out by reading the fine manual." We're in Unix land here. As such, my answers thus far are totally appropriate. -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] building packages remotely
On Fri, Jun 12, 2009 at 14:08, Maxim Wexler wrote: > Hi group, > > I've read references here and in other forums to building packages on > a desktop PC and installing them on a note/netbook remotely as a way > of relieving stress on the smaller machine. > > Can someone point me to the documentation or howto? I can't seem to > come up with the proper google input that doesn't lead to garbage. > You may use DISTCC so other (more powerful) rigs can help compiling stuff (keep in mind some packages don' t use this, as it can lead to errors, gcc and openoffice, for instance). Or you can build binary packages and use a binary mirror so you can use "emerge -k". -- Daniel da Veiga
[gentoo-user] Introduce Manual manipulation during an emerge
I'm really sorry to keep beating on this portage stuff and I guess I must be something of a dimwit since I find just about anything to do with portage and emerge that is outside `emerge -flags whatever' to be really hard to catch on to, even though (and shouldn't admit this) I've been running gentoo for at least 5 yrs and probably more. I read up on the parts I need at times when I run into trouble and get much coaching and pointers here. But then in a few mnths, with no problems I've forgotten the vast bulk of whatever I picked up. And I mean even when I've made notes... Anyway cutting to the chase: I want to do some manual manipulation to a failing emerge of procmail. emerge -v procmail (details here were posted elsewhere but involves somekind missnaming of getline in the ebuild) It fails but has a simple enough fix. It needs to have a sed exchange take place in the unpacked source. Running `sed -i -e 's/getline/get_line/g' src/*.c src/*.h' In top level of the upacked source will allow it to compile fine. So once all instances of getline are changed to get_line... it works. Now how do I go about doing that right during the emerge? For something like this it seems way overkill to have a separate custom /usr/local/portage where I build my own. Especially since I find that whole process difficult and way overkill for this problem. Further its something that will almost certainly be fixed soon and won't be something I have to attend to again. I'm pretty sure there is some way to introduce custom action during emerge and remember doing something like that before to fix cvs so root could commit. It envolved either a ./configure or make flag that portage didn't employ... somehow, with coaching here... I was able to introduce a non-default flag. Isn't there a simple way to introduce the sed run in procmail sources during emerge? Oh, and to fix cvs again too... with this new install I've started from scratch after nursing one original install and upating over yrs.
Re: [gentoo-user] conflict in fstab w/ lvm?
Am Freitag 12 Juni 2009 21:54:45 schrieb Daniel da Veiga: > > Is this legal? Mounting two things at the same place? > > AFAIK, no. Yes. However, unless you do union mounts, you'll only see what's mounted last. Bye... Dirk signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] Introduce Manual manipulation during an emerge
On 12 Jun 2009, at 20:59, Harry Putnam wrote: ... For something like this it seems way overkill to have a separate custom /usr/local/portage where I build my own. Especially since I find that whole process difficult and way overkill for this problem. Further its something that will almost certainly be fixed soon and won't be something I have to attend to again. If you think that's the case then just mask the current version of procmail in /etc/portage/package.mask & remerge procmail. An older version without this bug will be installed & when an updated version is available portage will install it automagically. I consider /usr/local/portage to be essential, because there's most always a package or three on my systems which I can't get through the regular tree, or which i wish to version bump myself. I don't consider this a hassle because I could create the updated ebuild with patch in a few minutes. Because I most always check some reference resources when I do this, I can't quickly explain to you how to maintain this tree yourself. I think the time it would take you to learn this would be well-spent, as I'm sure you'll eventually need to use a /usr/local/portage package again in the future. Like I say, once you know how, it becomes easy. But if you consider it a hassle, just mask the buggy version of procmail & forget about the problem. Stroller.
[gentoo-user] kde 3.5 packages blocking each other
Hi
As I'm still on KDE 3.X, I've got several package updates, but most of them
are blocking each other:
[blocks B ] kde-base/kcontrol:3.5 ("kde-base/kcontrol:3.5" is blocking
kde-base/kdebase-3.5.9-r4)
[blocks B ] kde-base/kicker:3.5 ("kde-base/kicker:3.5" is blocking
kde-base/kdebase-3.5.9-r4)
[blocks B ] kde-base/libkonq:3.5 ("kde-base/libkonq:3.5" is blocking
kde-base/kdebase-3.5.9-r4)
[blocks B ] kde-base/kdeedu ("kde-base/kdeedu" is blocking
kde-base/kdelibs-3.5.10-r6)
[blocks B ] kde-base/konqueror:3.5 ("kde-base/konqueror:3.5" is blocking
kde-base/kdebase-3.5.9-r4)
[blocks B ] kde-base/kdialog:3.5 ("kde-base/kdialog:3.5" is blocking
kde-base/kdebase-3.5.9-r4)
[blocks B ] =kde-base/kdebase-3.5* ("=kde-base/kdebase-3.5*" is blocking
kde-base/kmenuedit-3.5.10, kde-base/kdialog-3.5.10,
kde-base/kdebase-kioslaves-3.5.10-r1, kde-base/kicker-3.5.10-r1,
kde-base/kdebase-data-3.5.10, kde-base/kfind-3.5.10,
kde-base/kcminit-3.5.10, kde-base/kdesu-3.5.10, kde-base/konqueror-3.5.10,
kde-base/khotkeys-3.5.10, kde-base/khelpcenter-3.5.10,
kde-base/libkonq-3.5.10, kde-base/kcontrol-3.5.10)
[blocks B ] kde-base/kdeaddons ("kde-base/kdeaddons" is blocking
kde-base/kdelibs-3.5.10-r6)
[blocks B ] kde-base/kdeutils ("kde-base/kdeutils" is blocking
kde-base/kdelibs-3.5.10-r6)
[blocks B ] kde-base/kcminit:3.5 ("kde-base/kcminit:3.5" is blocking
kde-base/kdebase-3.5.9-r4)
[blocks B ] kde-base/kde ("kde-base/kde" is blocking
kde-base/kdelibs-3.5.10-r6)
[blocks B ] kde-base/kdetoys ("kde-base/kdetoys" is blocking
kde-base/kdelibs-3.5.10-r6)
[blocks B ] kde-base/kdebase-kioslaves:3.5
("kde-base/kdebase-kioslaves:3.5" is blocking kde-base/kdebase-3.5.9-r4)
[blocks B ] kde-base/kfind:3.5 ("kde-base/kfind:3.5" is blocking
kde-base/kdebase-3.5.9-r4)
[blocks B ] kde-base/kdepim ("kde-base/kdepim" is blocking
kde-base/kdelibs-3.5.10-r6)
[blocks B ] kde-base/kdebase ("kde-base/kdebase" is blocking
kde-base/kdelibs-3.5.10-r6)
[blocks B ] kde-base/kmenuedit:3.5 ("kde-base/kmenuedit:3.5" is blocking
kde-base/kdebase-3.5.9-r4)
[blocks B ] kde-base/kdegraphics ("kde-base/kdegraphics" is blocking
kde-base/kdelibs-3.5.10-r6)
[blocks B ] kde-base/kdewebdev ("kde-base/kdewebdev" is blocking
kde-base/kdelibs-3.5.10-r6)
[blocks B ] kde-base/kdegames ("kde-base/kdegames" is blocking
kde-base/kdelibs-3.5.10-r6)
[blocks B ] kde-base/khotkeys:3.5 ("kde-base/khotkeys:3.5" is blocking
kde-base/kdebase-3.5.9-r4)
[blocks B ] kde-base/kdenetwork ("kde-base/kdenetwork" is blocking
kde-base/kdelibs-3.5.10-r6)
[blocks B ] kde-base/kdebase-data:3.5 ("kde-base/kdebase-data:3.5" is
blocking kde-base/kdebase-3.5.9-r4)
[blocks B ] kde-base/kdemultimedia ("kde-base/kdemultimedia" is blocking
kde-base/kdelibs-3.5.10-r6)
[blocks B ] kde-base/kdeadmin ("kde-base/kdeadmin" is blocking
kde-base/kdelibs-3.5.10-r6)
[blocks B ] kde-base/kdesu:3.5 ("kde-base/kdesu:3.5" is blocking
kde-base/kdebase-3.5.9-r4)
[blocks B ] kde-base/kdeartwork ("kde-base/kdeartwork" is blocking
kde-base/kdelibs-3.5.10-r6)
[blocks B ] kde-base/khelpcenter:3.5 ("kde-base/khelpcenter:3.5" is
blocking kde-base/kdebase-3.5.9-r4)
It looks like there's some issue in some packages being 3.5.10 and all the
rest being 3.5.9, but I have no keywords set in
/etc/portage/packages.keywords - perhaps I should.
Any hints?
Thanks a lot
Francisco
--
"If you have an apple and I have an apple and we exchange apples then you
and I will still each have one apple. But if you have an idea and I have one
idea and we exchange these ideas, then each of us will have two ideas." -
George Bernard Shaw
Re: [gentoo-user] Lost free space on /
Am Freitag 12 Juni 2009 21:56:04 schrieb Alan McKinnon: > There's a few guidlelines one can give (but only a few). The variables tend > to be large than the amounts with guidelines though. > > /var/tmp/portage should be at least 1G on a modern system, 6G+ if building > mozilla stuff and OOo is something you intend to do. BTW: OOo 3.1 seems to raise the bar to 8.5G. > portage cache is about 200M > ccache needs as much space as it was given in make.conf > /var/mail or /var/spool/mail is completely dependant on number of user and > how much mail they get. Often none for a desktop and huge amounts for a > mail store. > mysql and postgres need as much as the amount of data intended to be > stored. log space is very big or not too much depending on what you do. And, not to forget, one can always either resize it later (if LVM is used) or simply mount more space below /var, which also gives the opportunity to use different filesystems, depending on what kind of data is stored. Bye... Dirk signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
>>> OK, thank you Norman. The reason I'm trying to switch (this is my >>> third serious attempt) is some kind of a bug that crashes the system >>> when SMP is enabled and the madwifi driver is in master mode. I've >>> been running without SMP, but I could really use the extra power. >> >> That's interesting. I had an old 4 x processor machine running as an >> access-point (madwifi or madwifi-ng) running in master mode for at least a >> year or two. It was unstable as heck, and I never attributed it to this. It >> would, however, stay up for days or weeks at a time. >> >> Maybe this bug has crept in more recently? I'm not sure that it will apply >> to my new system (on which I'd like to run an AP, as soon as I get round to >> it) as that is a single processor P4. >> >>> Do you know if there is better Gentoo support for this on the horizon? >> >> I did find the dev uberlord immensely helpful when I was first doing this. >> He was the baselayout guy at the time, although I don't know if he still is >> or if you might be able to get hold of him. >> >> IMO the first thing to do is get the AP up & running without resort to the >> Gentoo init.d scripts. Try allocating an IP address to wlan0 just using >> `ifconfig` as root. If that works then you know the hardware & principles of >> operation are all ok. >> >> Stroller. >> >> > I would recomment the same thing. Play around manualy. Find out what works > and what does not. And if you found a manual way, you can start > scriptworkarounds and automating things. Thanks everyone. This system is critical so I think I'm better off sticking with madwifi and no SMP for now. I just upgraded to 2.6.29 so I thought things might be ready. Which software component should I be on the lookout for as far as Gentoo being ready to integrate smoothly with ath5k? baselayout? - Grant
Re: [gentoo-user] kde 3.5 packages blocking each other
Am Freitag 12 Juni 2009 22:24:26 schrieb Francisco Ares: > It looks like there's some issue in some packages being 3.5.10 and all the > rest being 3.5.9, but I have no keywords set in > /etc/portage/packages.keywords - perhaps I should. Partly. It seems to be a mix of split ebuilds and monolithic ones. You should decide which one to use. The other part is indeed a version mix, because the monolithic ebuilds are still at 3.5.9. HTH... Dirk signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] conflict in fstab w/ lvm?
On Fri, 12 Jun 2009 13:45:04 -0600 Maxim Wexler wrote: > #shm /dev/shmtmpfs nodev,nosuid,noexec 0 0 I wonder, what's the rationale behind commenting out shm? -- Mike Kazantsev // fraggod.net signature.asc Description: PGP signature
Re: [gentoo-user] Can I exclude a package from --depclean's consideration?
On Fri, 12 Jun 2009 15:11:07 -0400
David wrote:
> What would I add to /etc/portage/sets.conf to exclude gcc from depclean?
> thanks
I'd add these to sets:
[gcc-preserve]
class = portage.sets.shell.CommandOutputSet
command = /usr/local/sbin/gcc-list
This to /usr/local/sbin/gcc-list:
#!/bin/sh
for PKG in `ls -1 /var/db/pkg/sys-devel | grep -E '^gcc-[[:digit:].]+(-r.)?$'`
do echo '=sys-devel/'${i}
done
And '@gcc-preserve' to /var/lib/portage/world_sets
Alternatively, you can define set as files in /usr/libexec/gcc:
[gcc-preserve]
class = portage.sets.dbapi.OwnerSet
world-candidate = False
files = /usr/libexec/gcc
Looks simplier, but somewhat dirty and probably a bit slower.
--
Mike Kazantsev // fraggod.net
signature.asc
Description: PGP signature
Re: [gentoo-user] kde 3.5 packages blocking each other
And how do I tell if an ebuild is monolithic or not? Thanks Francisco -- "If you have an apple and I have an apple and we exchange apples then you and I will still each have one apple. But if you have an idea and I have one idea and we exchange these ideas, then each of us will have two ideas." - George Bernard Shaw
Re: [gentoo-user] Introduce Manual manipulation during an emerge
On Fri, 12 Jun 2009 14:59:53 -0500 Harry Putnam wrote: > Isn't there a simple way to introduce the sed run in procmail sources > during emerge? man 1 ebuild ebuild /usr/portage/mail-filter/procmail configure (do-some-sed-in-/var/tmp) ebuild /usr/portage/mail-filter/procmail merge Ta da! ;) -- Mike Kazantsev // fraggod.net signature.asc Description: PGP signature
Re: [gentoo-user] Can I exclude a package from --depclean's consideration?
Mike Kazantsev wrote:
On Fri, 12 Jun 2009 15:11:07 -0400
David wrote:
What would I add to /etc/portage/sets.conf to exclude gcc from depclean?
thanks
I'd add these to sets:
[gcc-preserve]
class = portage.sets.shell.CommandOutputSet
command = /usr/local/sbin/gcc-list
This to /usr/local/sbin/gcc-list:
#!/bin/sh
for PKG in `ls -1 /var/db/pkg/sys-devel | grep -E '^gcc-[[:digit:].]+(-r.)?$'`
do echo '=sys-devel/'${i}
done
And '@gcc-preserve' to /var/lib/portage/world_sets
Alternatively, you can define set as files in /usr/libexec/gcc:
[gcc-preserve]
class = portage.sets.dbapi.OwnerSet
world-candidate = False
files = /usr/libexec/gcc
Looks simplier, but somewhat dirty and probably a bit slower.
Thanks Mike,
I was hoping it was as simple as the kernel-sources example. I got
caught a few months back by dep cleaning gcc :( I did not want to
rebuild everything with the latest gcc at that time but my hand was
forced at that point.
-david
--
Powered by Gentoo GNU/Linux
http://linuxcrazy.com
[gentoo-user] Re: Introduce Manual manipulation during an emerge
Stroller writes: > But if you consider it a hassle, just mask the buggy version of > procmail & forget about the problem. Thanks... I thought of something like that but then noticed there's only one version available in portage. At that point I downloaded the previous version *21* tar ball and attempted to build it just manually but hit the same getline problem. (this was outside emerge) The next version back behind 21 is quite old. (I then built an installed version 22* by hand using the sed tools ... more below about that) But can't I do something when the emerge breaks and tells me where the sources are in /var/tmp... can't I go there and finish the build somehow? First the sed run, then ebuild (I know those ebuild commands) and have the finished product installed with emerge? And similarly with CVS. I remember there being some way I could tell emerge to let me set a ./configure flag by hand... then finish the install (with emerge). May have been some flag set right at the emerge cmd like: # SOMEFLAG=something emerge -v cvs Just for information... I have built procmail *22* by hand using the afore mentioned sed command then symlinked /usr/local/bin/procmail to /usr/bin/procmail, where sendmail expects it to be, and so am able to run it fine and get mail working. So not really a big problem... just seems there'd be a fairly easy way to get this done with emerge short of going the personal overlay route. To me... the personal overlay just does not fall under `fairly easy'.
Re: [gentoo-user] kde 3.5 packages blocking each other
Am Freitag 12 Juni 2009 22:45:49 schrieb Francisco Ares: > And how do I tell if an ebuild is monolithic or not? The monolithic ones install larger parts of KDE, and usually have the same names as the original source packages offered at KDE.org. The split ebuilds, well, split those packages into their individual applications, so you have ebuilds for konqueror (which is also part of kdenetwork) or kmail (kdepim). In addition, there are the "-meta" ebuilds, which have the same name as the monolitic ones, but with -meta appended (kdepim-meta). Those usually install the same applications than monolithic ebuilds, but as split ebuilds. So, when you install kde, you get a complete KDE from monolithic ebuilds and when you install kde-meta, you get a complete KDE from split ebuilds. That's also the reason why they block each other. When you have kdepim installed, you already got kmail, so you shouldn't install kmail from the split ebuild again. HTH... Dirk signature.asc Description: This is a digitally signed message part.
[gentoo-user] Re: Introduce Manual manipulation during an emerge
Mike Kazantsev writes: > On Fri, 12 Jun 2009 14:59:53 -0500 > Harry Putnam wrote: > >> Isn't there a simple way to introduce the sed run in procmail sources >> during emerge? > > man 1 ebuild > ebuild /usr/portage/mail-filter/procmail configure > (do-some-sed-in-/var/tmp) > ebuild /usr/portage/mail-filter/procmail merge > > Ta da! ;) Yess indeed... Ta Da.. thank you... I new there was a simple way.
Re: [gentoo-user] Re: How to veiw absolute latest on partage without syncing
Harry Putnam wrote: > Dale writes: > > >> You could always sync then do a emerge -uv procmail. then it would only >> upgrade procmail and any friends that need to be updated. That would >> mostly likely miss most of the other updates that you are wanting to >> skip for the moment. >> > > yeah... its a thought... but why emerge world then procmail again when > I know it won't emerge unless something is done about the getline bug? > > I wanted to check in most recent portage to see if any changes were > made to procmail. > > > > > I was thinking about NOT doing the emerge -u world part. That would skip updating everything that has updates applied on your system. Doing just a emerge -u procmail would only update procmail and the dependencies if any are needed. Keep in mind, you can upgrade packages as needed without updating the whole system. I wouldn't recommend this long term tho. Those upgrades can add up pretty quick and come back to bite you later. Going longer than a month or two on a Gentoo system without a update is not recommended. Dale :-) :-)
[gentoo-user] Re: Introduce Manual manipulation during an emerge
Mike Kazantsev writes: > man 1 ebuild > ebuild /usr/portage/mail-filter/procmail configure > (do-some-sed-in-/var/tmp) > ebuild /usr/portage/mail-filter/procmail merge Many thanks for the tips... and hugely usefull. But: Yikes I may have jumped the gun thinking I was good to go but at least I now have the commands to get a little further... unfortunately the build breaks for a different reason now. ebuild \ /usr/portage/mail-filter/procmail/procmail-3.22-r10.ebuild configure cd to source: /var/tmp/portage/mail-filter/procmail-3.22-r10/work/procmail-3.22 sed -i -e 's/getline/get_line/g' src/*.c src/*.h (all good so far) ebuild \ /usr/portage/mail-filter/procmail/procmail-3.22-r10.ebuild merge [...] tail of error below: Whoeaaa! There's something fishy going on here. You have a look and see if you detect anything uncanny: *** i486-pc-linux-gnu-gcc -O2 -march=i486 -pipe -fno-inline-functions2 -march=i486 -pipe -fno-inline-functions _autotst.c -o _autotst -Wl,-O1 cc1: error: unrecognized command line option "-fno-inline-functions2" cc -O2 -march=i486 -pipe -fno-inline-functions2 -march=i486 -pipe -fno-inline-functions _autotst.c -o _autotst -Wl,-O1 cc1: error: unrecognized command line option "-fno-inline-functions2" gcc -O2 -march=i486 -pipe -fno-inline-functions2 -march=i486 -pipe -fno-inline-functions _autotst.c -o _autotst -Wl,-O1 cc1: error: unrecognized command line option "-fno-inline-functions2" *** I suggest you take a look at the definition of CFLAGS* and CC in the Makefile before you try make again. make[1]: *** [init] Error 1 make[1]: Leaving directory `/var/tmp/portage/mail-filter/procmail-3.22-r10/work/procmail-3.22' make: *** [src/Makefile] Error 2 * * ERROR: mail-filter/procmail-3.22-r10 failed. * Call stack: * ebuild.sh, line 49: Called src_compile * environment, line 2531: Called die * The specific snippet of code: * emake CC="$(tc-getCC)" || die * The die message: * (no error message)
[gentoo-user] Re: How to veiw absolute latest on partage without syncing
Dale writes: > I was thinking about NOT doing the emerge -u world part. That would > skip updating everything that has updates applied on your system. Doing > just a emerge -u procmail would only update procmail and the > dependencies if any are needed. > > Keep in mind, you can upgrade packages as needed without updating the > whole system. I wouldn't recommend this long term tho. Those upgrades > can add up pretty quick and come back to bite you later. Going longer > than a month or two on a Gentoo system without a update is not > recommended. Sorry I failed to mention a world upgrade at this point involved 1 file, but I take your point still mine also holds. Why keep banging on procmail when by looking up the newest ebuild online I can see it is not any different... hence the bug will remain. I still say its handy to be able to look at the latest portage pkgs when you feel the urge. Nicer and less labor intensive than upgrading anyghing really.
Re: [gentoo-user] How to veiw absolute latest on partage without syncing
On Fri, Jun 12, 2009 at 13:45, Mark Knecht wrote: > On Fri, Jun 12, 2009 at 9:07 AM, Stroller > wrote: >> >> On 12 Jun 2009, at 15:40, Paul Hartman wrote: >>> >>> On Fri, Jun 12, 2009 at 9:29 AM, Justin wrote: Harry Putnam schrieb: > > Is there a way to veiw the very latest packages on portage without > syncing my OS? > > http://sources.gentoo.org/viewcvs.py/gentoo-x86/ >>> >>> also http://packages.gentoo.org/ >>> >>> or http://gentoo-portage.com/Newest >> >> The problem with this is that it's difficult to determine which packages on >> one's own system have updated. One must check individually for each atom in >> world. >> >> >> Harry: >> >> I'm not sure if it's possible _without_ syncing, but you can `cp -a >> /usr/portage /usr/portage.orig`, sync, `emerge -pv world` and then move the >> original tree back if you want to. >> >> It's not really clear why you're asking, or why you're unable to sync. If >> the PC has no internet connection, for instance, security updates are >> unimportant. >> >> Stroller. > > I've wanted a way to do something like this for a long time. One > problem with the way portage works with ( I guess) rsync or whatever > it uses is that when someone decides to remove a package from portage > that I'm currently using syncing removes it from my system also. > Unfortunately before I do the sync I have no idea it has been removed > so I don't know that it's going to get taken off my system. Once it > does I can go find a copy and put it in a personal overlay but that > requires I do the work after the damage is done. It would be nice if > there was a message ahead of time that told me certain packages were > going to be removed, etc., before it was actually done, but I > understand from previous conversations that syncing doesn't work that > way. > > This has come up numerous times for me on older hardware where, for > instance, maybe some on-board graphics chip only works with older ATI > drivers, and that ATI driver only works with older kernels. By the > time sync is done I've lost the code for what my system is running, > and unfortunately there's no messages that this is happening when I'm > doing the sync so maybe I only figure it out a few weeks later and > then have to mess around building an overlay using the attic. > Portage keeps a copy of installed packages under /var/db/pkg, AFAIK. So, even if sync removes it from the tree, you can move it from /var to your local overlay and keep using it... If you are doying a fresh install, you can get the old ebuilds from the attic. -- Daniel da Veiga
Re: [gentoo-user] conflict in fstab w/ lvm?
On 6/12/09, Daniel da Veiga wrote: > On Fri, Jun 12, 2009 at 16:45, Maxim Wexler wrote: >> Hi group, >> >> Following the LVM2 gentoo doc I have in fstab: >> >> ... >> /dev/vg/tmp /tmp ext2 noatime 0 2 >> ... >> >> But also(suggested by the eee forum): >> >> ... >> #shm/dev/shmtmpfs nodev,nosuid,noexec 0 0 >> tmpfs /tmptmpfs defaults,noatime,mode=1777 0 0 >> >> Is this legal? Mounting two things at the same place? >> > > AFAIK, no. > First off, what do you want to do? The EEE forum suggested mounting I want to create a useful, trouble-free genteee box. mw
Re: [gentoo-user] conflict in fstab w/ lvm?
On 6/12/09, Mike Kazantsev wrote: > On Fri, 12 Jun 2009 13:45:04 -0600 > Maxim Wexler wrote: > >> #shm /dev/shmtmpfs nodev,nosuid,noexec 0 0 > > I wonder, what's the rationale behind commenting out shm? > Good question. I was given to understand the new line was intended to replaced the default, which I commented out. Perhaps that's a mistake. That's how I configured the previous iteration of genteee before it went south; maybe the new line had something to do with it. Should I use both? mw
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
On Friday 12 June 2009, Stroller wrote: > At that time the only other 802.11g driver that did master mode was, I > think, Prism54 and it was a little difficult to get hold of cards > featuring that chipset (consequently I got into the side-business of > selling them, and probably have 20 left here). madwifi was better > because it featured "virtual APs" (VAPs) and allowed you to run > separate WEP & unencrypted wireless networks on the same card (and run > iptables rules on the interface allocated to each VAP). > > So I'm not sure why you're changing from madwifi to ath5k. Well, I understand that once you move to 2.6.29 there's no choice of madwifi any more? I tried to emerge it and from what I recall was told to enable ath5k in the kernel - which as you say is not as powerful as madwifi was. -- Regards, Mick signature.asc Description: This is a digitally signed message part.
[gentoo-user] Prioritizing mpd
When I use the medium quality libsamplerate resampler with mpd, my CPU is around 15% and all is well. When I try to use the best quality resampler, the CPU stays around 99% and the sound frequently falls apart. Can I give mpd CPU priority? - Grant
Re: [gentoo-user] Re: Introduce Manual manipulation during an emerge
On Fri, 2009-06-12 at 16:22 -0500, Harry Putnam wrote: > i486-pc-linux-gnu-gcc -O2 -march=i486 -pipe -fno-inline-functions2 > -march=i486 -pipe -fno-inline-functions _autotst.c -o _autotst -Wl,-O1 > cc1: error: unrecognized command line option "-fno-inline-functions2" What do put in your CFLAGS (e.g., in /etc/make.conf)? Typos there? -- Arttu V.
[gentoo-user] Re: Introduce Manual manipulation during an emerge
"Arttu V." writes:
> On Fri, 2009-06-12 at 16:22 -0500, Harry Putnam wrote:
>
>> i486-pc-linux-gnu-gcc -O2 -march=i486 -pipe -fno-inline-functions2
>> -march=i486 -pipe -fno-inline-functions _autotst.c -o _autotst
>> -Wl,-O1 cc1: error: unrecognized command line option
>> "-fno-inline-functions2"
>
> What do put in your CFLAGS (e.g., in /etc/make.conf)? Typos there?
Here is what is in there right now.. apparently from the stage3
pulled down during install a few days ago.
CFLAGS="-O2 -march=i486 -pipe"
CXXFLAGS="-O2 -march=i486 -pipe"
CHOST="i486-pc-linux-gnu"
Looking at make.conf from an old backup I see:
CFLAGS="-O2 -march=i686 -pipe"
CHOST="i686-pc-linux-gnu"
CXXFLAGS="${CFLAGS}"
So trying that now.
Haa finished up smoothly with those flags changed.
Do you think I do something like `emerge -vuDN world' since everthing
was compiled up to now with the old flags shown above?
Re: [gentoo-user] kde 3.5 packages blocking each other
Thanks a lot! Francisco On Fri, Jun 12, 2009 at 5:58 PM, Dirk Heinrichs wrote: > Am Freitag 12 Juni 2009 22:45:49 schrieb Francisco Ares: > > > And how do I tell if an ebuild is monolithic or not? > > The monolithic ones install larger parts of KDE, and usually have the same > names as the original source packages offered at KDE.org. > > The split ebuilds, well, split those packages into their individual > applications, so you have ebuilds for konqueror (which is also part of > kdenetwork) or kmail (kdepim). In addition, there are the "-meta" ebuilds, > which have the same name as the monolitic ones, but with -meta appended > (kdepim-meta). Those usually install the same applications than monolithic > ebuilds, but as split ebuilds. > > So, when you install kde, you get a complete KDE from monolithic ebuilds > and > when you install kde-meta, you get a complete KDE from split ebuilds. > > That's also the reason why they block each other. When you have kdepim > installed, you already got kmail, so you shouldn't install kmail from the > split ebuild again. > > HTH... > >Dirk > > > -- "If you have an apple and I have an apple and we exchange apples then you and I will still each have one apple. But if you have an idea and I have one idea and we exchange these ideas, then each of us will have two ideas." - George Bernard Shaw
Re: [gentoo-user] Re: Introduce Manual manipulation during an emerge
On Fri, 2009-06-12 at 18:39 -0500, Harry Putnam wrote:
> Here is what is in there right now.. apparently from the stage3
> pulled down during install a few days ago.
>
> CFLAGS="-O2 -march=i486 -pipe"
> CXXFLAGS="-O2 -march=i486 -pipe"
> CHOST="i486-pc-linux-gnu"
Is there a reason to use i486 stage3? I think an i686 one might have
been available and a better hit if your system is/was set up as an i686
before this? Well, not that it counts now, gotta go with what you have
unpacked.
> Looking at make.conf from an old backup I see:
>
> CFLAGS="-O2 -march=i686 -pipe"
> CHOST="i686-pc-linux-gnu"
> CXXFLAGS="${CFLAGS}"
>
> So trying that now.
>
> Haa finished up smoothly with those flags changed.
>
> Do you think I do something like `emerge -vuDN world' since everthing
> was compiled up to now with the old flags shown above?
The friendly gentoo devs have whole guide for this:
http://www.gentoo.org/doc/en/change-chost.xml
And, as you can see, it starts with the soothing words "Changing the
CHOST is a big issue that can seriously screw up your system" ...
I'd go through that guide first to get to a (hopefully) sane system, and
continue with other emergings only then.
--
Arttu V.
Re: [gentoo-user] building packages remotely
On Fri, Jun 12, 2009 at 1:08 PM, Maxim Wexler wrote: > Hi group, > > I've read references here and in other forums to building packages on > a desktop PC and installing them on a note/netbook remotely as a way > of relieving stress on the smaller machine. > > Can someone point me to the documentation or howto? I can't seem to > come up with the proper google input that doesn't lead to garbage. > > Maxim Well, if your systems are VERY similar (chost, cflags, very similar selection of packages, etc) you can use: emerge --buildpkgonly some/thing to build packages, but I personally recommend putting together a chroot to build in for your netbook (I recommend it only, really, because I know it to work as I use it with virtual systems), and using it to build packages. The process isn't too difficult... and is really a lot like any other install. make a directory to hold it, extract an appropriate stage3 (might look at the weekly builds to save a lot of time on updating things), add buildpkg to your FEATURES, build anything you need, possibly even taking the time to do an -- emerge -ev --buildpkgonly world to get up to date packages for everything, then make those packages available to your weaker system through some means (ftp, http, or nfs mounted over /usr/portage/packages). And make sure to always use "emerge -k whatever" to make sure it uses the packages. Also, USE flags should match between the real weaker system and the chroot you built for it. You could also reinstall the weaker system from scratch by treating the chroot as, basically, a stage4 ... leaving you only a need to worry about bootloader, config files, and the kernel being configured and built properly for your needs. A similar, but secondary, option would be to start building for a second system using the host system's compiler and portage, building into a secondary 'ROOT', which I tend to do with systems that have no need at all for a compiler, portage tree, etc, and building packages out of those in the process. -- Poison [BLX] Joshua M. Murphy "After three days without programming, life becomes meaningless." - The Tao Of Programming
Re: [gentoo-user] conflict in fstab w/ lvm?
On Fri, Jun 12, 2009 at 18:46, Maxim Wexler wrote: > On 6/12/09, Daniel da Veiga wrote: >> On Fri, Jun 12, 2009 at 16:45, Maxim Wexler wrote: >>> Hi group, >>> >>> Following the LVM2 gentoo doc I have in fstab: >>> >>> ... >>> /dev/vg/tmp /tmp ext2 noatime 0 2 >>> ... >>> >>> But also(suggested by the eee forum): >>> >>> ... >>> #shm /dev/shm tmpfs nodev,nosuid,noexec 0 0 >>> tmpfs /tmp tmpfs defaults,noatime,mode=1777 0 0 >>> >>> Is this legal? Mounting two things at the same place? >>> >> >> AFAIK, no. >> First off, what do you want to do? The EEE forum suggested mounting > > I want to create a useful, trouble-free genteee box. > You have only two choices, being an eee user myself, and having it upgraded to 2GB RAM, I choose the tempfs filesystem for /tmp (RAM) instead of keeping temporary files writen and deleted from my poor SSD. If you have low RAM, you can decide to leave it on the SSD and thus give more room for app data on RAM. -- Daniel da Veiga
Re: [gentoo-user] conflict in fstab w/ lvm?
On Fri, Jun 12, 2009 at 5:52 PM, Maxim Wexler wrote: > On 6/12/09, Mike Kazantsev wrote: >> On Fri, 12 Jun 2009 13:45:04 -0600 >> Maxim Wexler wrote: >> >>> #shm /dev/shm tmpfs nodev,nosuid,noexec 0 0 >> >> I wonder, what's the rationale behind commenting out shm? >> > > Good question. I was given to understand the new line was intended to > replaced the default, which I commented out. Perhaps that's a mistake. > That's how I configured the previous iteration of genteee before it > went south; maybe the new line had something to do with it. Should I > use both? > > mw Hmm. 1) a tmpfs space is, by default, mounted on /dev/shm to meet some standard somewhere (can't recall, FHS I think). The important thing to note is that the name 'shm' is basically an unused placeholder (tmpfs doesn't operate on an actual block device like /dev/hda1), and that /dev/shm is the mount *point*. It should be there, and uncommented. 2) Yes it's 'legal' to mount the lvm volume onto /tmp *and* tmpfs space as you have your fstab lines there, but I can't say for sure which would truly be mounted first and which second, and in turn which would actually be used in the running system. IF you intend to use your system RAM to reduce read/write on your drive for temporary files, comment out the use of the LVM volume on /tmp and just leave the tmpfs mount on that point active (commenting leaves you free to change your mind anytime you like). 3) Vaguely related to your mention of it 'taking its place' about the /dev/shm and /tmp tmpfs mounts, the only time I've seen that mentioned was in a conversation somewhere about 'why not just use a --bind mount of /dev/shm onto /tmp to put it in tmpfs' ... which was answered with the simple fact that, by default everywhere I've seen it, /dev/shm is mounted noexec, while it's not altogether uncommon for things to be decompressed into /tmp before execution (which would fail if /tmp were mounted noexec). -- Poison [BLX] Joshua M. Murphy "Without a struggle, there can be no progress." - Frederick Douglass
Re: [gentoo-user] KDE menu missing: not solved !
090612 Mick wrote: > I've rebooted twice with same kernel (gentoo-2.6.29-r5) > and the settings seem to have stuck. > Not sure what I've done differently before to cause the settings to be lost. No changes of kernel here (same as yours). I've experimented with the same experience as you : (1) (with damaged Krusader 'open with' in Fluxbox): restart in KDE: won't start KDE Control Centre; restore 'open with' via Krusader: ok; (2) restart KDE: ok; (3) restart FB: Krusader 'open with' ok, incl use of Kview for JPG, which I didn't tell it to do under (1) ! (4) restart KDE: ok; (5) logout user, login, restart KDE: ok, restart FB: ok; (6) reboot, restart KDE: ok ! (7) restart FB: Krusader 'open with' ok ! So it's not something happening in the login or boot processes as such. -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
Re: [gentoo-user] KDE menu missing: not solved !
Further steps reveal something re the problem: (8) reboot: restart KDE: ok; (9) restart FB: Krusader 'open with' ok (yes, these repeat steps 6-7 in the previous msg); (10) reboot: restart FB : try Krusader: "Cannot talk to klauncher" ! (click to close msg box) when Krusader starts, 'open with' fails ! So something is being started by the KDE desktop which persists for FB, but if I start FB straight after rebooting, it doesn't get started & there's a failure of communication somewhere which affects Krusader. Does anyone have any suggestions ? -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
Re: [gentoo-user] KDE menu missing: not solved !
Yet another step reveals the locus of the problem here at least: (11) try to run 'klauncher' from CLI : "must be started by Kdeinit"; (12) restart in KDE, fix Krusader + Apwal ; (13) reboot & restart FB but without 'kdeinit &' in ~/.xinitrc : Krusader 'open with' ok ! So my problem seems to lie in starting 'kdeinit' without the KDE desktop. It will take another couple of reboots (tomorrow) to confirm this. Mick mb doing things a bit differently & his problem mb elsewhere. -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
[gentoo-user] Keyboard handling weird... in 2.6.30?
I don't know if it is from kernel 2.6.30 or if something else changed, but my keyboard does not behave as normal. It seems like the "key up" signal from the previous key is causing the repeat of the current key to get interrupted. This happens everywhere, not only in X, but in console as well. For example, try this: Type some text, like X and then hold left arrow. Your cursor will move to the left until you release the key. Now try to hold the right arrow after you've already been holding the left arrow. It will start to move to the right. Previously, it would continue moving to the right after you released the left arrow. Now, the keyboard repeat STOPS once you release the PREVIOUS key. So, in other words, the repeating of the right arrow is stopped when I release the left arrow... which causes the cursor to stop, and causes me to become aggravated. :P Has anyone else noticed this? I hope it's not a new "feature" :) Maybe later tonight I'll try to go back to 2.6.29 and see if this truly was the thing that brought on this change.
Re: [gentoo-user] conflict in fstab w/ lvm?
On Fri, 12 Jun 2009 21:52:20 -0400
Joshua Murphy wrote:
> On Fri, Jun 12, 2009 at 5:52 PM, Maxim Wexler wrote:
> > On 6/12/09, Mike Kazantsev wrote:
> >> On Fri, 12 Jun 2009 13:45:04 -0600
> >> Maxim Wexler wrote:
> >>
> >>> #shm /dev/shm tmpfs nodev,nosuid,noexec 0 0
> >>
> >> I wonder, what's the rationale behind commenting out shm?
> >
> > Good question. I was given to understand the new line was intended to
> > replaced the default, which I commented out. Perhaps that's a mistake.
> > That's how I configured the previous iteration of genteee before it
> > went south; maybe the new line had something to do with it. Should I
> > use both?
>
> Hmm.
> 1) a tmpfs space is, by default, mounted on /dev/shm to meet some
> standard somewhere (can't recall, FHS I think). The important thing to
> note is that the name 'shm' is basically an unused placeholder (tmpfs
> doesn't operate on an actual block device like /dev/hda1), and that
> /dev/shm is the mount *point*. It should be there, and uncommented.
>
...
>
> 3) Vaguely related to your mention of it 'taking its place' about the
> /dev/shm and /tmp tmpfs mounts, the only time I've seen that mentioned
> was in a conversation somewhere about 'why not just use a --bind mount
> of /dev/shm onto /tmp to put it in tmpfs' ... which was answered with
> the simple fact that, by default everywhere I've seen it, /dev/shm is
> mounted noexec, while it's not altogether uncommon for things to be
> decompressed into /tmp before execution (which would fail if /tmp were
> mounted noexec).
Indeed it should be there, it's as a shared memory for inter-process
communication (IPC). Many stuff uses shared memory, notably gcc and
multi-process daemons like apache, so you should give it to them.
And, as noted, tmpfs is not real device or even some single virtual
device. By "mount -t tmpfs none /tmp" you mount some piece of virtual
memory to a place but it's never the same piece, so you can have two,
ten or hundred tmpfs mounts completely independent of each other.
mkdir /mnt/{tmp1,tmp2}
mount -t tmpfs none /mnt/tmp1
mount -t tmpfs none /mnt/tmp2
touch /mnt/tmp1/some_file
ls -la /mnt/tmp1 (shows "some_file"
ls -la /mnt/tmp2 (empty)
So you don't have to bind everything into one tmpfs, just create as
many as you want, but, once again, especially if you chose not to have
swap, limit their size so they won't eat all your RAM!
Imagine scenario like this (or do "sync" and run it, but it should hang
your machine!):
mount -t tmpfs none /mnt/tmp1
dd if=/dev/zero of=/mnt/tmp1/some_file bs=1024 count=10
Your VM should go away and kernel 'll go on a killing spree, wiping
out all the runnuing processes, but, since tmpfs itself is not a
process, it'll just kill everything until panic or nothing's left at
all.
"-o size=512M" will just give you "No free space on disk" instead of
nasty crash. /tmp is world-writable, anything can choose to ditch a gig
or two into it for whatever reasons...
--
Mike Kazantsev // fraggod.net
signature.asc
Description: PGP signature
Re: [gentoo-user] Prioritizing mpd
On Fri, 12 Jun 2009 16:02:00 -0700
Grant wrote:
> When I use the medium quality libsamplerate resampler with mpd, my CPU
> is around 15% and all is well. When I try to use the best quality
> resampler, the CPU stays around 99% and the sound frequently falls
> apart. Can I give mpd CPU priority?
Yes, it's usually done via nice/renice commands:
renice -n -10 -p `pgrep mpd`
You can tune it's priority up to -20 (most real-time priority).
I'd suggest looking at load-average it generates ("top" shows it, at the
top)) first.
After running mpd for 15 minutes or so, if any of the three (5/10/15)
will go above number of physical CPU cores you have (and that's
probably the case if you see full load at any given time), tuning it's
priority up will make the rest of the system extremely sluggish, since
mpd won't let any other process to execute and just doing "ls" may take
ages, not to mention whole X operation...
--
Mike Kazantsev // fraggod.net
signature.asc
Description: PGP signature
Re: [gentoo-user] Prioritizing mpd
>> When I use the medium quality libsamplerate resampler with mpd, my CPU
>> is around 15% and all is well. When I try to use the best quality
>> resampler, the CPU stays around 99% and the sound frequently falls
>> apart. Can I give mpd CPU priority?
>
> Yes, it's usually done via nice/renice commands:
>
> renice -n -10 -p `pgrep mpd`
>
> You can tune it's priority up to -20 (most real-time priority).
>
> I'd suggest looking at load-average it generates ("top" shows it, at the
> top)) first.
> After running mpd for 15 minutes or so, if any of the three (5/10/15)
> will go above number of physical CPU cores you have (and that's
> probably the case if you see full load at any given time), tuning it's
> priority up will make the rest of the system extremely sluggish, since
> mpd won't let any other process to execute and just doing "ls" may take
> ages, not to mention whole X operation...
Thanks Mike. I tried:
renice -20 -p `pgrep mpd`
but my Athlon 2.2Ghz still can't handle it for more than a few
seconds. I don't have SMP enabled because of a bug in madwifi, and
I'm hoping when I get that fixed I'll be able to run the best
libsamplerate resampler. Any other ideas for making this work?
- Grant
Re: [gentoo-user] Prioritizing mpd
On Fri, 12 Jun 2009 21:45:02 -0700 Grant wrote: > renice -20 -p `pgrep mpd` > > but my Athlon 2.2Ghz still can't handle it for more than a few > seconds. I don't have SMP enabled because of a bug in madwifi, and > I'm hoping when I get that fixed I'll be able to run the best > libsamplerate resampler. Any other ideas for making this work? AFAIK resampling is expensive operation that's only necessary when your sound card can't handle native stream sample rate, furthermore, it's a lossy operation (degrading quality). So, I'd look for the answer to the question "why mpd is doing it and why I allow it to do that?". For example, you might have enabled it to resample stream to 32 bits depth, while your built-in card can only handle 16 and the stream has also 16, so what happens is userspace-level conversion (with some loss of quality) to 32, loading your CPU, then this stream goes to alsa, and, provided that your card can't play this, driver or the card itself converts it back to 16. Note that the latter case would probably mean "card offloads conversion to your CPU as well", so you'll get CPU load for both ways' conversion anyway, only reducing sound quality, no matter how good converters are. To avoid any processing, try disabling resampling in mpd, since it'll probably be done for you anyway, if necessary (you'll hear "white noise" otherwise). And you can pre-convert all the streams to any given samplerate, but note that you'll probably get far worse results if the target format isn't lossless (flac, ape), even if the source one is lossy, than with worst resampling. And you can get worse CPU/IO load with lossless format in the end, since it's harder to decode and the input data stream is much heavier than with lossy mp3s or oggs. -- Mike Kazantsev // fraggod.net signature.asc Description: PGP signature
[gentoo-user] Idle Process Scheduling
I'm having a strange problem on my Q6600 that cropped up starting with the 2.6.29 series of the kernel, and is still present in 2.6.30. Essentially, at all times, I have four nice 19 processes running, which for the sake of this post, we'll call "dnetc". All four cores are utilized. At this point, if I start another CPU-bound process that isn't niced, it begins to take up an entire core. This is expected. What isn't expected, however, is that another core begins idling inexplicably. As a result, despite 5 processes currently available to run, only 3 are actually running at any given time (the non-niced process, and two instances of dnetc). I have no idea where to begin diagnosing this, so if anyone has any pointers or knows anything, I'd like to hear about it. I've done numerous searches of mailing lists, bug trackers, etc., but haven't found anything. Maybe I just can't find the right keywords.
