4 replies to this topic

[IBBoard]

My VPS just went down for ten minutes. Everything was unresponsive - no HTTP, email or SSH. I logged in to Virtuozzo and only iptables was up. I did a full restart and everything is okay now, but nothing seems obvious in the logs. I'm able to work out what is wrong if I have info, but I can't work out how to get the info.

/var/log/messages shows the following at the time of the outage (with the line before included as well):
[code]
Mar 22 16:01:15 vps sshd(pam_unix)[20330]: session closed for user ***
Mar 27 11:57:54 vps shutdown: shutting down for system halt
Mar 27 11:57:56 vps init: Switching to runlevel: 0
Mar 27 12:57:57 vps saslauthd[11277]: server_exit     : master exited: 11277
Mar 27 11:57:57 vps saslauthd: saslauthd shutdown succeeded
Mar 27 11:57:58 vps sshd: sshd shutdown succeeded
Mar 27 11:57:58 vps postfix: Shutting down postfix:
Mar 27 11:57:59 vps postfix:  succeeded
Mar 27 11:57:59 vps postfix: [
Mar 27 11:57:59 vps postfix:
Mar 27 11:57:59 vps rc: Stopping postfix:  succeeded
Mar 27 11:58:00 vps postfix: postfix stop failed
Mar 27 11:58:00 vps dovecot: dovecot shutdown succeeded
Mar 27 11:58:01 vps psmon: psmon shutdown succeeded
Mar 27 11:58:01 vps spawn-php: Stopping spawn-php:
Mar 27 11:58:02 vps spawn-php:  succeeded
Mar 27 11:58:02 vps spawn-php:
Mar 27 11:58:02 vps rc: Stopping spawn-php:  succeeded
Mar 27 11:58:03 vps postfix:  succeeded
Mar 27 11:58:03 vps svnserve: svnserve shutdown succeeded
Mar 27 11:58:07 vps mysqld: Stopping MySQL:  succeeded
Mar 27 12:58:07 vps xinetd[9724]: Exiting...
Mar 27 11:58:08 vps xinetd: xinetd shutdown succeeded
Mar 27 11:58:08 vps crond: crond shutdown succeeded
Mar 27 12:58:08 vps kernel: Kernel logging (proc) stopped.
Mar 27 12:58:08 vps kernel: Kernel log daemon terminating.
Mar 27 11:58:10 vps syslog: klogd shutdown succeeded
Mar 27 12:58:10 vps exiting on signal 15
Mar 27 12:10:17 vps syslogd 1.4.1: restart.
Mar 27 12:10:17 vps syslog: syslogd startup succeeded
Mar 27 12:10:17 vps kernel: klogd 1.4.1, log source = /proc/kmsg started.
Mar 27 12:10:17 vps syslog: klogd startup succeeded
...
[code]
(I'm assuming the rogue 12:58 time is a timezone issue). dmesg is empty, bootlog just shows the same info but without the timestamps, and I can't think whatelse might have useful information. I'm not even sure whether 11:58 was everything shutting down by itself or whether that was the restart and it took a while to process (although 12 minutes seems far too long).

Any ideas?

Thanks.

[davidatfuzzylime]

Hi Frank, I was just coming on here to ask much the same question, as mine went down. Is there any chance of VPS users getting a little bit of notice of this in future - even 5mins - just in case someone phones and asks us what's up with their/our website? An e-mail would be ideal, but even a post here would be helpful …

[IBBoard]

It wasn't terribly inconvenient, although it did seem like things went down and didn't come back up again. There wasn't anything in the logs that looked obvious, but the site went down and Virtuozzo seemed to let me restart services without obviously saying "server is being rebooted" or anything.

As David says, a bit of warning would be good Installing KSplice sounds like it'll have been worth it, though.

Now to work out whether to ask about a move to a CentOS5 server over Easter or wait until the new server details are announced...