On Mon, May 28, 2007 at 10:53:40PM +0100, Andrew Doran wrote:
> > The problem is however similar. The system works perfectly stable (weeks
> > of uptime) if I leave the RAIDframe mirror in its current state (one
> > component failed). As soon as I try to rewrite the parity the system
> > freezes in less than an hour and needs to be resetted.
> 
> Just as a datapoint I have a system with a RAID 1 set that has been running
> 24x7 for the last 2.5 years, without a panic or any noticed data corruption
> problems.

I've used RAIDframe without problems on other hardware, too. I've even
used on this hardware without problems when I was still using it as my
server (it's now a friend's server).

But in this particular case it just doesn't want to work:
- If I try to at the second disk to the RAIDframe mirror the machine freezes
  (no crash, it just hangs) within a few minutes.
- If I try to simulate the parity rewrite e.g. by copying data from the
  first to the second harddisk with "dd" nothing happens.
- The machine survives high I/O load (e.g. a "cvs update") without
  any problems.

There are still things we want to try to make sure it is not a problem
with the hardware:
- change the IDE cables
- move the disk from the secondary onboard IDE channel to the secondary
  channel of the PCI IDE hostadapter
But it's somewhat suspicious that somebody else has similar problems
on different hardware.

	Kind regards

-- 
Matthias Scheler                                  http://zhadum.org.uk/