Sendmail's host status cacheing (HostStatusDirectory)
Hi, Is it just me and my Sendmail (the one in stable), or does it seem to cache rather aggressively the host status of other mail relays? We've got an intermediate SMTP relay that relays down to clients, and a client's SMTP server blew up and was repaired at 2pm, and at 5pm, our Sendmail was still considering it to be down, purely because of what was in the HostStatusDirectory. I notice that the Sendmail default is to have this switched off, however the Debian default is to turn it on. Would the Sendmail maintainer like to comment on the efficacy of this option, and indeed the rationale behind enabling at the default? Andrew -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Using a DPT SmartRAID V controller under Linux
Hi, Has anyone out there tried to get a DPT SmartRAID V card working under Debian (slink)? The DPT site has installation boot disks for Red Hat, and from what I can determine, kernel source modifications for 2.2.5, but I'm having dramas getting it working. The documentation that DPT provides is pretty non-existent. Has anyone been there, done that? Andrew Andrew Pollock Technical Director [EMAIL PROTECTED] http://staff.bit.net.au/apollock Brisbane Internet Technology Pty Ltd.
Small faux pas with LILO
Hi, I have a PC with two hard drives, /dev/hda and /dev/hdb I installed Linux on /dev/hdb and inadvertantly set boot=/dev/hda1 instead of /dev/hda in /etc/lilo.conf and now I'm having trouble booting that partition (a Windows 95 partition). Have I done inrepairable damage to that partition or can I recover from this boo boo? Andrew Andrew Pollock Technical Director [EMAIL PROTECTED] http://staff.bit.net.au/apollock Brisbane Internet Technology Pty Ltd.
RE: Small faux pas with LILO
> I've made this mistake too! > > LILO should have backed up the boot sector of /dev/hda1 to > /boot/boot.0301. So, this incantation of dd should get things back the > way they were: > dd if=/boot/boot.0301 of=/dev/hda1 bs=512 count=1 That is so cool. It fixed it good and proper. Thankyou so much, I'll have to file this one away for future reference. Andrew
Setuid Perl script works on one slink box, doesn't work on the other
Hi, We've got two relatively new slink boxes, both running 2.2.12, and I've got a setuid Perl script that doesn't work on one, but does on the other. Permissions are the same, everything (that I can think of) is the same... It's Perl 5.004_04. On the box that won't run the script, it barfs: Can't do setuid As soon as you try to run the script. Anyone got any bright ideas? Andrew Andrew Pollock Technical Director [EMAIL PROTECTED] http://staff.bit.net.au/apollock Brisbane Internet Technology Pty Ltd.
Checking quotas and the bootup process
Hi, Currently /etc/init.d/quota runs at position 20 in runlevel 2 (default installation) and this tends to take a very long time to run at bootup on one particular server, and because a lot of the other services tend to startup afterwards, either for alphabetic reasons within position 20, or they are in a higher position, this particular Linux box is pretty much inaccessible and useless whilst checking it's quotas. Is there any harm in making the quota checking process further down the line, after the likes of sendmail and apache have started, or does it run at this position for a particular reason (i.e. it's better not to be modifying the filesystem whilst it's running?). Andrew -------- Andrew Pollock Technical Manager [EMAIL PROTECTED] http://www.bit.net.au/ Phone: +61 7 3252 1600 Fax: +61 7 3252 1900 Brisbane Internet Technology, an Asia Online Company
Last attempt, Debian 2.2 with X, on a UltraSPARC 5, with a Type 6 keyboard
Hi, If I can't get this resolved soon I'm going to be forced to hose Debian and go back to Solaris (weep). I've installed Debian on my Ultra 5. Not a problem in the world. Installed X. Started X. The keyboard mapping is all messed up (mainly the bottom line). I've tried playing with xkeycaps. It only knows about Sun Type 5 keyboards, not Type 6 keyboards, which is what I have. I've heard on the grapevine, with very little detail, that there isn't a problem with Type 6 keyboards under Redhat, so what's the problem with Debian? (I installed Debian because I wanted nice looking X, I've got that but now it's unusable. I can't win. At least I get usable ugly X under Solaris). Andrew ------------ Andrew Pollock Systems Integrator [EMAIL PROTECTED] http://www.asiaonline.net/ Phone: +61 2 6267 5610 Fax: +61 2 6200 2700 Asia Online
Purpose of shutdown and halt users
Hi, Is there any purpose for the shutdown and halt users? Is the intention there so that you can give operators an alternative means of shutting down a server other than giving them the root password? Thanks in advance Andrew Andrew Pollock Asia Online ABC Project [EMAIL PROTECTED]
php4-pgsql package
Hi, Anyone know what's happened to the php4-pgsql package in testing and unstable? [EMAIL PROTECTED]:~$ sudo apt-get install php4-pgsql Reading Package Lists... Done Building Dependency Tree... Done Some packages could not be installed. This may mean that you have requested an impossible situation or if you are using the unstable distribution that some required packages have not yet been created or been moved out of Incoming. Since you only requested a single operation it is extremely likely that the package is simply not installable and a bug report against that package should be filed. The following information may help to resolve the situation: Sorry, but the following packages have unmet dependencies: php4-pgsql: Depends: php4 (= 4.0.3pl1-0potato1.1) but 3:4.0.5-2 is to be installed E: Sorry, broken packages
Biometric devices supported by Linux
Hi, Does anyone know of any biometric (fingerprint recognition systems in particular) that can be made to work under Linux? I've got in mind something USB based... I'm going through the motions of looking at the feasibility of making a Linux-based attendance tracking system that will use fingerprints as the identifying key. Any pointers would be appreciated. I'm currently rummaging through www.biometrics.org Andrew
Bus mastering and dual port ethernet
Hi, I bought an Intel PRO/100 S Dual Port Server adapter to put in my PC. (This is supported by Linux isn't it?) and chucked it in my PC, and got lovely stuff like this when I booted up: eth0: Invalid EEPROM checksum 0xffc0, check settings before activating this device! eth0: OEM i82557/i82558 10/100 Ethernet at 0x5800, FF:FF:FF:FF:FF:FF, IRQ 0. Receiver lock-up bug exists -- enabling work-around. Board assembly ff-255, Physical connectors present: RJ45 BNC AUI MII Primary interface chip unknown-15 PHY #31. Secondary interface chip i82555. Self test failed, status : Failure to initialize the i82557. Verify that the card is a bus-master capable slot. eth1: Invalid EEPROM checksum 0xffc0, check settings before activating this device! eth1: OEM i82557/i82558 10/100 Ethernet at 0x5820, FF:FF:FF:FF:FF:FF, IRQ 0. Receiver lock-up bug exists -- enabling work-around. Board assembly ff-255, Physical connectors present: RJ45 BNC AUI MII Primary interface chip unknown-15 PHY #31. Secondary interface chip i82555. Self test failed, status : Failure to initialize the i82557. Verify that the card is a bus-master capable slot. The PC is an early Pentium 75, running kernel 2.2.14 with a kernel with Intel EtherExpress PRO support compiled in. Is it simply a case of my computer being too old, or am I doing something wrong? Andrew
ADSL problem with large packets
Hi, I just got ADSL running on my home network with a Debian gateway (two ethernet cards). I think I've got an MTU problem. Large transfers just don't work, but I can make an SSH connection out (for example) and pinging and the likes works. I've got pty "pppoe -I eth0 -T 80 -m 1412" and mtu 1412 in my peers file for the DSL connection as per documentation. I haven't messed with the MTUs on anything behind the Linux box. Should I? I thought the whole idea of this "MTU clamping" was this wasn't required? Andrew
Re: ADSL problem with large packets
On Thu, 23 Aug 2001, Andrew Pollock wrote: > Hi, > > I just got ADSL running on my home network with a Debian gateway (two ethernet > cards). > > I think I've got an MTU problem. Large transfers just don't work, but I can > make > an SSH connection out (for example) and pinging and the likes works. > > I've got > > pty "pppoe -I eth0 -T 80 -m 1412" > and > mtu 1412 > > in my peers file for the DSL connection as per documentation. I haven't messed > with the MTUs on anything behind the Linux box. Should I? I thought the whole > idea of this "MTU clamping" was this wasn't required? For the record, I'm pretty certain that simply upgrading pppoe to the version that's in unstable fixed the problem. Andrew
Looking for a wireless ethernet solution...
Hi, I'm looking for a wireless ethernet solution for my home network. I've got: 1 Linux box with two ethernet interfaces as my DSL gateway. 1 Windoze PC 1 Laptop (Windoze/Debian) I currently have the Linux gateway connected to hub and the Windoze PC and laptop go into the hub. Ideally, I'm after something like a "wireless hub" where I can eliminate the ethernet connection between the Linux box and the hub, still use a cable from the hub to the Windoze PC and use a wireless card in the laptop so I can take it anywhere in the house. Failing that I guess it's wireless cards for everything, but then if someone visits with a PC/laptop, they can't plug in... What's available that is supported by Linux? Can you mix and match 802.11(b) gear and expect it to cooperate? Are there any good sites for general reference? I'm finding new vendors daily, I'm after a definitive list... Andrew
Anyone else having problems with file uploads and PHP4?
Hi, Sometime after an upgrade, file uploads seem to have died with PHP4 (for me). Is anyone else having problems? I've upgraded all the way to 4.1-2 to try and shake the problem without success. The configuration file has the option switched on. Just wondering if it's "me" or PHP in general... Andrew
Apache dies on SIGUSR1 and SIGHUP
Hi, I'm running Apache 1.3.9-13.2 (is this a security update? It's not in stable, 13.1 is) and every time I do a "apachectl graceful" (send a USR1 signal) it just dies. Come to think of it, an "apachectl restart" kills it dead too... Even in debug mode, all it says is: [Thu Apr 12 22:42:15 2001] [notice] SIGUSR1 received. Doing graceful restart Is it just me? Andrew
Re: OpenSSH: How to do RhostsRSAAuthentication?
Hi, I've only got experience with SSH1, but I believe the process to be the same with SSH2. You've pretty much done everything right except you have to add the public identity you generated with ssh-keygen into ~/.ssh/authorized_keys on the box you want to SSH into with no password. I think you added it to the known hosts file instead. Hope this helps Andrew On 20 Apr 2001, Mario Vukelic wrote: > Hi, > > please help before I tear my hair out. I'm trying to get > RhostsRSAAuthentication to work. What I want is to be able to ssh > between the machines on my home network without having to supply a > passphrase/-word (also supplying it once with ssh-agent I'd like to > avoid). The docs I've found on OpenSSH don't say much about this special > method, but from what I gleaned from them, RhostsRSAAuthentication would > give me what I want. However, since the info is scarce, I'm not even > sure if it in fact does what I think it does. Although I'm on a rather > secure home network I don't want to use RhostsAuthentication, since I > want to learn how to configure OpenSSH properly, and rhosts-only > authentication is insecure. Also, there's always the possibility that > one time I'll allow ssh access from my external interface, and I don't > want to have to reconfigure it then. In any way, if I set > "RhostsAuthentication yes" in sshd_config it doesn't work either. > > This is what I've done: > > I've generated the host keys with > [EMAIL PROTECTED]:/etc/ssh# ssh-keygen -t dsa -f ssh_host_dsa_key (with empty > passphrase) > (now send ONE's /etc/ssh/ssh_host_dsa_key.pub to [EMAIL PROTECTED]) > [EMAIL PROTECTED]:~# mv ssh_host_dsa_key.pub /etc/ssh/ssh_known_hosts2 > > I did this for the other host, too. Then I prepared > /etc/ssh/ssh_known_hosts2 on both hosts by adding the hostname field as > described in man sshd (SSH_KNOWN_HOSTS FILE FORMAT). > > I've also generated user keys and distributed them > [EMAIL PROTECTED]:~/.ssh$ ssh-keygen -t dsa > (now send ~/.ssh/id_dsa.pub to [EMAIL PROTECTED]) > [EMAIL PROTECTED]:~$ mv id_dsa.pub .ssh/authorized_keys2 > (and vice versa) > > This is my config: > [EMAIL PROTECTED]:/etc/ssh# cat sshd_config > (excerpt) > Protocol 2,1 > HostKey /etc/ssh/ssh_host_dsa_key > IgnoreRhosts yes > IgnoreUserKnownHosts yes > RhostsAuthentication no > RhostsRSAAuthentication yes > RSAAuthentication yes > PasswordAuthentication yes > > [EMAIL PROTECTED]:/etc/ssh# cat ssh_config > (excerpt) > Host ONE > RhostsAuthentication no > RhostsRSAAuthentication yes > RSAAuthentication yes > PasswordAuthentication yes > FallBackToRsh no > UseRsh no > IdentityFile ~/.ssh/id_dsa > Protocol 2,1 > > [EMAIL PROTECTED]:/etc# cat hosts.equiv > +TWO > [EMAIL PROTECTED]:/etc# ls -l hosts.equiv > -rw-r--r-- 1 root root 13 Apr 20 12:17 ../hosts.equiv > > [EMAIL PROTECTED]:/etc# cat hosts.equiv > +ONE > [EMAIL PROTECTED]:/etc# ls -l hosts.equiv > -rw-r--r-- 1 root root 13 Apr 20 12:18 ../hosts.equiv > > Now I can ssh from TWO to ONE, and the host is already known to ssh, > although there is no ~/.ssh/known_hosts2. Therefore I think that the > host keys work. However, I still get asked for authentication: > [EMAIL PROTECTED]:~$ ls .ssh > authorized_keys2 id_dsa id_dsa.pub > [EMAIL PROTECTED]:~$ ssh ONE > Enter passphrase for key '/home/user/.ssh/id_dsa':[Enter] > [EMAIL PROTECTED]'s password:[Enter] > Permission denied, please try again. > [EMAIL PROTECTED]'s password:[Enter] > Permission denied, please try again. > [EMAIL PROTECTED]'s password:[Enter] > Permission denied (publickey,password). > [EMAIL PROTECTED]:~$ > > > It would be very nice if someone reviewed my config and told me if I've > commited mistakes somewhere (I'm afraid I wouldn't see it myself by now, > I'm already a bit dizzy after staring at the config files for hours). > Do I need a /etc/ssh/authorized_keys2, too. That is not mentioned in man > sshd, but still. > Any input is greatly appreciated. > > -- > > I did not vote for the Austrian government > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] >
Differences between distros in different architectures
Hi, Is it normal for the same distro (in this case, woody) on different architectures (in this case, i386 and SPARC) to be different? I got caught out rather severely today when I installed the sendmail package from woody on a SPARC box, and then on an i386 box. On the SPARC box, sendmail 8.11.3+8.12.0.Beta7-3 is linked against Berkeley DB version 2. iaamail:/usr/lib/cgi-bin/vmail# ldd /usr/lib/sendmail libdb2.so.2 => /lib/libdb2.so.2 (0x7002c000) libnsl.so.1 => /lib/libnsl.so.1 (0x7008a000) libldap.so.2 => /usr/lib/libldap.so.2 (0x700b) liblber.so.2 => /usr/lib/liblber.so.2 (0x700ea000) libresolv.so.2 => /lib/libresolv.so.2 (0x70106000) libwrap.so.0 => /lib/libwrap.so.0 (0x70128000) libsasl.so.7 => /usr/lib/libsasl.so.7 (0x7014) libc.so.6 => /lib/libc.so.6 (0x7015c000) libdl.so.2 => /lib/libdl.so.2 (0x70292000) libcrypt.so.1 => /lib/libcrypt.so.1 (0x702a6000) libpam.so.0 => /lib/libpam.so.0 (0x702e4000) /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x7000) On the i386 box, sendmail 8.11.3+8.12.0.Beta7-3 is linked against BerkeleyDB version 3 also, and includes the libdb3 package as a dependency, which forces a few other bits and pieces to get upgraded as well. [EMAIL PROTECTED]:~$ ldd /usr/lib/sendmail libdb3.so.3 => /usr/lib/libdb3.so.3 (0x4001c000) libnsl.so.1 => /lib/libnsl.so.1 (0x400c8000) libldap.so.2 => /usr/lib/libldap.so.2 (0x400dd000) liblber.so.2 => /usr/lib/liblber.so.2 (0x40103000) libresolv.so.2 => /lib/libresolv.so.2 (0x4010d000) libwrap.so.0 => /lib/libwrap.so.0 (0x4011f000) libsasl.so.7 => /usr/lib/libsasl.so.7 (0x40126000) libc.so.6 => /lib/libc.so.6 (0x40131000) libdb2.so.2 => /lib/libdb2.so.2 (0x40244000) libdl.so.2 => /lib/libdl.so.2 (0x40285000) libcrypt.so.1 => /lib/libcrypt.so.1 (0x40289000) libpam.so.0 => /lib/libpam.so.0 (0x402b7000) /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x4000) So I got very caught out when I installed sendmail from testing on the i386 box (after installing it on the SPARC box), it blew away libdb2-dev and I think it was the Perl BerkeleyDB module that I compiled myself broke. Just wondering if the mistake was mine in assuming equilibrium between same package versions in the same distribution versions on different architectures. Andrew
unable to write /var/run/sendmail/sendmail.pid: Permission denied
I get this everytime I start Sendmail, yet ls -ld /var/run/sendmail drwxr-s---2 mail mail 4096 May 2 16:17 /var/run/sendmail What am I doing wrong? I'm running 8.11.3+8.12.0.Beta7-3 Andrew
Re: ReiserFS + 2.4.4
On Wed, 2 May 2001, Oki DZ wrote: > William Leese wrote: > > oh, i didn't backup /proc and /tmp because tar spewed out a few error > > messages. but from what i can recall /proc is created by the kernel(?) and > > for /tmp the directory just needs to be recreated, correct? > > You'd need to recreate the /proc directory; the system uses it as the > mount point (of the proc filesystem). You can safely recreate the /tmp > directory; its content is always deleted on reboot. > > BTW, I have all my partitions running on reiserfs; problem is, when the > system booted up, / partition is always mounted read-only, so that the > transaction log is always replayed on, well, read-only filesystem. I > have done update-rc.d -f checkfs.sh remove, so that fsck wouldn't be > done on the system. But that doesn't make the root partition mounted > read-write on boot. The question is, how can I set the system so that > the / partition mounted read-write? Using LILO? image=/vmlinuz label=Linux read-only Lose the read-only directive. Andrew
Package configuration with /tmp mounted noexec
Hi, I've got /tmp mounted rw,noexec,nosuid,nodev because I think I read somewhere that that was a good way to go security-wise. It seems that some package related configuration stuff writes temporary scripts into /tmp, which then don't run because /tmp's mounted noexec Should perhaps such scripts be placed elsewhere? /var/tmp? Is mounting /tmp noexec a bit pointless? Andrew
Re: Package configuration with /tmp mounted noexec
On Tue, 23 Oct 2001, Karsten M. Self wrote: > on Tue, Oct 23, 2001 at 02:37:23PM +1000, Andrew Pollock ([EMAIL PROTECTED]) > wrote: > > Hi, > > > > I've got /tmp mounted rw,noexec,nosuid,nodev because I think I read > > somewhere > > that that was a good way to go security-wise. > > It is, but... > > > It seems that some package related configuration stuff writes > > temporary scripts into /tmp, which then don't run because /tmp's > > mounted noexec > > ...it creates problems. > > Incidentally, what package is doing this? I'd been asked this onece > after sugesting 'noexec' and wasn't aware of specific executables. I've > also found that the PCMCIA cardmgr wants to put a device file on /tmp, > and had to modify the init.d script for it to do a remount. This particular occasion was the faqomatic package, I was upgrading to the version in unstable. I'm not sure whether it's a debconf thing or a Perl thing. I'm still learning the internals of packages, and the scripts internal to the package don't make a lot of sense to me at the moment. > > Should perhaps such scripts be placed elsewhere? /var/tmp? Is mounting > > /tmp noexec a bit pointless? > > If you *do* specify a "TEMP=/var/tmp", most (but not all) applications > will respect it (though not necessarially in the morning). > > Note that *any* mount option is going to be relatively easy to change > with the -remount option -- this can be done without umounting the > partition. I'd prolly aquiesce and mount /tmp executable, seeing as > there are several pretty trivial ways of getting around this exclusion, > so it is somewhat pointless. Yeah, I think I'll do that. Andrew
SSL support for Sendmail
Hi, There doesn't appear to be a separate package for Sendmail that has SSL support, and I presume the standard Sendmail package doesn't have it, is that correct? Is it possible to do SSL'ed Sendmail easily with Debian (like for example, it is with IMAPS) Andrew
Troubleshooting sensible-mda
Hi, I've got some messages sitting in my mail queue that are deferred, with this message: local mailer (/usr/lib/sm.bin/sensible-mda) exited with EX_TEMPFAIL I'm not really sure how to troubleshoot the sensible-mda program, and therefore can't work out where the problem might be. I'm using a built-from-source version 8.12.1-2 of sendmail-tls Any help would be appreciated. Andrew
sendmail and disk quotas (was Re: Troubleshooting sensible-mda)
On Sun, 11 Nov 2001, Andrew Pollock wrote: > Hi, > > I've got some messages sitting in my mail queue that are deferred, with this > message: > > local mailer (/usr/lib/sm.bin/sensible-mda) exited with EX_TEMPFAIL > > I'm not really sure how to troubleshoot the sensible-mda program, and > therefore > can't work out where the problem might be. > > I'm using a built-from-source version 8.12.1-2 of sendmail-tls > > Any help would be appreciated. It would appear that this problem was related to the fact that the specific user in question had reached their disk quota. I'm assuming it's actually the MDA sendmail is configured to use and not so much sendmail's problem to deal with quotas? In this case, how can I configure the MDA to bounce messages with a fatal error if the specific user's disk quota has been exceeded? Andrew
Recpmmended Sendmail resources?
Is it just me, or does sendmail.org really suck? Especially in terms of up to date information? I can think of a few questions that I can never seem to easily find the answer to, like: how to change the banner that Sendmail spits out how to make it listen on only one interface how to disable the VRFY and EXPN commands for example. Is there a better, more up to date resource than sendmail.[org|net|com] or should I continue to Google the net for answers? Andrew
Good mail management techniques?
Hi, I'm just wondering how people manage their email... I've basically got a .procmailrc happening, which shoves each mailing list into it's own file in my home directory, and Pine, IMAP (via Webmail) accesses it all nicely. I have looked at Mutt, but being a Pine weeny it freaks me out everytime I try to use it too much. My main inbox is in /var/spool/mail The main problem I have is that each mailbox/folder/whatever you want to call it, grows without bounds. I wouldn't mind something to automatically shoved mail in a folder for each month or something like that, but I don't think that IMAP/Pine etc support multilevel folders, or do they? Anyway, I'm just interested in seeing how other people do it, and what is considered "best practise" Andrew
Samba problem with user level share permissions under Windoze
Hi, I'm trying to configure a Windoze 98 PC on my home network to use user level share permissions instead of share level permissions, and when I go to add users to a particular share, Windows barfs: You cannot view the list of users at this time. Please try again. and I get: [2001/09/03 14:21:29, 1] smbd/ipc.c:api_fd_reply(3314) api_fd_reply: INVALID PIPE HANDLE: 0 [2001/09/03 14:21:29, 1] smbd/ipc.c:api_fd_reply(3314) api_fd_reply: INVALID PIPE HANDLE: 0 [2001/09/03 14:24:15, 1] smbd/service.c:close_cnum(583) in /var/log/smb I haven't been able to find any mention of this in the Samba FAQ, and nothing overly useful with Google. Microsoft Knowledge Base article Q177607 refers to something related to Outlook Express, but I'm not sure if it affects me or not. Has anyone else been there, done that? Andrew
Sendmail trusted users
Hi, I'm using JAWmail and it calls sendmail directly with the -f option when sending mail. I've added www-data to /etc/mail/trusted-users but still sendmail complains about the -f option. I'm wondering if Sendmail doesn't like usernames with hypens in them? I'm running Sendmail 8.12.0.Beta7 Andrew
Re: Sendmail trusted users
On 05.09.2001 at 23:43:10, Richard A Nelson <[EMAIL PROTECTED]> wrote: > Did you `/etc/init.d/sendmail restart` after editing the file ? > > Any non-DB file update requires a restart I'm *sure* I've restarted it several times, but for the heck of it I've killed it dead and started from scratch. Does this message have warning headers? Andrew
Re: Sendmail trusted users
On 06.09.2001 at 02:50:23, Richard A Nelson <[EMAIL PROTECTED]> wrote: > Indeed, it does... > > Do you have > FEATURE(`use_ct_file')dnl # trusted-users > in /etc/mail/sendmail.mc ? Yup. and in sendmail.cf: Ft/etc/mail/trusted-users %[^\#] Troot Tdaemon Andrew
ide-smart and proactive drive monitoring
Hi,
It just crossed me as being a good idea (tm) to monitor my hard drives in case
of imminent failure, and sure enough the ide-smart package exists. Cool. One
problem, I have no idea how to read the output:
daedalus:/usr/doc/ide-smart# ide-smart /dev/hda
Id= 1, Status=41 {PreFailure , OffLine}, Value=100, Threshold= 20, Passed
Id= 3, Status=39 {PreFailure , OnLine }, Value= 57, Threshold= 20, Passed
Id= 4, Status=50 {Advisory, OnLine }, Value=100, Threshold= 8, Passed
Id= 5, Status=51 {PreFailure , OnLine }, Value=100, Threshold= 20, Passed
Id= 7, Status=11 {PreFailure , OnLine }, Value=100, Threshold= 23, Passed
Id= 9, Status=18 {Advisory, OnLine }, Value= 94, Threshold= 1, Passed
Id= 10, Status=38 {Advisory, OnLine }, Value=100, Threshold= 20, Passed
Id= 11, Status=19 {PreFailure , OnLine }, Value=100, Threshold= 20, Passed
Id= 12, Status=50 {Advisory, OnLine }, Value=100, Threshold= 8, Passed
Id= 13, Status=11 {PreFailure , OnLine }, Value=100, Threshold= 23, Passed
Id=199, Status=26 {Advisory, OnLine }, Value=200, Threshold= 0, Passed
Id=196, Status=16 {Advisory, OffLine}, Value=100, Threshold= 20, Passed
Id=197, Status=50 {Advisory, OnLine }, Value=100, Threshold= 20, Passed
Id=198, Status=16 {Advisory, OffLine}, Value=100, Threshold= 0, Passed
OffLineStatus=0 {NeverStarted}, AutoOffLine=No, OffLineTimeout=0 minutes
OffLineCapability=27 {Immediate Auto SuspendOnCmd}
SmartRevision=11, CheckSum=189, SmartCapability=3 {SaveOnStandBy AutoSave}
Does anyone know how to decipher it, or of any good SMART references?
Andrew
Re: ide-smart and proactive drive monitoring
To answer my own question, the ucsc-smartutils package provides a SMART
monitoring daemon and another tool for also retrieving SMART statistics. I
guess I just need to employ one of the log watching tools to look for
errors from the daemon.
I still wouldn't mind a better understanding of the actual output,
smartctl produces something similar but in a different format:
Vendor Specific SMART Attributes with Thresholds:
Revision Number: 11
AttributeFlag Value Worst Threshold Raw Value
( 1)Raw Read Error Rate 0x0029 100 253 020
( 3)Spin Up Time0x0027 057 057 020 153f
( 4)Start Stop Count0x0032 100 100 008 0015
( 5)Reallocated Sector Ct 0x0033 100 100 020
( 7)Seek Error Rate 0x000b 100 100 023
( 9)Power On Hours 0x0012 094 094 001 10af
( 10)Spin Retry Count0x0026 100 100 020
( 11)Unknown Attribute 0x0013 100 100 020
( 12)Power Cycle Count 0x0032 100 100 008 0015
( 13)Unknown Attribute 0x000b 100 093 023
(199)UDMA CRC Error Count0x001a 200 200 000
(196)Reallocated Event Count 0x0010 100 253 020
(197)Current Pending Sector 0x0032 100 100 020
(198)Offline Uncorrectable 0x0010 100 253 000
regards
Andrew
On Mon, 10 Sep 2001, Andrew Pollock wrote:
> Hi,
>
> It just crossed me as being a good idea (tm) to monitor my hard drives in case
> of imminent failure, and sure enough the ide-smart package exists. Cool. One
> problem, I have no idea how to read the output:
>
> daedalus:/usr/doc/ide-smart# ide-smart /dev/hda
> Id= 1, Status=41 {PreFailure , OffLine}, Value=100, Threshold= 20, Passed
> Id= 3, Status=39 {PreFailure , OnLine }, Value= 57, Threshold= 20, Passed
> Id= 4, Status=50 {Advisory, OnLine }, Value=100, Threshold= 8, Passed
> Id= 5, Status=51 {PreFailure , OnLine }, Value=100, Threshold= 20, Passed
> Id= 7, Status=11 {PreFailure , OnLine }, Value=100, Threshold= 23, Passed
> Id= 9, Status=18 {Advisory, OnLine }, Value= 94, Threshold= 1, Passed
> Id= 10, Status=38 {Advisory, OnLine }, Value=100, Threshold= 20, Passed
> Id= 11, Status=19 {PreFailure , OnLine }, Value=100, Threshold= 20, Passed
> Id= 12, Status=50 {Advisory, OnLine }, Value=100, Threshold= 8, Passed
> Id= 13, Status=11 {PreFailure , OnLine }, Value=100, Threshold= 23, Passed
> Id=199, Status=26 {Advisory, OnLine }, Value=200, Threshold= 0, Passed
> Id=196, Status=16 {Advisory, OffLine}, Value=100, Threshold= 20, Passed
> Id=197, Status=50 {Advisory, OnLine }, Value=100, Threshold= 20, Passed
> Id=198, Status=16 {Advisory, OffLine}, Value=100, Threshold= 0, Passed
> OffLineStatus=0 {NeverStarted}, AutoOffLine=No, OffLineTimeout=0 minutes
> OffLineCapability=27 {Immediate Auto SuspendOnCmd}
> SmartRevision=11, CheckSum=189, SmartCapability=3 {SaveOnStandBy AutoSave}
>
> Does anyone know how to decipher it, or of any good SMART references?
>
> Andrew
>
>
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
>
Anyone played with Festival?
Hi, I had a bit of a play with Festival on my laptop today. I've got it working (or talking as the case may be) but I wanted to try and get the client/server mode part of it working, as the delay in initializing was annoying me. The server starts fine, but whenever I try to connect to port 1314, it closes the connection immediately. I did an strace on the server process, and the two lines that looked interesting where: shmat(8, 0x1, 0x1ptrace: umoven: Input/output error ) = ? shmat(8, 0x1, 0x2ptrace: umoven: Input/output error ) = ? Does this imply a shared memory problem? I scoured the documentation for any reference of shared memory and couldn't find anything. I also can't figure out how to turn on any logging, which is making troubleshooting difficult. I'm suffering from Scheme/LISP culture shock :-) Andrew
ReiserFS revisited
Hey, I'm soon to redo my home ADSL gateway/firewall/muckaround box and wanted to toy with ReiserFS. What's the status of ReiserFS Debian bootdisks? Where's the latest greatest unofficial ones obtainable from? There's some dated stuff on Debian Planet saying that it's still a bit dodgey, but I don't think that's the case any more is it? Andrew
/etc/resolv.conf disappears on reboot
Hi, I've got a Debian SPARC box that was initially setup to use DHCP, but now has a statically assigned IP address, and every time it boots, it wipes out /etc/resolv.conf I've been unable to track down what's doing it. Are there any other DHCP settings hanging around that might be doing it? Andrew
Re: /etc/resolv.conf disappears on reboot
On Thu, 3 May 2001, Andrew Pollock wrote: > Hi, > > I've got a Debian SPARC box that was initially setup to use DHCP, but now > has a statically assigned IP address, and every time it boots, it wipes > out /etc/resolv.conf > > I've been unable to track down what's doing it. Are there any other DHCP > settings hanging around that might be doing it? Found it. Didn't know that pump had replaced dhcpcd Andrew
Setting the time with Samba
Hi, I was being lazy, and was going to sync the time on my Windoze box against my Linux box, using the "net time" command from a DOS box. Here's what happened: C:\WINDOWS>net time \\caesar Current time at \\CAESAR is 5-26-2001 2:53A.M. The command was completed successfully. But on my Linux box: caesar:/home/apollock# date Sat May 26 17:56:20 EST 2001 I'm pretty sure I do have my hardware clock set to UTC (how can you tell?), but even then: caesar:/home/apollock# date --utc Sat May 26 07:57:06 UTC 2001 So I'm at a bit of a loss to work out what's going on with respect to the time differences. The timezone on the Linux box is same as the Windoze box. Any suggestions? Andrew
Re: Setting the time with Samba
Sorry if my problem was not clear... The Windoze box is reporting a totally different time to what's on the Linux box when I use the net time command. On Sat, 26 May 2001, Mike Egglestone wrote: > Hi... > The Samba list guys should know more about this stuff... > but you may want to try this in your netlogon batch file.. > > net use \\samba /set /y > > Hope this helps... > Mike > - Original Message - > From: "Andrew Pollock" <[EMAIL PROTECTED]> > To: > Sent: Saturday, May 26, 2001 12:58 AM > Subject: Setting the time with Samba > > > > Hi, > > > > I was being lazy, and was going to sync the time on my Windoze box against > > my Linux box, using the "net time" command from a DOS box. > > > > Here's what happened: > > > > C:\WINDOWS>net time \\caesar > > Current time at \\CAESAR is 5-26-2001 2:53A.M. > > The command was completed successfully. > > > > But on my Linux box: > > > > caesar:/home/apollock# date > > Sat May 26 17:56:20 EST 2001 > > > > I'm pretty sure I do have my hardware clock set to UTC (how can you > > tell?), but even then: > > > > caesar:/home/apollock# date --utc > > Sat May 26 07:57:06 UTC 2001 > > > > So I'm at a bit of a loss to work out what's going on with respect to the > > time differences. > > > > The timezone on the Linux box is same as the Windoze box. > > > > Any suggestions? > > > > Andrew > > > > > > -- > > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > > with a subject of "unsubscribe". Trouble? Contact > [EMAIL PROTECTED] > > > > > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] >
/etc/mail/nsswitch.conf (and /etc/mail/service.switch*)
Are these used by sendmail or are they there for demonstration purposes? I'm running 8.11.3+8.12.0.Beta7-8, and it's doing some kooky stuff with respect to DNS resolution and mail delivery. It's swearing black and blue that a domain isn't resolving so it's queuing it, but I can resolve it by hand. Very confusing, and very annoying. Even if you send a test message by connecting to port 25, and manually talking SMTP, after going RCPT TO: it goes "Recipient ok, (will queue)", so it categorically thinks the domain's uncontactable (or at least unresolveable). Just wondering if the switch files have anything to do with it. Andrew
Re: /etc/mail/nsswitch.conf (and /etc/mail/service.switch*)
On Tue, 29 May 2001, Richard A Nelson wrote: > > On Tue, 29 May 2001, Andrew Pollock wrote: > > > Subject: /etc/mail/nsswitch.conf (and /etc/mail/service.switch*) > > > > Are these used by sendmail or are they there for demonstration purposes? > > They are indeed used, and should mirror your /etc/nsswitch.conf (barring > the small syntatic change). > > > I'm running 8.11.3+8.12.0.Beta7-8, and it's doing some kooky stuff with > > respect to DNS resolution and mail delivery. It's swearing black and blue > > that a domain isn't resolving so it's queuing it, but I can resolve it by > > hand. Very confusing, and very annoying. > > I'll bet you've got a b0rked name server in your path... > > Install Beta10-1,Beta7-9, or if they're not yet available, add this to > /etc/mail/sendmail.mc: > define(`confBIND_OPTS', `+WorkAroundBroken')dnl > > run sendmailconfig (or make) > > restart/reload sendmail That was amazing. The mail queue is like nearly empty now. What does it do exactly? In what way is one of my nameservers b0rked? > > Even if you send a test message by connecting to port 25, and manually > > talking SMTP, after going RCPT TO: it goes "Recipient ok, (will queue)", > > so it categorically thinks the domain's uncontactable (or at least > > unresolveable). > > > > Just wondering if the switch files have anything to do with it. > > Not with this problem... > > -- > Rick Nelson > you people are all insane. > knight: sure, that's why we work on Debian. > Knghtbrd: get in touch with your inner nutcase. >
apache: dl-close.c:119: _dl_close: Assertion `new_opencount[0] == 0' failed.
I added a new virtual webserver to Apache today and restarted it, and now it won't start, just barfs this out. Even when I back out the config changes I made :-( Any ideas why it's happening? I'm running 1.3.19-1 from unstable. Andrew
Re: ls nitpick
On Sat, Jan 03, 2004 at 10:31:41PM -0500, Russ Schneider wrote: > When you do an ls on Debian, you see something like the following: > file1 file2 file3 dir1 > dir2 file4 > > etc. > > When you do the same on Mandrake, you get > file1 file2 file3 dir1/ > dir2/ file4 > > You see how there's a / at the end of each directory name, making it > really easy to tell at a glance what's a directory and what's not? > > Any way to config Debian's ls to do that? I realize it's just a nitpick, > but I am curious. In your home directory, you should have a .bashrc file put there from /etc/skel when your account was created. It contains a lot of goodies that are commented out, in particular: # enable color support of ls and also add handy aliases eval `dircolors -b` alias ls='ls --color=auto' If you uncomment the above two lines, and add a -F option to the ls alias, it will walk and talk similarly to how Mandrake's ls presumably does. Run 'alias ls' on your Mandrake box to determine exactly how they've aliased it if you want to make it identical. Naturally, you will need to log in and log out again (or close and open a fresh xterm, as the case may be). You may also wish to uncomment the lines in your .bash_profile that source the .bashrc if it exists. HTH Andrew signature.asc Description: Digital signature
Re: MMap ran out of room
On Tue, Jan 06, 2004 at 08:53:13PM -0500, Kevin Mark wrote: > On Tue, Jan 06, 2004 at 12:51:56PM -0800, Karsten M. Self wrote: > > # echo 'APT::Cache-Limit "25165824";' >> /etc/apt/apt.conf > > > Hi list, > it seems that random newbies come across this. And it seems like a > simple fix to set this Apt variable. Coun't Apt be set with a more > reasonable value so that this simple problem can be avoided in MOST > cases? Or how do you figure this value? > -Kev Kevin, APT has been fixed up in newer versions, but it's not a security issue, and so stable doesn't get updated with a newer version. Andrew -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Woody with ext3, 2.4 kernel + custom install questions
Hi, I've been given the opportunity at my place of work (which is currently a Mandrake shop) to tout Debian. I was intending to use Woody, and have created the 8 CD's. I work for a managed security provider, and one of the reasons that they are using Mandrake over the likes of Red Hat is because of the control Mandrake allows over what gets installed. (i.e. when you say you want nothing, you get exactly that. The exact example that was told to me was with Red Hat you'd say you wanted nothing installed, but the thing would still listen on port 25. I have to say that even a base install of Debian has port 25 open, which is going to unimpress some people here...) Any, question #1: Where can I get a boot disk for Woody that has ext3 support (and a 2.4 kernel). More to the point, where is it _documented_? I rummaged around on this list and found that if you boot off disc 3, you apparently get ext3 (and presumably a 2.4 kernel) however I tried this on the SCSI system that I was playing with and that kernel doesn't support SCSI. and question #2: Is it possible to automate the installation process of Debian at all? Red Hat has KickStart, and Mandrake has some autoinst.img thingy. I'd like to be able to provide a set of packages for it to pre install. Could I create my own task package or something? Is there a set of documentation for customising the installation at all? Any help would be appreciated, I'd really like to see Debian get up here. Andrew -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Woody with ext3, 2.4 kernel + custom install questions
On Fri, 26 Apr 2002, Shawn McMahon wrote: > begin Andrew Pollock quotation: > > > > I work for a managed security provider, and one of the reasons that they > > are using Mandrake over the likes of Red Hat is because of the control > > Mandrake allows over what gets installed. (i.e. when you say you want > > nothing, you get exactly that. The exact example that was told to me was > > with Red Hat you'd say you wanted nothing installed, but the thing would > > still listen on port 25. I have to say that even a base install of Debian > > has port 25 open, which is going to unimpress some people here...) > > Define what they meant by "port 25 open". If you don't install an SMTP > daemon of any kind, such as sendmail or exim, you won't have anything > listening on that port, but "open" means different things in different > contexts. Debian installs exim by default. i.e. it doesn't ask you if you'd like an SMTP server, it installs it. Sure, one of the very next things it asks you is how do you want exim configured, but I believe even if you choose the "do nothing" response, it leaves exim activated via inetd. > Also, "want nothing installed" is irrational. If NOTHING is installed, > you won't have any ports listening, because you'll have a blank hard > drive. You can't say "when I installed RedHat (or Mandrake or Debian > etc.) I told it to install nothing." It's nonsensical. Okay, "nothing" vs "minimalist", meaning you get a bare bones system with just the bare necessities, and anything beyond that you explicitly choose to install. > Either you're misremembering what was said, or the person saying it was > very very confused. Basically what was said about Red Hat's "minimalist" install was it included too much. Andrew > -- > Shawn McMahon| McMahon's Laws of Linux support: > http://www.eiv.com | 1) There's more than one way to do it > AIM: spmcmahonfedex, smcmahoneiv | 2) Somebody thinks your way is wrong > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Sendmail: How to get at the envelope in Procmail?
Hi, I've got a setup whereby I've got an entire subdomain being directed at one mailbox (using a virtusertable). I'd like the procmail recipe in that mailbox to be able to access the intended recipient, which may not necessarily be what's in the "To:" header. Can anyone offer any suggestions on how to go about that? Andrew -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Sendmail: How to get at the envelope in Procmail?
On Tue, 14 May 2002, Carel Fellinger wrote: > I'm not sure I understand the setup, so I may be way off here. > > The best solution is to have your IPS (or who ever fills that mailbox) > add a Delivered-To header with the envelop To header. It's my mail server, so I control it. Your suggestion sounds good. I'll do some Googling, but can you tell me how to configure Sendmail to add this header? (Preferably what to add to my sendmail.mc so it'll persist) Andrew -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Pentium optimised vs not
Hi, I'm working in what is a Mandrake shop, but relatively open minded about Debian. I'm already making inroads into getting Debian used as the distro of choice for infrastructure boxes. Thanks to FAI. Nice work Thomas. Mandrake alledgedly compile all their binaries with Pentium optimisation. I was recently asked if Debian's binaries were optimised. I didn't think they were. I was wondering, in reality, if there were any significant performance gain from doing it? Let's say I wanted to keep a local package repository build from the source packages, how trivial is it when (re) building the packages to enable Pentium optimisation? Andrew -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
"Debian way" for permanent static routes?
Hi, What's the "Debian way" of having static routes added when an (ethernet) interface comes up, every time? Andrew -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bringing up an interface without an IP address
Hi, I have a requirement for an IDS, that the interface it listens on be up, and in promiscuous mode without an IP address assigned to it. I'm trying to work out how best to achieve this, preferably, doing it "the Debian way". >From reading interfaces(5) there appears to be no way to directly stipulate >that an interface be brought up in promiscuous mode, so I figured adding an "up ifconfig eth0 promisc" to the stanza for that interface would do the trick. I'm also giving the interface the address 0.0.0.0, which ifup complains about, but still brings up the interface. The problem I'm having is the "up" part isn't being run, I assume because of the error returned from trying to set the interface address to 0.0.0.0 Can I do what I need to do using the interfaces file or am I reduced to a more clumsy hack? Andrew -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Satisfying dependencies on virtual packages
Hi, I'm wondering how APT satisfies the dependencies a package has on a virtual package. (Specially, the netsaint package depends on httpd). On a box I was installing that had no package installed that provided httpd, APT chose to install the aolserver server package, and I'm wondering why of the 9 or so packages that provide httpd (is there an easier way of finding this out than grepping /var/lib/dpkg/available?) aolserver was chosen? It's not the first package in my available file to provide httpd, and it's not even the first package alphabetically to provide it. Andrew -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
