On Sun, Jan 14, 2007 at 08:05:23AM -0500, Stef Caunter wrote:
> Hope this is the appropriate list; there isn't much mention of this in 
> anything recent. A Dell 2800 runs fine but logs this during disk to disk 
> backups when a job first starts up (directory with around 50,000 files 
> gets backed up with tar and gzip). File operation succeeds. Later a much 
> larger directory is done with no log message:
> 
> Jan 14 03:16:15 mx1 /netbsd: sd0(mpt0:0:0:0): unable to allocate scsipi_xfer
> Jan 14 03:16:15 mx1 /netbsd: sd0(mpt0:0:0:0): unable to allocate scsipi_xfer
> Jan 14 03:16:15 mx1 last message repeated 339 times
> 
> The box does not seem to be affected; the error will occur on either drive 
> (there are only 2). Is this something that could be problematic?

The scsipi subsystem should be able to recover from this condition;
but this probably means that this system could benefit from some more
RAM for this workload ....

-- 
Manuel Bouyer <bouyer@antioche.eu.org>
     NetBSD: 26 ans d'experience feront toujours la difference
--