Re: which files took the space

[tomas]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Thu, Mar 03, 2016 at 11:17:19PM -0500, Cindy-Sue Causey wrote:
> On 3/3/16, Andrew McGlashan  wrote:
> >
> > On 4/03/2016 3:07 AM, Adam Wilson wrote:
> >> On Fri, 4 Mar 2016 03:03:53 +1100 Andrew McGlashan
> >>  wrote:
> >>> It also may have been files in the file system, but where another file
> >>> system mount hides them
> >>
> >> What does this mean? Mounts overlapping and hiding other mounts?
> >>
> >> Explain, please.
> >
> > Yes, this is more likely to happen to the root file system.
> >
> > Say you have a bunch of files in /boot, but for some reason you have a
> > /boot partition that wasn't mounted when those files were created 
> > then you mount the /normal/ boot partition over it and now the other
> > files are now hidden from view, but still taking up space.
> 
> 
> Is that behavior as designed and thus expected, or is it a glitch?

If memory serves, long, long time ago, "mount" (the system call)
refused to comply unless the directory was empty (kernel 2.6.mumble;
so long ago. Expect bit flips and glitches in my wetware, yadda).

It was perceived as an artificial limitation which has no place in
the kernel. Actually, I can conceive use cases for mount shadowing
the content of a directory.

> My brain's thinking it should either complain and refuse to continue
> else obliterate and replace.

Nothing gets obliterated. Only shadowed...

> To me it would be... safer that it halt and complain rather than
> destroy, but all that shows is that I most likely just don't
> understand the function.
> 
> Do (and/or should) the original files "reappear" later?

They do -- at unmount. That's where the "missing" space is used,
after all :-)

> Guess I'm just thinking out loud again mostly because I actually
> understand the circumstance as presented. :)

Maling lists are just the hum of many people thinking loud. Keep doing
it :-)

regards
- -- t
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)

iEYEARECAAYFAlbZO3sACgkQBcgs9XrR2kYwNgCfdoVhuenZV6/wf+hGj2KkOIVd
KYsAniD1kTT85SjV1BTpGhR4cghb9yr8
=RjOC
-END PGP SIGNATURE-

Re: which files took the space

[jdd]

Le 04/03/2016 08:38, to...@tuxteam.de a écrit :


If memory serves, long, long time ago, "mount" (the system call)
refused to comply unless the directory was empty (kernel 2.6.mumble;
so long ago. Expect bit flips and glitches in my wetware, yadda).


if it existed, it's much older than that, never seen it since 1997...

man mount:

This  tells  the  kernel  to attach the filesystem found on device 
(which is of type type) at the directory dir.  The previous contents (if 
any) and owner and mode of dir become  invisible,  and  as  long  as 
this  filesystem  remains mounted, the pathname dir refers to the root 
of the filesystem on device."



jdd

Stable rare earth materials fr Ganzhou Wanfeng

[wayne]

Dear Purchasing Manager,
Glad to hear you are in the rare earth high-tech application market from 
Gooogle. 
This is Wayne Wang from Ganzhou Wanfeng Advanced Materials Tech. Co., Ltd., 
supplier of high quality& stable rare earth materials.
 We have our own professional engineers to meet any of your requirements. If 
any question, pls contact us for details.
Looking forward to your reply.
Best regards. 
Sincerely yours,
Wayne
Ganzhou Wanfeng Advanced Materials Technology Co., Ltd.
Office 1706, Yunxing Century Center, No. 8, Ganjiangyuan Ave., Zhanggong 
District, Ganzhou, Jiangxi, China  341000
Tel:+86 797 8286977  Mobile:+86 18979730852
Web site: www.gzwfen.comSkype: wa...@gzwfen.com
WhatsApp: +86 18979730852 E-mail: wa...@gzwfen.com

Re: Netinstall senza rete

[Lisi Reisz]

On Friday 04 March 2016 06:31:21 Fabrizio Carrai wrote:
> Ciao a tutti,
> Ho fatto una installazione Debian usando la Netinstall. Purtroppo al
> momento dell'installazione la rete non era disponibile e quindi potuto fare
> solo una installazione minimale. Tutto ok. Ora la connessione di rete è
> funzionante.
>
> Che cosa devo fare per completare l'installazione e scegliere gli altri
> pacchetti ?
>
> Grazie

This is the English speaking list. You want:

https://lists.debian.org/debian-italian/

HTH
Lisi

Re: Bluetooth headphones mistaken for a keyboard!?!

[deloptes]

Mark Fletcher wrote:

> deloptes  gmail.com> writes:
> 
>> 
>> Mark Fletcher wrote:
>> 
> 
>> > PC is
>> > a 7-year-old self-built desktop box running Jessie with Gnome as the
>> > DE,
>> 
>> Which version of debian are you on and which kernel?
>> 
> 
> Thanks for your time and attention. As I mentioned in my original post, I
> am running Jessie. The kernel is the stock x86_64 kernel currently in
> Jessie. I'm sorry but I'm not in front of the computer right now and am
> not sure exactly what version. I last updated last Sunday.
> 
>> You may need to tell udev what kind of device it is, but we don't know if
>> your system is configured properly
>> 
> 
> I'm fairly confident in my system configuration since I have been able to
> get my system to work as s Pulse SINK for an iPhone and Android SOURCE,
> which is reputed to be harder than what I am trying to do here. But I take
> your point, no proof there isn't some misconfiguration lurking somewhere.
> And I suppose it's possible that I jiggered something up, if you'll pardon
> the technical term, when I was getting the iPhone / Android thing working.
> I'm not sure HOW to educate udev on a device it doesn't just know -- and
> the few google queries on that I have tried so far have left me cross-
> eyed...
> 
>> I couldn't find any information on that
>> head set and linux except that head set is pretty expensive
>> 
> 
> Yes it's a very nice pair of headphones -- the sound quality is amazing,
> and the noise cancellation isn't bad either. And comfortable for spectacle
> wearers which not all makers can claim.
> 
>> https://lwn.net/Articles/531133/
>> http://jfcarter.net/~jimc/documents/blue-music-1504.html
>> 
> 
> Thanks for these -- I have only been able to review the second of these as
> the first is being blocked by my company firewall for some reason. I'll
> look at it when I get home tonight. Unfortunately, the second, like
> everything I've read on the internet about this problem in the last 24
> hours or so, skips over the problem I'm having -- from pairing (no
> problem) to connecting (no ERRORS, _probably_ no problem) to configuring /
> setting up Pulse (which I don't get to because my system thinks a keyboard
> just connected to it, not a pair of headphones).
> 
> Mark

>From my experience with bluetooth worst case would be that something is not
supported, but it is rear as BT is mostly standardized in terms of
communication.

So what I would do is to check the profiles reported via the hcitool
It would be most probably A2DP.
Next is to find out how to configure udev to map the profile to an output.
Jessie stock is 3.16 - it could be worth trying something newer - live
ubuntu pre release or so. Not only because of the kernel, but because udev
could be smarter.
I had to install blueman in order to easily configure bluetooth devices in
gnome on jessie recently. And I read that A2DP is located now in
gstreamer ... pulse something package (bluez5).

I think you misinterpret the messages - it is normal that you get keyboard
as input, because there are buttons you press to control the device.
The problem is you are missing the audio link. So there is a missing part -
not wrongly configured one. I would say this part is correctly configured,
but the audio link is not configured at all.

However it could be also that the expensive brand "Bang & Olufsen" provided
you with a CD with windows drivers for something not supported in linux -
this would be your worst case.

I hope this helps

Re: Using an SSD with strictatime mount option

[Andrew McGlashan]

On 4/03/2016 12:36 PM, Joel Roth wrote:
> On Tue, Mar 01, 2016 at 01:53:11PM +0100, Jochen Spieker wrote:
>> Joel Roth:
>>>
>>> mount -o remount,strictatime /
>>>
>>> Fixes it.
>>
>> Great! But wouldn't that still be a good idea to switch to Maildir? :)
> 
> Yes, it just takes a little work to change over :)

Perhaps this will help?

Migrate mbox data as follows:

mb2md -f /var/mail/mboxfile -d /tmp/imap-tmp

Adjust file ownership for the new files in /tmp/imap-tmp

Then change to user to use Maildir and move the files from the temp
location to the end desired Maildir folder for the user.

Cheers
A.

Avoiding to mount /run/lock

[Marlen Caemmerer]

Hello,

I want to avoid mounting /run/lock and I found there is an option in 
/etc/default/tmpfs.
I have set RAMLOCK=no but after reboot /run/lock is still a partition.
Do you have any hints on how to fix this?

Cheers
nosy

Nvidia Optimus in Debian

[Stephan Seitz]

Hi!

I have an Optimus notebook Acer Aspire V3 773G and using Debian testing.

lspci:
00:02.0 VGA compatible controller: Intel Corporation 4th Gen Core Processor 
Integrated Graphics Controller (rev 06)
01:00.0 3D controller: NVIDIA Corporation GK106M [GeForce GTX 760M] (rev a1)

The system is working, for now X is using the Intel chip.

While I have heard of Bumblebee I want to always use the Nvidia chip. It 
seems Ubuntu can do this (I think they are using something called nvidia 
prime) to switch between the chips.


Has anyone done this in Debian?

Many greetings,

Stephan

--
| Stephan Seitz  E-Mail: s...@fsing.rootsland.net |
| Public Keys: http://fsing.rootsland.net/~stse/keys.html |


smime.p7s
Description: S/MIME cryptographic signature

Re: which files took the space

[Gene Heskett]

On Friday 04 March 2016 03:39:03 jdd wrote:

> Le 04/03/2016 08:38, to...@tuxteam.de a écrit :
> > If memory serves, long, long time ago, "mount" (the system call)
> > refused to comply unless the directory was empty (kernel 2.6.mumble;
> > so long ago. Expect bit flips and glitches in my wetware, yadda).
>
> if it existed, it's much older than that, never seen it since 1997...

Me either, and I go back to '97 myself in linux usage.

> man mount:
>
> This  tells  the  kernel  to attach the filesystem found on device
> (which is of type type) at the directory dir.  The previous contents
> (if any) and owner and mode of dir become  invisible,  and  as  long 
> as this  filesystem  remains mounted, the pathname dir refers to the
> root of the filesystem on device."
>
>
> jdd


Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

systemd and php5-fpm

[basti]

Hello,

I have some errors with systemd and php5-fpm.
I want to create a dir /var/run/php5-fpm where I can place my sockets.

With the old sysv-init all is fine I just add a "mkdir" in it.

In systemd there is a bit different.
I try ExecStartPre=/bin/mkdir ... but always get errors that the socket
cant be created. (No such file and directory) /var/run/php5-fpm is not
created.

Now i try to use the ExecStartPre=/usr/lib/php5/php5-fpm-checkconf and
put my mkdir in there but it seems that this script isn't also working.

can someone please try it?
I use systemd 215-17+deb8u3

thanks a lot.

systemd + mkdir

[basti]

Hello,

I have some errors with systemd and php5-fpm.
I want to create a dir /var/run/php5-fpm where I can place my sockets.

With the old sysv-init all is fine I just add a "mkdir" in it.

In systemd there is a bit different.
I try ExecStartPre=/bin/mkdir ... but always get errors that the socket
cant be created. (No such file and directory) /var/run/php5-fpm is not
created.

Now i try to use the ExecStartPre=/usr/lib/php5/php5-fpm-checkconf and
put my mkdir in there but it seems that this script isn't also working.

can someone please try it?
I use systemd 215-17+deb8u3

thanks a lot.

Re: Nvidia Optimus in Debian

[Sven Arvidsson]

On Fri, 2016-03-04 at 14:26 +0100, Stephan Seitz wrote:
> Hi!
> 
> I have an Optimus notebook Acer Aspire V3 773G and using Debian
> testing.
> 
> lspci:
> 00:02.0 VGA compatible controller: Intel Corporation 4th Gen Core
> Processor Integrated Graphics Controller (rev 06)
> 01:00.0 3D controller: NVIDIA Corporation GK106M [GeForce GTX 760M]
> (rev a1)
> 
> The system is working, for now X is using the Intel chip.
> 
> While I have heard of Bumblebee I want to always use the Nvidia chip.
> It 
> seems Ubuntu can do this (I think they are using something called
> nvidia 
> prime) to switch between the chips.
> 
> Has anyone done this in Debian?

I think this is what you're looking for?
https://wiki.archlinux.org/index.php/PRIME#Reverse_PRIME

Instructions are for Arch. but I'm guessing it's the same in Debian.


-- 
Cheers,
Sven Arvidsson
http://www.whiz.se
PGP Key ID 6FAB5CD5



signature.asc
Description: This is a digitally signed message part

Re: Avoiding to mount /run/lock

[Sven Hartge]

Marlen Caemmerer  wrote:

> I want to avoid mounting /run/lock and I found there is an option in
> /etc/default/tmpfs.  I have set RAMLOCK=no but after reboot /run/lock
> is still a partition.  Do you have any hints on how to fix this?

What version of Debian are you using?

Grüße,
Sven.

-- 
Sigmentation fault. Core dumped.

What else must I do to avoid installing "suggests"? Was: How to get a "minimal font set"?

[Kynn Jones]

On Tue, Mar 1, 2016 at 11:03 AM, Stefan Monnier
 wrote:
>> Find a file somehow belonging to the font you are interested in.
>> Do "dpkg -L " on it.
>^^
>-S

Thank you, that is a useful tool.

When I ran `dpkg -S` over all the files reported by `fc-list`, and
then ran `aptitude why` on the reported packages, I discovered that
most of the superfluous fonts in my system came through packages that
were only suggested by other packages.

In fact, in many cases, the font-containing package is at the end of a
chain of suggests, where each link in the chain is a suggests of the
previous link.

IOW, I have a ton of junk in my system, not just superfluous fonts.

How can this be, given that I never use the `--install-suggests` flag
when I run `apt-get install`, and I don't have `Item:
APT::Install-Suggests` in my `/etc/apt/apt.conf`?

I must have some big gap or error in my understanding of Debian
package management and/or how `apt-get` works...

What else do I have to do to avoid having all these suggested packages
somehow get through to my system?

Thanks in advance!

kj

Re: systemd and php5-fpm

[Sven Hartge]

basti  wrote:

> I have some errors with systemd and php5-fpm.
> I want to create a dir /var/run/php5-fpm where I can place my sockets.

Why a separate directory? What do you want to achieve?

> In systemd there is a bit different.
> I try ExecStartPre=/bin/mkdir ... but always get errors that the socket
> cant be created. (No such file and directory) /var/run/php5-fpm is not
> created.

Any logs? What is the output of "journalctl -u php5-fpm.service"?

> Now i try to use the ExecStartPre=/usr/lib/php5/php5-fpm-checkconf and
> put my mkdir in there but it seems that this script isn't also working.

That is not wise, as that script gets overwritten on every package
update.


Grüße,
Sven.

-- 
Sigmentation fault. Core dumped.

Re: What else must I do to avoid installing "suggests"? Was: How to get a "minimal font set"?

[tomas]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Fri, Mar 04, 2016 at 09:59:54AM -0500, Kynn Jones wrote:
> On Tue, Mar 1, 2016 at 11:03 AM, Stefan Monnier
>  wrote:
> >> Find a file somehow belonging to the font you are interested in.
> >> Do "dpkg -L " on it.
> >^^
> >-S
> 
> Thank you, that is a useful tool.
> 
> When I ran `dpkg -S` over all the files reported by `fc-list`, and
> then ran `aptitude why` on the reported packages, I discovered that
> most of the superfluous fonts in my system came through packages that
> were only suggested by other packages.
> 
> In fact, in many cases, the font-containing package is at the end of a
> chain of suggests, where each link in the chain is a suggests of the
> previous link.
> 
> IOW, I have a ton of junk in my system, not just superfluous fonts.
> 
> How can this be, given that I never use the `--install-suggests` flag
> when I run `apt-get install`, and I don't have `Item:
> APT::Install-Suggests` in my `/etc/apt/apt.conf`?

Hm. Try "apt-config dump | grep -i suggests" -- it must be coming
from somewhere?

regards
- -- t
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)

iEYEARECAAYFAlbZnUEACgkQBcgs9XrR2kZ9PgCfQdgVIIr72sGKNlASAXzEb3c8
fIgAn1wiQm8MSOrlB5B7+VFDuPAqxqvE
=gHOF
-END PGP SIGNATURE-

Re: What else must I do to avoid installing "suggests"? Was: How to get a "minimal font set"?

[Linux-Fan]

[Fri, 4 Mar 2016 09:59:54 -0500] Kynn Jones  wrote:

[...]

> When I ran `dpkg -S` over all the files reported by `fc-list`, and
> then ran `aptitude why` on the reported packages, I discovered that
> most of the superfluous fonts in my system came through packages that
> were only suggested by other packages.
> 
> In fact, in many cases, the font-containing package is at the end of a
> chain of suggests, where each link in the chain is a suggests of the
> previous link.

[...]

I suggest you to try to uninstall some of the superflous packages.
Because whenever I found some chain of suggests, it always had one of
these reasons:

1. Aptitude did not report a stronger dependency. Often, a package is
   suggested by one but also recommended by another package. IIRC,
   when one tries to uninstall the package, Aptitude reports that some
   ``Recommends''-dependencies will be broken.
2. Sometimes, packages are pulled in as dependencies for one package
   and not removed when you uninstall the package because of
   ``Suggests''. This could probably be resolved with `deborphan`.

HTH
Linux-Fan

-- 
http://masysma.lima-city.de/


pgpjVVcqPySqe.pgp
Description: OpenPGP digital signature

Re: error "sh: 1: exec: ...: File exists" with dash

[tomas]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Thu, Mar 03, 2016 at 12:12:45PM +0100, Vincent Lefevre wrote:
> On 2016-03-03 11:06:55 +0100, to...@tuxteam.de wrote:
> > On Thu, Mar 03, 2016 at 10:18:55AM +0100, Vincent Lefevre wrote:
> > > I've got the following error in a script yesterday on a Debian 8
> > > machine:
> > > 
> > >   sh: 1: exec: /home/vlefevre/bin/mymaple: File exists
> > > 
> > > where sh is dash. The command that led to this error is equivalent to:
> > 
> > Just a cheap shot (I see you've dug up a couple of things which
> > might make my hunch obsolete). Do you have perhaps the noclobber
> > option set?

[...]

> As you can see the error message is different. It is at the shell
> level (before the "exec" command gets executed).

Yeah -- I said it was a cheap shot. Judging by the rest of the thread,
it's far more "interesting".

regards
- -- t
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)

iEYEARECAAYFAlbZoBgACgkQBcgs9XrR2kZVVgCfboXyUx7jd8CgwuOyrm6fOcUX
N20AnR4g8/IaouewpnU3XjE4cJ9S64lT
=DL/p
-END PGP SIGNATURE-

Re: Bluetooth headphones mistaken for a keyboard!?!

[Mark Fletcher]

On Fri, 2016-03-04 at 11:08 +0100, deloptes wrote:

> So what I would do is to check the profiles reported via the hcitool
> It would be most probably A2DP.

When I run hcitool info on the headphones I get:

$ hcitool info XX:XX:XX:XX:XX:XX
Requesting information ...
BD Address:  XX:XX:XX:XX:XX:XX
OUI Company: Bang & Olufsen A/S (00-09-A7)
Device Name: BeoPlay H8
LMP Version: 4.1 (0x7) LMP Subversion: 0x2812
Manufacturer: Cambridge Silicon Radio (10)
Features page 0: 0xff 0xff 0x8f 0xfe 0xdb 0xff 0x5b 0x87
<3-slot packets> <5-slot packets>   

 
 
   
   
   

  <3-slot EDR ACL> 
<5-slot EDR ACL>   
   
 <3-slot EDR eSCO>  
   

 
Features page 1: 0x03 0x00 0x00 0x00 0x00 0x00 0x00 0x00

(Address elided just in case I should care about that)

"sdptool records" gives me the following profiles supported (scraped
from the output, can post it all if it helps):

  "Handsfree" (0x111e)
Version: 0x0106
  "Headset" (0x1108)
Version: 0x0102
  "Advanced Audio" (0x110d)
Version: 0x0103
  "AV Remote" (0x110e)
Version: 0x0105

I suspect that "Advanced Audio" is the one I want...

> Next is to find out how to configure udev to map the profile to an output.

Uh... yeah. Unfortunately I am not sure what that means. Does that mean
create new udev rules? Know of any good resources I can use to learn how
to do that? I imagine I am gonna need that 0x110d code though, right?

> Jessie stock is 3.16 - it could be worth trying something newer - live
> ubuntu pre release or so. Not only because of the kernel, but because udev
> could be smarter.

The headphones are about a year old -- I mean this product has been on
the market for about a year, this particular pair is brand new -- does
that make any difference? I am willing to try a newer kernel if you
think it will really make a difference.

> I had to install blueman in order to easily configure bluetooth devices in
> gnome on jessie recently. And I read that A2DP is located now in
> gstreamer ... pulse something package (bluez5).
> 

Now that's interesting -- you are not the first person to recommend
Blueman since I started looking into this but I installed Blueman the
other day and don't see it telling me anything remotely useful that the
gnome bluetooth applet doesn't, except for signal strengths etc which
seem to be strong / optimal but aren't actually helping. Am I being
stupid?

I don't see A2DP in the list of profiles -- unless that is "Advanced
Audio"?

And gstreamer is installed. I see a lot of packages for gstreamer and
they aren't all installed, but the ones that look important to me are.
Specifically gstreamer-1.0-pulse is installed.

> I think you misinterpret the messages - it is normal that you get keyboard
> as input, because there are buttons you press to control the device.
> The problem is you are missing the audio link. So there is a missing part -
> not wrongly configured one. I would say this part is correctly configured,
> but the audio link is not configured at all.
> 

Ah, I see, that makes sense I guess.

> However it could be also that the expensive brand "Bang & Olufsen" provided
> you with a CD with windows drivers for something not supported in linux -
> this would be your worst case.
> 

It came with a small USB cable for charging, a regular audio cable for
situations where Bluetooth is not allowed, a manual in about 10
languages, and that's it. No Windows drivers. It's a pair of headphones
primarily designed for use generically with bluetooth enabled devices,
not primarily designed for use with a computer specifically. And I've
been using them connected to my iPhone while writing this, and I have to
say they are swet. So I am glad I bought 'em even if it takes me a
few days to get them working with my PC.

(Obviously I was not expecting them to work with my PC while they were
being used by my iPhone, I am not _that_ stupid :-) )

> I hope this helps
> 

Thanks for your ongoing help, I appreciate you taking the time.

Mark

Re: Bluetooth headphones mistaken for a keyboard!?!

[Sven Arvidsson]

On Fri, 2016-03-04 at 08:12 +0900, Mark Fletcher wrote:
> I took delivery of a very nice pair of Bang & Olufsen BeoPlay H8
> bluetooth headphones yesterday. I've verified they work properly by
> connecting them to my iPhone.
> 
> When I try to connect them to my PC, I am not having so much luck. PC
> is
> a 7-year-old self-built desktop box running Jessie with Gnome as the
> DE,
> with Intel Core i7-920 CPU and 24GB RAM. Bluetooth capability comes
> from
> a rather younger (bought about 6 months ago) bluetooth USB dongle
> which
> I routinely use to connect my iPhone and Android tablets to the
> computer
> to play audio from them through the computer's speakers. So I know it
> works fine too.
> 
> Turning on bluetooth from the gnome applet, and putting the
> headphones
> into pairing mode, the computer quickly finds the headphones. When I
> ask
> the computer to connect it quickly does so, apparently successfully,
> and
> I hear the tones in the headphones indicating they too know they are
> connected. But, pulseaudio doesn't recognise a new audio sink has
> been
> installed -- and I don't blame it, because of the below.
> 
> Looking in the Gnome Log Viewer in messages, at the moment of
> connecting
> the headphones, I see this:
> 
> Mar  4 08:07:17 kazuki gdm-Xorg-:0[1040]: (II) config/udev: Adding
> input
> device 00:09:A7:09:BE:22 (/dev/input/event17)
> Mar  4 08:07:17 kazuki gdm-Xorg-:0[1040]: (**) 00:09:A7:09:BE:22:
> Applying InputClass "evdev keyboard catchall"
> Mar  4 08:07:17 kazuki gdm-Xorg-:0[1040]: (II) Using input driver
> 'evdev' for '00:09:A7:09:BE:22'
> Mar  4 08:07:17 kazuki gdm-Xorg-:0[1040]: (**) 00:09:A7:09:BE:22:
> always
> reports core events
> Mar  4 08:07:17 kazuki gdm-Xorg-:0[1040]: (**) evdev:
> 00:09:A7:09:BE:22:
> Device: "/dev/input/event17"
> Mar  4 08:07:17 kazuki gdm-Xorg-:0[1040]: (--) evdev:
> 00:09:A7:09:BE:22:
> Vendor 0 Product 0
> Mar  4 08:07:17 kazuki gdm-Xorg-:0[1040]: (--) evdev:
> 00:09:A7:09:BE:22:
> Found keys
> Mar  4 08:07:17 kazuki kernel: [382097.651490] input:
> 00:09:A7:09:BE:22
> as /devices/virtual/input/input30
> Mar  4 08:07:17 kazuki gdm-Xorg-:0[1040]: (II) evdev:
> 00:09:A7:09:BE:22:
> Configuring as keyboard
> Mar  4 08:07:17 kazuki gdm-Xorg-:0[1040]: (**) Option "config_info"
> "udev:/sys/devices/virtual/input/input30/event17"
> Mar  4 08:07:17 kazuki gdm-Xorg-:0[1040]: (II) XINPUT: Adding
> extended
> input device "00:09:A7:09:BE:22" (type: KEYBOARD, id 13)
> Mar  4 08:07:17 kazuki gdm-Xorg-:0[1040]: (**) Option "xkb_rules"
> "evdev"
> Mar  4 08:07:17 kazuki gdm-Xorg-:0[1040]: (**) Option "xkb_model"
> "pc105"
> Mar  4 08:07:17 kazuki gdm-Xorg-:0[1040]: (**) Option "xkb_layout"
> "jp,us"
> Mar  4 08:07:17 kazuki gdm-Xorg-:0[1040]: (**) Option "xkb_variant"
> ","
> Mar  4 08:07:17 kazuki gdm-Xorg-:0[1040]: (**) Option "xkb_options"
> "grp:alt_shift_toggle,grp_led:scroll"
> 
> 
> That doesn't look healthy. First of all it is detecting an INPUT
> device,
> presumably the headphone controls for volume and so on, and no OUTPUT
> device. Second, it seems to think that what it has found is a
> keyboard.
> Pulseaudio can, it seems to me, be excused for thinking a keyboard
> has
> been added and not being accustomed to playing sound through a
> keyboard... :-)
> 
> I'm thinking this could be a udev problem of some kind -- perhaps I
> am
> missing appropriate udev rules? I am not sure how to diagnose the
> problem from here, far less fix it, so any advice would be
> appreciated.
> Google has, unusually, turned up zilch.
> 
> Thanks in advance
> 
> Mark
> 

Wouldn't headphones be pretty much like setting up a Bluetooth
speaker? 

The only thing I did to get it working was making sure that pulseaudio-
module-bluetooth was installed, everything else went without a hitch.

-- 
Cheers,
Sven Arvidsson
http://www.whiz.se
PGP Key ID 6FAB5CD5



signature.asc
Description: This is a digitally signed message part

Re: Avoiding to mount /run/lock

[Marlen Caemmerer]

On Fri, 4 Mar 2016, Sven Hartge wrote:



I want to avoid mounting /run/lock and I found there is an option in
/etc/default/tmpfs.  I have set RAMLOCK=no but after reboot /run/lock
is still a partition.  Do you have any hints on how to fix this?


What version of Debian are you using?


Debian Jessie

Re: What else must I do to avoid installing "suggests"? Was: How to get a "minimal font set"?

[Patrick Bartek]

On Fri, 04 Mar 2016, Kynn Jones wrote:

> On Tue, Mar 1, 2016 at 11:03 AM, Stefan Monnier
>  wrote:
> >> Find a file somehow belonging to the font you are interested in.
> >> Do "dpkg -L " on it.
> >^^
> >-S
> 
> Thank you, that is a useful tool.
> 
> When I ran `dpkg -S` over all the files reported by `fc-list`, and
> then ran `aptitude why` on the reported packages, I discovered that
> most of the superfluous fonts in my system came through packages that
> were only suggested by other packages.
> 
> In fact, in many cases, the font-containing package is at the end of a
> chain of suggests, where each link in the chain is a suggests of the
> previous link.
> 
> IOW, I have a ton of junk in my system, not just superfluous fonts.
> 
> How can this be, given that I never use the `--install-suggests` flag
> when I run `apt-get install`, and I don't have `Item:
> APT::Install-Suggests` in my `/etc/apt/apt.conf`?
> 
> I must have some big gap or error in my understanding of Debian
> package management and/or how `apt-get` works...
> 
> What else do I have to do to avoid having all these suggested packages
> somehow get through to my system?

My system doesn't install "Suggests" by default.  Never has.  Only true
dependencies.  How did you install your initial minimal system?

I started with just the Basic Install, terminal and networking from
a Wheezy 64-bit Netinstall CD (Jessie didn't exist), then added
piece by piece to end up with a lean X and window manager system.  Is
that what you basically did?

B

stretch boot problems

[Carlos Davila]

Hi,

I am having problems booting Stretch (Debian 9). During boot, I get a
series of messages:

(1 of 4) A start job is running for .
(2 of 4) A start job is running for.
.
(4 of 4) A start job is running for Network Manager (47 s / 1min 35s)

and after waiting for several minutes, I get:

[FAILED] Failed to start Login Service.
See 'systemctl status systemd-logind.service' for details
.
[FAILED] Failed to start Accounts Service.
See 'systemctl status accounts-daemon.service' for details
[FAILED] Failed to start Modem Manager.
See 'systemctl status ModemManager.service' for details
[FAILED] Failed to start Avahi.
See 'systemctl status
[FAILED] Failed to start Network Manager.
See 'systemctl status NetworkManager.service' for details

I have upgraded regularly with Synaptic. I often had to try several times
to get the os to boot, sometimes even having to run fsck from another
distro on my dual boot machine, and eventually it would bootuntil now.

Would appreciate any suggestions.
Carlos

Re: systemd + mkdir

[Sven Joachim]

On 2016-03-04 14:48 +0100, basti wrote:

> I have some errors with systemd and php5-fpm.
> I want to create a dir /var/run/php5-fpm where I can place my sockets.
>
> With the old sysv-init all is fine I just add a "mkdir" in it.
>
> In systemd there is a bit different.
> I try ExecStartPre=/bin/mkdir ... but always get errors that the socket
> cant be created. (No such file and directory) /var/run/php5-fpm is not
> created.
>
> Now i try to use the ExecStartPre=/usr/lib/php5/php5-fpm-checkconf and
> put my mkdir in there but it seems that this script isn't also working.

Use systemd-tmpfiles to create the directory.  To do that, place a file
in /etc/tmpfiles.d, say php5-fpm.conf, with the following content:

d /run/php5-fpm 0775 root root - -

See tmpfiles.d(5).

Cheers,
   Sven

Re: Nvidia Optimus in Debian

[Bret Busby]

On 04/03/2016, Stephan Seitz  wrote:
> Hi!
>
> I have an Optimus notebook Acer Aspire V3 773G and using Debian testing.
>
> lspci:
> 00:02.0 VGA compatible controller: Intel Corporation 4th Gen Core Processor
> Integrated Graphics Controller (rev 06)
> 01:00.0 3D controller: NVIDIA Corporation GK106M [GeForce GTX 760M] (rev
> a1)
>
> The system is working, for now X is using the Intel chip.
>
> While I have heard of Bumblebee I want to always use the Nvidia chip. It
> seems Ubuntu can do this (I think they are using something called nvidia
> prime) to switch between the chips.
>
> Has anyone done this in Debian?
>
> Many greetings,
>
>   Stephan
>

Hello, Stephan.

In searching for information about the V3-773G, I found very little
that is in english.

I have a V3-772G, and it took me two years to get it working with a
non-MS operating system.

My V3-772G shows on the case, as having an intel i7-4702MQ CPU (with
inbuilt graphics); I believe that it is of the Haskell architecture,
and, an nVidia GeForce GT750M graphics thing.

The only non-MS operating system that I could find, to drive the
hardware, was Ubuntu Linux, and I am now using UbuntuMATE 15.10 on it,
and, have been using UbuntuMATE, since some time last year.

Apart from having increased the RAM to 32GB, the computer hardware is
as it was  in its standard form.

There is, I think, much, in the list archives, about my adventure in
trying to get the computer to work, and, a particular problem, that
led to me finding about the drivers problems, was that, with Debian, I
could not get an external monitor to work.

The computer is a fine computer, but I think that you may have to
install Ubuntu Linux on your computer, to get it to work properly.

Whilst I was not trying Debian testing, in its current form, I had
tried to apply bumblebee to it, in my adventure with it, and, the
version of bumblebee that I tried (a later version may exist - I have
no idea), it did not work.

The only solution for my computer, was to install Ubuntu Linux.

And, it works. Mostly.

-- 

Bret Busby
Armadale
West Australia

..

"So once you do know what the question actually is,
 you'll know what the answer means."
- Deep Thought,
 Chapter 28 of Book 1 of
 "The Hitchhiker's Guide to the Galaxy:
 A Trilogy In Four Parts",
 written by Douglas Adams,
 published by Pan Books, 1992

Automounting query

[Ron]

Would a kind soul remind me the invocation needed to have removable drives 
automount to /media/label instead of /media/user/label ?

TIA
 
Cheers,
 
Ron.
-- 
 The first draft of anything is shit.
  -- Ernest Hemingway

   -- http://www.olgiati-in-paraguay.org --
 

Converting to Maildir (was: Re: Using an SSD with strictatime mount option)

[Joel Roth]

Hi Andrew,

Andrew McGlashan wrote:
> 
> On 4/03/2016 12:36 PM, Joel Roth wrote:
> > On Tue, Mar 01, 2016 at 01:53:11PM +0100, Jochen Spieker wrote:
> >> Joel Roth:
> >>>
> >>> mount -o remount,strictatime /
> >>>
> >>> Fixes it.
> >>
> >> Great! But wouldn't that still be a good idea to switch to Maildir? :)
> > 
> > Yes, it just takes a little work to change over :)
> 
> Perhaps this will help?
> 
> Migrate mbox data as follows:
> 
> mb2md -f /var/mail/mboxfile -d /tmp/imap-tmp

Thanks for offering this. For some reason, I ended up
using this:

mb2md -s Mail -d Maildir

What is interesting, is that mb2md prepends a '.' to all of
the existing mbox folder names it converts. Also, dots in
existing folder names are converted to underscore.

Is there something magical about the dot in the Maildir format?

Otherwise, I will just remove the dots, since my
mail filter saves to foldernames without the dots.

Also is there any special meaning for the new, cur, other,
queue and tmp folders that mb2md created?

Regards,

Joel

 
> Adjust file ownership for the new files in /tmp/imap-tmp
> 
> Then change to user to use Maildir and move the files from the temp
> location to the end desired Maildir folder for the user.
> 
> Cheers
> A.
> 

-- 
Joel Roth
  

Re: NVIDIA GeForce problem

[Terence]

Thanks, Bret,

I was thinking of switching to Intel when I next upgrade, so this will
either help me to decide on change, or confirm my  continuation with AMD.

"Tidy as you go!" The Starship Titanic, Douglas Adams...

Terence

On 3 March 2016 at 05:24, Bret Busby  wrote:

> On 03/03/2016, Levi S. Darrell  wrote:
> > Hello,
> >
> > I am trying to get my NVIDIA video hardware to work properly with X. I am
> > running Debian Testing, which I just installed a few days ago.
>
> 
>
> Hello.
>
> I have, some time ago, posted a series of messages, relating to a
> problem where it took me two years to get a portable computer working
> with a non-MS operating system.
>
> As it was, I found only two non-MS operating systems to then have the
> drivers for the intel i7 CPU; DragonflyBSD and Ubuntu Linux.
>
> Of those, only Ubuntu had a driver that worked the nVIDIA graphics thing.
>
> So, I am now using UbuntuMATE 15.10 on the system.
>
> My suggestion to you, is to try (possibly first in live mode, rather
> than installed) running Ubuntu on the system, and, finding how it
> works, with the system hardware. I would suggest downloading a
> UbuntuMATE 15.10 iso, and trying that, but, whilst my preference is
> for the GNOME2 styled interface that I get with UbuntuMATE, others may
> prefer different interfaces.
>
> --
>
> Bret Busby
> Armadale
> West Australia
>
> ..
>
> "So once you do know what the question actually is,
>  you'll know what the answer means."
> - Deep Thought,
>  Chapter 28 of Book 1 of
>  "The Hitchhiker's Guide to the Galaxy:
>  A Trilogy In Four Parts",
>  written by Douglas Adams,
>  published by Pan Books, 1992
>
> 
>
>

Re: Converting to Maildir (was: Re: Using an SSD with strictatime mount option)

[Joe]

On Fri, 4 Mar 2016 09:12:47 -1000
Joel Roth  wrote:


> 
> Also is there any special meaning for the new, cur, other,
> queue and tmp folders that mb2md created?
> 

Not sure about the others, cur, new and tmp are created in all
subdirectories by Courier IMAP, so presumably are relevant to it. New
contains mail which has not yet been touched, even enumerating the mail
moves it into cur(rent). I've never seen anything in a tmp yet.

Not that I spend much of my life snooping on my mail server's habits...

More Than You Wanted To Know (tm):
https://en.wikipedia.org/wiki/Maildir

-- 
Joe

Re: Avoiding to mount /run/lock

[Sven Hartge]

Marlen Caemmerer  wrote:

> I want to avoid mounting /run/lock and I found there is an option in
> /etc/default/tmpfs.  I have set RAMLOCK=no but after reboot /run/lock
> is still a partition.  Do you have any hints on how to fix this?

Why do you want to disable /run/lock? What do you want to accomplish by
doing so? It is a 5MB tmpfs, what could be wrong with it?

Grüße,
Sven.

-- 
Sigmentation fault. Core dumped.

Re: Bluetooth headphones mistaken for a keyboard!?!

[deloptes]

Mark Fletcher wrote:

> On Fri, 2016-03-04 at 11:08 +0100, deloptes wrote:
> 
>> So what I would do is to check the profiles reported via the hcitool
>> It would be most probably A2DP.
> 
> When I run hcitool info on the headphones I get:
> 
> $ hcitool info XX:XX:XX:XX:XX:XX
> Requesting information ...
> BD Address:  XX:XX:XX:XX:XX:XX
> OUI Company: Bang & Olufsen A/S (00-09-A7)
> Device Name: BeoPlay H8
> LMP Version: 4.1 (0x7) LMP Subversion: 0x2812
> Manufacturer: Cambridge Silicon Radio (10)
> Features page 0: 0xff 0xff 0x8f 0xfe 0xdb 0xff 0x5b 0x87
> <3-slot packets> <5-slot packets>  
>
> 
> 
>   
>   
>   
>
>   <3-slot EDR ACL>
> <5-slot EDR ACL>  
>   
>  <3-slot EDR eSCO> 
>   
>
> 
> Features page 1: 0x03 0x00 0x00 0x00 0x00 0x00 0x00 0x00

I can not check at the moment. I have not tried to configure headset. I did
a lot with mobile phones and recall there was a2dp. I'll try to check later
on that.

> 
> (Address elided just in case I should care about that)
> 
> "sdptool records" gives me the following profiles supported (scraped
> from the output, can post it all if it helps):
> 
>   "Handsfree" (0x111e)
> Version: 0x0106
>   "Headset" (0x1108)
> Version: 0x0102
>   "Advanced Audio" (0x110d)
> Version: 0x0103
>   "AV Remote" (0x110e)
> Version: 0x0105
> 
> I suspect that "Advanced Audio" is the one I want...

Yes it looks like 
https://en.wikipedia.org/wiki/List_of_Bluetooth_profiles#Advanced_Audio_Distribution_Profile_.28A2DP.29

> 
>> Next is to find out how to configure udev to map the profile to an
>> output.
> 
> Uh... yeah. Unfortunately I am not sure what that means. Does that mean
> create new udev rules? Know of any good resources I can use to learn how
> to do that? I imagine I am gonna need that 0x110d code though, right?

https://wiki.archlinux.org/index.php/Bluetooth_headset

yes, perhaps check this article here (section: Legacy documentation: ALSA,
bluez5 and PulseAudio method) - seems a big discussion and article had some
critics, but there are few ideas on how to setup udev rules
example
$ bluetoothctl 
produces output with UUID

> 
>> Jessie stock is 3.16 - it could be worth trying something newer - live
>> ubuntu pre release or so. Not only because of the kernel, but because
>> udev could be smarter.
> 
> The headphones are about a year old -- I mean this product has been on
> the market for about a year, this particular pair is brand new -- does
> that make any difference? I am willing to try a newer kernel if you
> think it will really make a difference.

No difference - question is the level of support by the OS. I must admit
that before bluez5 all this BT related stuff was a nightmare.
The good thing is that BT devices are more or less standard, so the question
is how good is linux for the B&O product.

> 
>> I had to install blueman in order to easily configure bluetooth devices
>> in gnome on jessie recently. And I read that A2DP is located now in
>> gstreamer ... pulse something package (bluez5).
>> 
> 
> Now that's interesting -- you are not the first person to recommend
> Blueman since I started looking into this but I installed Blueman the
> other day and don't see it telling me anything remotely useful that the
> gnome bluetooth applet doesn't, except for signal strengths etc which
> seem to be strong / optimal but aren't actually helping. Am I being
> stupid?
> 
> I don't see A2DP in the list of profiles -- unless that is "Advanced
> Audio"?

I think blueman helped me somehow get the proper profile loaded for the
phones I tested with

> 
> And gstreamer is installed. I see a lot of packages for gstreamer and
> they aren't all installed, but the ones that look important to me are.
> Specifically gstreamer-1.0-pulse is installed.

what about pulse-bluetooth ?

> 
>> I think you misinterpret the messages - it is normal that you get
>> keyboard as input, because there are buttons you press to control the
>> device. The problem is you are missing the audio link. So there is a
>> missing part - not wrongly configured one. I would say this part is
>> correctly configured, but the audio link is not configured at all.
>> 
> 
> Ah, I see, that makes sense I guess.
> 
>> However it could be also that the expensive brand "Bang & Olufsen"
>> provided you with a CD with windows drivers for something not supported
>> in linux - this would be your worst case.
>> 
> 
> It came with a small USB cable for charging, a regular audio cable for
> situations where Bluetooth is not allowed, a manual in about 10
> languages, and that's it. No Windows drivers. It's a pair of headphones
> primarily designed for use generically with bluetooth enabled devices,
> not primarily designed for use with a computer specifically. And I've
> been using them connected to my iPhone while writing this, and I have to
> say they are swet. So I am glad I bought 'em even if it takes me a
> few days to get them working with my PC.

Thats good - you could use a windows 

Re: Avoiding to mount /run/lock

[Marlen Caemmerer]

On Fri, 4 Mar 2016, Sven Hartge wrote:



Why do you want to disable /run/lock? What do you want to accomplish by
doing so? It is a 5MB tmpfs, what could be wrong with it?



I run a samba server that is heavily used. It currently uses 46MB in /run/lock. 
When the partition is full all samba service dies.
I dont see any real advantage in having an extra tmpfs partition - if it is not 
a separate mount point I never run into any issues.
Currently I unmount the partition after boot but  I guess thats not an optimal 
solution.

Grüße
nosy

Re: Avoiding to mount /run/lock

[Sven Hartge]

Marlen Caemmerer  wrote:
> On Fri, 4 Mar 2016, Sven Hartge wrote:

>> Why do you want to disable /run/lock? What do you want to accomplish by
>> doing so? It is a 5MB tmpfs, what could be wrong with it?

> I run a samba server that is heavily used. It currently uses 46MB in
> /run/lock. When the partition is full all samba service dies.

Aha. That is a valid reason. But I'd rather increase the size instead of
just eliminating it.

But this could also be a bug with Samba, for example not deleting old
lock files. Have you checked what causes this?

Grüße,
Sven.

-- 
Sigmentation fault. Core dumped.

Re: which files took the space

[Jörg-Volker Peetz]

The list seems nowadays to have a very short memory ;-)
A similar problem was discussed last time hardly a month ago:
https://lists.debian.org/debian-user/2016/02/threads.html#00267

So in order to avoid the pit you fall into you could've used
a command like

  du /home -hx --max-depth=1

Regards,
jvp.

Re: Avoiding to mount /run/lock

[tomas]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Fri, Mar 04, 2016 at 10:15:40PM +0100, Marlen Caemmerer wrote:
> On Fri, 4 Mar 2016, Sven Hartge wrote:
> 
> >
> >Why do you want to disable /run/lock? What do you want to accomplish by
> >doing so? It is a 5MB tmpfs, what could be wrong with it?
> >
> 
> I run a samba server that is heavily used. It currently uses 46MB in 
> /run/lock. When the partition is full all samba service dies.
> I dont see any real advantage in having an extra tmpfs partition - if it is 
> not a separate mount point I never run into any issues.
> Currently I unmount the partition after boot but  I guess thats not an 
> optimal solution.

Sven has a point: perhaps increasing the size or finding out what's up
with samba. That said, if you are using sysvinit, the script responsible
for mounting /run/lock is /etc/init.d/mountkernfs.sh; if you're on
systemd I'm out of my depth, sorry.

regards
- -- t
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)

iEYEARECAAYFAlbZ9NUACgkQBcgs9XrR2kaqmgCdGBcZIwL5/6SxnnyDgK4MVPyt
X2YAn2YHUD7AP6vJwXsYd24OcaUcb7I6
=Rtsi
-END PGP SIGNATURE-

Re: Avoiding to mount /run/lock

[Sven Hartge]

to...@tuxteam.de wrote:
> On Fri, Mar 04, 2016 at 10:15:40PM +0100, Marlen Caemmerer wrote:
>> On Fri, 4 Mar 2016, Sven Hartge wrote:

>>> Why do you want to disable /run/lock? What do you want to accomplish by
>>> doing so? It is a 5MB tmpfs, what could be wrong with it?


>> I run a samba server that is heavily used. It currently uses 46MB in
>> /run/lock. When the partition is full all samba service dies.  I dont
>> see any real advantage in having an extra tmpfs partition - if it is
>> not a separate mount point I never run into any issues.  Currently I
>> unmount the partition after boot but  I guess thats not an optimal
>> solution.

> Sven has a point: perhaps increasing the size or finding out what's up
> with samba. That said, if you are using sysvinit, the script responsible
> for mounting /run/lock is /etc/init.d/mountkernfs.sh; if you're on
> systemd I'm out of my depth, sorry.

I see reports from Samba users dating back quite a while about big
locking.tdb files in /run/lock or /var/lock, for example
https://lists.samba.org/archive/samba/2013-August/175075.html

The Sernet guys, whor provided the Samba 4 packages back then changed
the location for the locking.tdb to /var/cache/samba to avoid that
problem: 
https://lists.samba.org/archive/samba/2013-August/175356.html

Maybe Debian should to the same?

Looking at smb.conf, you can also do this, just change "lock directory"
to a sane value. The default is "lock directory = /var/run/samba".

Grüße,
Sven

-- 
Sigmentation fault. Core dumped.

Re: NVIDIA GeForce problem

[Sven Arvidsson]

On Wed, 2016-03-02 at 16:33 -0700, Levi S. Darrell wrote:
> Hello,
> 
> I am trying to get my NVIDIA video hardware to work properly with X.
> I am running Debian Testing, which I just installed a few days ago. I
> installed nvidia-detect using apt. I then installed Bumblebee using
> the instructions at 
> https://wiki.debian.org/Bumblebee#Installation. I added my login user
> to the bumblebee grouop. However, attempting '$ optirun xorg -info'
> gives the following error:
> 
> [15211.548537] nouveau :01:00.0: priv: HUB0: 10ecc0 
> (1d40822c)
> [15211.140632] [ERROR]Cannot access secondary GPU - error: [XORG]
> (EE) No devices detected.
> 
[..]
> Section "Device"
> Identifier  "DiscreteNvidia"
> Driver  "nvidia"
> VendorName  "NVIDIA Corporation"
> BusID "PCI:01:00.0"

What did you do, or what are you trying to do? You can't match nvidia
with nouveau.

-- 
Cheers,
Sven Arvidsson
http://www.whiz.se
PGP Key ID 6FAB5CD5



signature.asc
Description: This is a digitally signed message part

Jenkins on Debian 8 jessie

[Nicholas Geovanis]

Is anyone running jenkins under Debian 8 jessie? If so, which JDK are you
using? Did you need to update /etc/default/jenkins to point at the right
JDK location, or it did fine "out of the box"? Many thanks.Nick Geo

Re: Avoiding to mount /run/lock

[Marlen Caemmerer]

On Fri, 4 Mar 2016, Sven Hartge wrote:




Sven has a point: perhaps increasing the size or finding out what's up
with samba. That said, if you are using sysvinit, the script responsible
for mounting /run/lock is /etc/init.d/mountkernfs.sh; if you're on
systemd I'm out of my depth, sorry.


Yes its systemd now...



I see reports from Samba users dating back quite a while about big
locking.tdb files in /run/lock or /var/lock, for example
https://lists.samba.org/archive/samba/2013-August/175075.html



The biggest file is a file called unexpected.tdb.
Samba documentation says it a file for "Unexpected packet queue needed to support 
windows clients that respond on a different port that the originating reques."



The Sernet guys, whor provided the Samba 4 packages back then changed
the location for the locking.tdb to /var/cache/samba to avoid that
problem:
https://lists.samba.org/archive/samba/2013-August/175356.html

Maybe Debian should to the same?

Looking at smb.conf, you can also do this, just change "lock directory"
to a sane value. The default is "lock directory = /var/run/samba".


Sounds like a reasonable workaround. What would be a proper path? 
/var/lib/samba?

Still I guess its not really an option to not know where those tempdirs can 
really be configured.

Grüße
nosy

Re: Avoiding to mount /run/lock

[Sven Hartge]

Marlen Caemmerer  wrote:
> On Fri, 4 Mar 2016, Sven Hartge wrote:

>> Looking at smb.conf, you can also do this, just change "lock directory"
>> to a sane value. The default is "lock directory = /var/run/samba".

> Sounds like a reasonable workaround. What would be a proper path? 
> /var/lib/samba?

Either that or /var/cache/samba, because those locks are non-persistant
data.

> Still I guess its not really an option to not know where those
> tempdirs can really be configured.

There is documentation, I have seen it. But I can't find or remember it
right now. If you are able to conjure Michael Bible to notice this
thread, I am sure he will be able to answer on how to change the size of
/run/lock.

S°

-- 
Sigmentation fault. Core dumped.

Re: What else must I do to avoid installing "suggests"? Was: How to get a "minimal font set"?

[David Christensen]

On 03/04/2016 06:59 AM, Kynn Jones wrote:

IOW, I have a ton of junk in my system, not just superfluous fonts.

How can this be, given that I never use the `--install-suggests` flag
when I run `apt-get install`, and I don't have `Item:
APT::Install-Suggests` in my `/etc/apt/apt.conf`?

I must have some big gap or error in my understanding of Debian
package management and/or how `apt-get` works...

What else do I have to do to avoid having all these suggested packages
somehow get through to my system?


I use apt-get.  About a year ago I discovered that installing a certain 
package pulled in 100's of MB of stuff that I didn't want or need.  My 
solution was to write a Perl wrapper script that invokes 'apt-get 
install' with the '--no-install-recommends' option.



HTH,

David

Re: What else must I do to avoid installing "suggests"? Was: How to get a "minimal font set"?

[Joel Roth]

On Fri, Mar 04, 2016 at 04:59:03PM -0800, David Christensen wrote:
> On 03/04/2016 06:59 AM, Kynn Jones wrote:
> >IOW, I have a ton of junk in my system, not just superfluous fonts.
> >
> >How can this be, given that I never use the `--install-suggests` flag
> >when I run `apt-get install`, and I don't have `Item:
> >APT::Install-Suggests` in my `/etc/apt/apt.conf`?
> >
> >I must have some big gap or error in my understanding of Debian
> >package management and/or how `apt-get` works...
> >
> >What else do I have to do to avoid having all these suggested packages
> >somehow get through to my system?
> 
> I use apt-get.  About a year ago I discovered that installing a certain
> package pulled in 100's of MB of stuff that I didn't want or need.  My
> solution was to write a Perl wrapper script that invokes 'apt-get install'
> with the '--no-install-recommends' option.

Hi,

cat /etc/apt/apt.conf.d/40nosuggestrecommends 
Apt::Install-Suggests false;
Apt::Install-Recommends false;
> 
> HTH,
> 
> David
> 

-- 
Joel Roth
  

Re: Bluetooth headphones mistaken for a keyboard!?!

[Mark Fletcher]

On Sat, Mar 5, 2016, 5:26 AM deloptes  wrote:

>
> I think you could fix this with proper udev setup though, but still it
> depends and I can not make promises - it could be you need some extra
> software
>
>


SOLVED -- This link was helpful -->
https://forums.gentoo.org/viewtopic-p-7872608.html?sid=1ebad1b4456ab28b24e0eb0228f7bcbf

The issue was that I had two instances of PulseAudio running. The first was
running as Debian-gdm and the second as my local user. Following the
instructions on the above link, I went to Debian-gdm's home directory and
created a pulse directory in .config. In there I created client.conf and
populated it according to the above link. Reboot the PC, reconnect the
headphones, and VOILA. Working perfectly. (Reboot probably wasn't strictly
necessary, a restart of GDM would probably have sufficed)

Now I don't know why there were two instances of Pulse Audio running, but
I'm loath to believe that it is purely past stupidity on my part, since
someone else had the same problem as evidenced by the above link. So
there's presumably some legitimate reason why this situation would apply,
and I don't know what I will have broken by making this change, but so far
nothing I can detect. Anyone with opinions on why stock Jessie would be set
up this way out of the box, please comment.

I want to particularly thank deloptes for your time and help, you got me
searching along the right lines to find the problem.

Mark

Re: What else must I do to avoid installing "suggests"? Was: How to get a "minimal font set"?

[Bob Bernstein]

On Fri, 4 Mar 2016, Joel Roth wrote:


cat /etc/apt/apt.conf.d/40nosuggestrecommends
Apt::Install-Suggests false;
Apt::Install-Recommends false;


Nice. Done. Thanks.

I had reason a couple of nights ago to consult 'man apt-get' 
because it seemed that lots of "suggested" pkgs were being 
installed. The two relevant options I found on the man page:


--no-install-recommends
and
--install-suggests

lead me to believe that the desired default behavior is to go 
ahead and install 'recommends' but not 'suggests'.


When as an experiment I added "--no-install-suggests" (not in 
the man page afaik) to an apt-get commandline, I was surprised 
that no error message leaped up i.e. I did not see "no such 
option."


I'm sorry I don't have better empirical data to offer as to what 
packages were involved, and how many, but my vote for the 
question "Has weirdness crept into apt-get? is "Yes."



All best,

--
I am animated by distrust of all high guesses, and by sympathy
with the old prejudices and workaday opinions of mankind: they
are ill expressed, but they are well grounded.

  George Santayana

Re: What else must I do to avoid installing "suggests"? Was: How to get a "minimal font set"?

[David Christensen]

On 03/04/2016 06:14 PM, Joel Roth wrote:

cat /etc/apt/apt.conf.d/40nosuggestrecommends
Apt::Install-Suggests false;
Apt::Install-Recommends false;


Thanks for the tip -- I was wondering if there was a configuration 
setting somewhere:


$ man apt.conf

$ zcat /usr/share/doc/apt/examples/configure-index.gz | less


David

Re: What else must I do to avoid installing "suggests"? Was: How to get a "minimal font set"?

[David Wright]

On Fri 04 Mar 2016 at 23:15:28 (-0500), Bob Bernstein wrote:
> On Fri, 4 Mar 2016, Joel Roth wrote:
> 
> >cat /etc/apt/apt.conf.d/40nosuggestrecommends
> >Apt::Install-Suggests false;
> >Apt::Install-Recommends false;
> 
> Nice. Done. Thanks.
> 
> I had reason a couple of nights ago to consult 'man apt-get' because
> it seemed that lots of "suggested" pkgs were being installed. The
> two relevant options I found on the man page:
> 
> --no-install-recommends
> and
> --install-suggests
> 
> lead me to believe that the desired default behavior is to go ahead
> and install 'recommends' but not 'suggests'.

That's logical. The options listed all make it do something different
from the normal behaviour with one exception: --list-cleanup is the
default. I don't know why they picked on that one.

> When as an experiment I added "--no-install-suggests" (not in the
> man page afaik) to an apt-get commandline, I was surprised that no
> error message leaped up i.e. I did not see "no such option."

My man page starts thus:

OPTIONS
   All command line options may be set using the configuration
   file, the descriptions indicate the configuration option to
   set. For boolean options you can override the config file by
   using something like -f-,--no-f, -f=no or several other   <--- 
second in the list is --no-f
   variations.

   --no-install-recommends
   Do not consider recommended packages as a dependency for
   installing. Configuration Item: APT::Install-Recommends.

   --install-suggests
   Consider suggested packages as a dependency for
   installing. Configuration Item: APT::Install-Suggests.

Cheers,
David.

Re: Automounting query

[Gary Dale]

On 04/03/16 12:42 PM, Renaud (Ron) OLGIATI wrote:

Would a kind soul remind me the invocation needed to have removable drives 
automount to /media/label instead of /media/user/label ?

TIA
  
Cheers,
  
Ron.


I believe that if you have the drive identified in /etc/fstab to mount 
where you want it, it will override the default location.


For a more general solution, look at 
http://www.reactivated.net/writing_udev_rules.html.

Re: What else must I do to avoid installing "suggests"? Was: How to get a "minimal font set"?

[David Wright]

On Fri 04 Mar 2016 at 20:41:25 (-0800), David Christensen wrote:
> On 03/04/2016 06:14 PM, Joel Roth wrote:
> >cat /etc/apt/apt.conf.d/40nosuggestrecommends
> >Apt::Install-Suggests false;
> >Apt::Install-Recommends false;
> 
> Thanks for the tip -- I was wondering if there was a configuration
> setting somewhere:
> 
> $ man apt.conf
> 
> $ zcat /usr/share/doc/apt/examples/configure-index.gz | less

I just write:less /usr/share/doc/apt/examples/configure-index.gz
because I have:

  eval $(lesspipe)

in my ~/.bashrc

Cheers,
David.

trying to optimize find's searches ...

[Albretch Mueller]

find works significantly faster when you bag various search patterns á la:

-iname '*.ko' -or -iname '*.bin' -or -iname '*.txt'

but, how could you do that using a batch strategy and setting a
variable with many "-[i]name ... -or -[i]name ..." cases?

line-by-line strategies are too slow

lbrtchx

$ _ODIR="<...>"

$ _SDIR="<...>"
$
$ _OFL="$_ODIR/$(basename "$_SDIR")_$(date +%Y%m%d%H%M%S)_fs_ko_bin_txt.log"
$ (time sudo find "${_SDIR}" -type f -iname '*.ko' -or -iname '*.bin'
-or -iname '*.txt')  > "${_OFL}" 2>&1
$
$ wc -l "${_OFL}"
1216 /media/sda6/tmp/sda1_20160304192857_fs_ko_bin_txt.log
$ tail -n 4 "${_OFL}"

real0m2.274s
user0m1.457s
sys 0m0.773s

$ _OFL="$_ODIR/$(basename "$_SDIR")_$(date +%Y%m%d%H%M%S)_fs_ko.log"
$ (time sudo find "${_SDIR}" -type f -iname '*.ko')  > "${_OFL}" 2>&1
$
$ wc -l "${_OFL}"
4 /media/sda6/tmp/sda1_20160304193029_fs_ko.log
$ tail -n 4 "${_OFL}"

real0m1.526s
user0m0.697s
sys 0m0.807s

$ _SDIR="/media/sda1"
$
$ _OFL="$_ODIR/$(basename "$_SDIR")_$(date +%Y%m%d%H%M%S)_fs_txt.log"
$ (time sudo find "${_SDIR}" -type f -iname '*.txt')  > "${_OFL}" 2>&1
$
$ wc -l "${_OFL}"
1213 /media/sda6/tmp/sda1_20160304193057_fs_txt.log
$ tail -n 4 "${_OFL}"

real0m1.641s
user0m0.757s
sys 0m0.797s

$ _OFL="$_ODIR/$(basename "$_SDIR")_$(date +%Y%m%d%H%M%S)_fs_bin.log"
$ (time sudo find "${_SDIR}" -type f -iname '*.bin')  > "${_OFL}" 2>&1
$
$ wc -l "${_OFL}"
7 /media/sda6/tmp/sda1_20160304193127_fs_bin.log
$ tail -n 4 "${_OFL}"

real0m1.535s
user0m0.750s
sys 0m0.763s

Re: What else must I do to avoid installing "suggests"? Was: How to get a "minimal font set"?

[David Christensen]

On 03/04/2016 09:10 PM, David Wright wrote:

I just write:less /usr/share/doc/apt/examples/configure-index.gz
because I have:

   eval $(lesspipe)

in my ~/.bashrc


Okay:

$ man lessfile

$ man lesspipe


David

Re: trying to optimize find's searches ...

[tomas]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Sat, Mar 05, 2016 at 05:23:55AM +, Albretch Mueller wrote:
> find works significantly faster when you bag various search patterns á la:
> 
> -iname '*.ko' -or -iname '*.bin' -or -iname '*.txt'
> 
> but, how could you do that using a batch strategy and setting a
> variable with many "-[i]name ... -or -[i]name ..." cases?
> 
> line-by-line strategies are too slow

[examples snipped: basically one complex find faster than three]

What exactly is what you want to do? I'd say "use the first strategy,
it's faster" -- but I have the feeling I haven't understood you
correctly.

Regards
- -- tomás
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)

iEYEARECAAYFAlbadrYACgkQBcgs9XrR2kadLACdE847fsxhCmh05yWDYZj3ZOqW
KTkAnA6m27Fa6XsuPZL2GAlVuhyJIONY
=uYHx
-END PGP SIGNATURE-