As James indicated, start looking at /var/log/messages and /var/log/syslog, which might show bad things occurring before the apocalyptic kernel panic (if that is what is happening).
Netxt, though I haven't done this sort of thing for a while, you can arrange for processes and/or the kernel to "dump core" into the file system which can be analysed with various utilities. >From what I have read, if you set up the kernel crashdump facility, in the event of a kernel panic another crashdump kernel can start and write a snapshot of the runtime kernel to disk. Have a look at this for a starting point. https://wiki.ubuntu.com/Kernel/CrashdumpRecipe However as James has pointed out it could be a host issue, in the way it presents the virtual hardware to your guest. If the host does do something bad you just may not see it (and the guest-based crashdump I have mentioned above won't help). Your provider could be able to provide some logs that would indicate at least the host sees concerning the untimely reincarnation of your guest ;-) Regards, Martin [email protected] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
