This would lead me to point to the drivers for, as well as the hardware
of, your RAID controller for possible problems.
The last time I came across similar issues, it has been, in one case, a
power supply unable to quite keep up, causing such flakiness when all the
drives were engaged, and in another case, RAM that ultimately tested bad.
I wouldn't necessarily rule out a bad kernel build, but BackupPC (and
other userspace programs) shouldn't be capable of crashing the kernel --
the key element of BackupPC is that it stresses a filesystem.
> Hi people,
> OK I've tried swapping out hardware. I found an identical server, known to
> be working, and moved the RAID controller and the five drives over to
> that. In a stroke that meant different memory, cpu, power supply and
> motherboard. The result though stays the same :-( Kernel panic within a
> few seconds of reaching the login prompt. Interactive boot with no
> services running seems to be fine.
> Has anyone on this list come across a similar issue? This is new to me so
> I'd appreciate any words of wisdom.
> Huw Wyn Jones
> Systems Administrator
> Coleg Meirion Dwyfor
> ----- Original Message -----
> From: "Huw Wyn Jones" <huw.jones@...>
> To: "General list for user discussion, questions and support"
> Sent: Tuesday, 26 January, 2010 16:30:47 GMT +00:00 GMT Britain, Ireland,
> Subject: Re: [BackupPC-users] ext3 filesystem limitations
>>> Any updates to packages recently that you can think of?
> None. Our network is like Fort Knox so once we have a box running reliably
> we tend to leave it alone. I'm pretty new to sysadmin'ing so I'm inclined
> to 'go-with-the-flow' at the moment.
> I also suspected hardware initially. However the RAID5 controller reports
> that all the disks are healthy, and (once the services have been switched
> off) it runs very reliably despite having rsync/scp copying files off it
> for most of the day. Similarly the processor is running at 20% load for
> the past 6 hours doing file transfers. :-/
> Still, I hear you all saying hardware! Best dig out the test disks once
> I've secured a copy of the data.
> Thanks for your input