tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Problems with hangs under NetBSD-5.x

Brian Buhrow writes:
>       Hello.  I've been able to track down my problem a little further.
> Here  is what I know:
> 1.  Server: I386

ditto here.

> 2.  Version NetBSD-5.0-stable as of July 11, 2009

5.0.1 kernel, 3.0 userland.

> 3.  FFS options: softdep, quota.

None of softdep, quota, or logging.  RAIDframe is used.  amd is also 
involved (one of the last messages before one of the hangs was that 
the pid237@foo:/tmp_mnt had become unavailable...)

> 4.  The machine doesn't hang, but any processes which need to write to the
> filesystem with quotas enabled get stuck in tstile after a while. 

I've had a box get stuck in tstile almost immediately after a reboot 
(it was likely processing a bunch of mail at the time).  (after 330+ 
days of uptime with a 3.0 kernel, the box died after 4 days with 
out-of-mbuf issues.  2 days later it died from tstile issues (I think 
-- NMBCLUSTERS was bumped to 4096) and then 2 minutes after that 
death it died again with known tstile issues.  At that point it's 
been reverted back to 3.0.

(The box is an IBM x336 with dual CPUs (that show up as duals in 3.0, 
but as hyperthreaded CPUs in 5.0.1), 1GB RAM, and a couple of 73GB 
disks in a RAIDframe RAID 1 config.)


Greg Oster

Home | Main Index | Thread Index | Old Index