Random crashes after upgrading to 2.2
-
Hi all,
I'm running pfsense via VMWare esxi and since upgrading i'm getting 1-2 complete system locks/day. In the console I don't see anything when it crashes and it is completely unresponsive. Couldn't find anything in the logs either. Today it happened at 12:01 so I thought it could've been related to a cronjob, but I can't figure out what would cause it exactly.
Has anyone encountered similar issues? Any ideas on how to debug? Since I don't seem to get a kernel panic..
Thanks!
-
do you have a virtual cd drive in the VM ? if yes –> remove it (not sure if its still an issue on 2.2 )
are you running acpi-safe? https://lists.freebsd.org/pipermail/freebsd-stable/2012-August/069070.html (not sure if its still an issue on 2.2)are you running carp? see upgrade notes https://doc.pfsense.org/index.php/Upgrade_Guide#pfSense_2.2_Upgrade_Notes
do you get anything in the console when it happens? might not be in logs, but there should be something in the console.
there is no well know issue with 2.2 running in esxi AFAIK (i run a couple myself) … so more info is needed to debug your situation.
-
do you have a virtual cd drive in the VM ? if yes –> remove it (not sure if its still an issue on 2.2 )
are you running acpi-safe? https://lists.freebsd.org/pipermail/freebsd-stable/2012-August/069070.html (not sure if its still an issue on 2.2)Take that one step farther: remove the virtual CD-ROM device entirely, not just the media.
-
The most likely cause of a lockup in ESX is a mismatched architecture. For example, running an i386 VM inside a VM container set to "FreeBSD 64-bit".
Make sure both the OS and the container are set to 64-bit/amd64.
-
do you have a virtual cd drive in the VM ? if yes –> remove it (not sure if its still an issue on 2.2 )
are you running acpi-safe? https://lists.freebsd.org/pipermail/freebsd-stable/2012-August/069070.html (not sure if its still an issue on 2.2)Take that one step farther: remove the virtual CD-ROM device entirely, not just the media.
thats what i meant :)
-
It appeared to be fixed for a few days, but sadly the issue re-occured just now! So I have removed the virtual CD completely and hopefully it'll be fixed with that.
I'm still running i386, which is what the ESX was configured as as well. Might just reinstall a clean x64 one and restore a config file/backup into that to see if that fixes it if the issues persist.
Thanks for suggestions will update if it happens again.
-
I'm also seeing this issue, also running a 'mismatched' architecture since the host is x64 and the VMs are x86.
In some cases a clean install is more difficult (since that'd leave the customer without internet access in many cases to perform the reinstall, so it requires being there), so I tried other options that could be done remotelly.
I eventually found out that disabling IP compression on the advanced ipsec options seems to have fixed the issue, went from at least a crash/reboot per hour to running more than 24h stable and counting, running both IKE v1 and v2 tunnels (so that's not the source of the issue in theory).
Clean reinstall on proper arch is probably the best solution long-term anyway, but if the issue is with IPcomp then disabling it might help others who are now facing issues. Will update if I see any more crashes.
