[ydl-gen] Syslog question regarding mysterious shutdowns

[Derick Centeno]

On Fri, 21 May 2010 14:50:37 -0500
Nick Jones <nicholas-jones at uiowa.edu> wrote:

> Hi,
> 
> I have an xserve G5 running YDL 4.1 (I'm pretty sure, can't
> find a way to check this easily nor can I remember, but did a
> package search and that is what came up).  

To find out what your system is running, just open a terminal
and do:

$ uname -r
2.6.29-3.ydl61.4

Explanation:
At the $ you execute (from within YDL) the uname command with
the r flag.  The system will respond it did above.  Here's a
basic label explaining what the system is reporting:

-> 2: the Kernel version
-> 6: Major revision # of the kernel
-> 29: Minor revision # of the kernel
-> 3: revision # correcting critical bugs
-> ydl61.4: YDL version #

> It started gracefully shutting down on me every once in awhile
> for no rhyme or reason that I can see (I say graceful because
> I've encountered no signs of a crash, such as an inconsistent
> filesystem).
> 
> I can't find anything in the logs regarding why it is shutting
> down.  My conclusion is that syslogd is not saying anything
> about the shutdown to /var/log/messages or anything else
> under /var/log/*

You may already be familiar with the following strategies which
I'll invoke on my own system as an example:

Explanation: The following checks the running of syslogd
$ ps aux |grep syslogd
root  3248  0.0  0.0   2036   688 ?  Ss 17:09  0:00 syslogd -m 0 
aguila 29234  0.0 0.0 4416  708 pts/2 S+ 19:12 0:00 grep syslogd 

Explanation: The following quickly checks the boot logs for
outputs from syslog/d. 
$ dmesg |grep syslogd
$ dmesg |grep syslog

Explanation: The following checks for any output from anything
named by "sys" just in case grep missed something.
$ dmesg |grep sys
SCSI subsystem initialized
SGI XFS Quota Management subsystem
EXT3-fs: mounted filesystem with ordered data mode.
EXT3-fs: mounted filesystem with ordered data mode.
warning: process `kudzu' used the deprecated sysctl system call
with 1.23. sysctl net.netfilter.nf_conntrack_acct=1 to enable it.
$ 

> It is on a battery backup with several other pieces of
> hardware, all of which exhibit no problems related to power.
> I disconnected the USB cable from the battery backup but the
> problem persists.
> 
> Any help would be appreciated.  I hope I'm not missing
> something obvious.
> 
> Nick

Is it possible that your server is being shut down by the backup
battery system you use because the power system is doing it's
job and shutting down when the server demands more power than
the backup power system can provide?  Maybe it's time to :

* check that the installed YDL version you has software
  monitoring the backup power supply system you use 

* check that your batteries on your back up system are "fresh"
  and don't require either upgrade or replacement as your power
  requirements may change or have changed depending upon the
  load of power your server requires as work intensifies.

I see that you are running YDL 4.1.  I checked the Fixstar
hardware pages for YDL 6.2 looking for any description
approaching what you described.  I found interesting comments,
but I'm not sure anything there applies to the model you have:

http://www.ydl.net/support/hardware/apple-xserve.shtml

Hopefully, the Fixstars webpage can help provide you enough
information to suggest whether an upgrade to YDL 6.2 would be
useful for you.

Besides that option I thought you should check to make sure
whether your version of yum is pointing to Fedora extras as
software supporting apc UPS systems exist there.  What follows
is a search I invoked for "apc"; I've already provided yum the
information it needs to search for it at Fedora Extras:

$ sudo yum info "*apc*"
Password:
Loading "protectbase" plugin
Loading "installonlyn" plugin
PS3Bodega                 100% |=========================| 1.9 kB    00:00     
livna-stable              100% |=========================| 2.1 kB    00:00     
base                      100% |=========================| 1.1 kB    00:00     
updates                   100% |=========================|  951 B    00:00     
fedora-extras             100% |=========================| 1.1 kB    00:00     
dribble                   100% |=========================|  951 B    00:00     
Excluding Packages from Livna for Fedora Core 6 - ppc - Base
Finished
Excluding Packages from Fedora Extras
Finished
121 packages excluded due to repository protections
Available Packages
Name   : apcupsd
Arch   : ppc
Version: 3.12.4
Release: 5.fc6
Size   : 449 k
Repo   : fedora-extras
Summary: APC UPS Power Control Daemon for Linux
Description:
Apcupsd can be used for controlling most APC UPSes. During a
power failure, apcupsd will inform the users about the power
failure and that a shutdown may occur.  If power is not restored,
a system shutdown will follow when the battery is exausted, a
timeout (seconds) expires, or the battery runtime expires based
on internal APC calculations determined by power consumption
rates.  If the power is restored before one of the above shutdown
conditions is met, apcupsd will inform users about this fact.
Some features depend on what UPS model you have (simple or smart).

Name   : apcupsd-cgi
Arch   : ppc
Version: 3.12.4
Release: 5.fc6
Size   : 150 k
Repo   : fedora-extras
Summary: Web interface for apcupsd
Description:
A CGI interface to the APC UPS monitoring daemon.

Name   : perl-Mail-IMAPClient
Arch   : noarch
Version: 2.2.9
Release: 3.fc6
Size   : 420 k
Repo   : fedora-extras
Summary: An IMAP Client API
Description:
This module provides perl routines that simplify a sockets connection
to and an IMAP conversation with an IMAP server.

Name   : php-pecl-apc
Arch   : ppc
Version: 3.0.12
Release: 1.fc6
Size   : 77 k
Repo   : fedora-extras
Summary: APC caches and optimizes PHP intermediate code
Description:
APC is a free, open, and robust framework for
caching and optimizing PHP intermediate code.

$ 

My system runs from a laptop so I don't need the ups utilities
as you may.  The ups utilities could provide a means for your
version of YDL to manage communications with your ups better
allowing the xserve to run longer.  

If you are interested I can share the code I use to program yum
to search the livna and Fedora extras repositories (and
install from them) while protecting the YDL base installation.  

All the best...
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://lists.fixstars.com/pipermail/yellowdog-general/attachments/20100521/302c0b79/attachment-0001.bin>