Page 1 of 1

[Sort-of-solved] Help - incessant kernel oops, freezing etc.

Posted: 24. Aug 2009, 17:12
by urilabob
Host: CentOS 5.3 64bit
VBox: 3.0.4
Guest: Fedora 11 64bit
Hardware: 4*4core intel, 16GB with guest getting 4core, 1GB
Bridged networking, with separate NICs for host and guest

I've been trying for a week or more to get a workable system running :cry: I'm about ready to give up, but I thought I'd make an appeal for any hints. For the third time around, I've now got to the stage of an installed fedora, and doing the post-install software update (got to here before and lost _all seven_ of my snapshots in the last crash :? ). To get to this stage has required many backups to previous snapshots. The system regularly gives kernel oopses, and just as regularly crashes (when it crashes, it generally shows almost exactly 300% or almost exactly400% cpu use in host top - maybe 398%, for example, but never 350% - strange...). We've installed exactly the same guest (identical .iso) on another, debian64 AMD, single NIC bridged host with absolutely no problems. So I'm assuming that the problem stems from either:
.CentOS host
.Intel multicore
.dual NIC bridged networking configuration (which was a devil to get working, and I'm not sure it'll survive a restart of the host)
or some combination of those with some other aspect of what we are doing.

Is anyone else encountering problems with any of these (I'm talking severe problems, crashes every few minutes, not a crash a day)? It might help to isolate where the problem is if we could see any commonalty in the systems that aren't working.

Vbox 3.0 looks like a great system, and a great achievement to get SMP working virtualised so nicely - the AMD machine is running like a dream. I just wish the intel machine would behave... :? And just in case it might make a difference, is there any news on when the next release might be coming out?

Thanks for listening...
Bob

Re: Help - incessant kernel oops, freezing etc.

Posted: 25. Aug 2009, 17:35
by urilabob
Sort-of-solved. Exactly the same installation, but with only one cpu in guest, runs perfectly. So it's presumably an interaction between SMP and intel cpus (intel X5472) and perhaps other aspects of hardware....