Last weekend I upgraded a Linode from Debian Squeeze to Wheezy via 'apt-get dist-upgrade'. Everything went well at the time but since then the VM has randomly been rebooting once or twice a day.

I have a backup so it's not too difficult to rollback or build a new Wheezy VM and migrate the data but it'd be nice to figure it out if I can. There's nothing obvious being logged anywhere and the timing isn't consistent nor is the reboot linked to a particular activity (that I can discern).

Whenever it reboots I just get a load of junk in the syslog and then the next message is it coming back up. The timing of the last message often doesn't line up with the reboot as you can see

Jun 22 14:21:51 ty-01 postfix/anvil[14511]: statistics: max cache size 1 at Jun 22 14:18:29
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@Jun 22 14:33:30 ty-01 ke
rnel: imklog 5.8.11, log source = /proc/kmsg started.
Jun 22 14:33:30 ty-01 rsyslogd: [origin software="rsyslogd" swVersion="5.8.11" x-pid="3466" x-info="http://www.rsyslog.com"] start

Does anyone have any suggestions before I blow it away and start again? None of the usual culprits seem to be at fault here

Cheers

I searched Google for "rsyslog random reboot" and found this old bug: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=509292

No idea if it's related. I'd check the Debian forums and check your system resources to make sure you're not maxing anything out. You can file a support ticket and ask for it to be migrated to another machine, file a ticket anyway since there could be some hardware issues on the machine your VM is on so Linode can investigate. That will also enable them to correlate your issue with any similar reports.

Does anything show up on the lish console around the time of the reboot? The 'logview' command will spit out the console logs.

_________________

/* TODO: need to add signature to posts */

Thanks for the tip, I didn't realise I could view historical data from the LISH console. It was also pretty helpful, looks like it's a kernel panic so now I've got somewhere to start digging further.



Good idea, I'll file a ticket so they can check. I don't think it's a resource issue, I do monitor the VM and haven't seen anything out of the ordinary. Initially when it started happening I thought it might have been snmpd as it was usually the last daemon to log anything before the crash but with snmpd disabled it's still happening.

Are you on the latest Linode kernel?

I was but at the suggestion of the support guys I dropped the linode back to the previous version. Currently running 3.8.4-x86_64-linode31 and it's been up solid for almost 2 days. Will keep an eye on things but it definitely seems to have helped.