Matthias Scheler writes:
> On Fri, Jun 01, 2007 at 11:02:07AM -0600, Greg Oster wrote:
> > You say the machine freezes -- it does that when you attempt to do a 
> > 'raidctl -F' or 'raidctl -R', yes?  Can you get into ddb at that 
> > point?  Can you ping the box at that point?  My next guess is that 
> > it's a kernel memory issue of some sort... 
> 
> And it seem you were right about that. After I increased the available
> kernel memory from 128MB (default for a 1GB machine) to 256MB I could
> rebuild the RAID without the machine getting stuck:
> 
> raid0: Reconstruction of disk at col 0 completed
> raid0: Recon time was 7976.858062 seconds, accumulated XOR time was 0 us (0.0
> 00000)
> raid0:  (start time 1180943157 sec 172287 usec, end time 1180951134 sec 30349
>  usec)
> raid0: Total head-sep stall count was 0
> raid0: 6678754 recon event waits, 4 recon delays
> raid0: 3244449136 max exec ticks
> 
> The machine is still up and running.

Interesting!!!  So I wonder if it's a case where RAIDframe gets its 
memory, and then something else runs out and is unhappy, or if 
RAIDframe is running out (for a rebuild, it should be willing to wait 
for memory to come free)...

Later...

Greg Oster