Re: bogus SCSI timeouts and ntpd...

On Sat, 4 Aug 2001, Troy Benjegerdes wrote:

I have recently started getting what appear to be 'bogus' SCSI timeouts
after a crash and reboot, when my raid1 is re-syncing. (which amounts to
copying disk1 of the raid to disk2 of the raid.)

This didn't start happening until I started running ntpd at bootup..

I suspect something is broken somewhere such that when ntp adjusts the
time by a small amount, the SCSI layer (and generic timer code??) gets
confused and thinks that SCSI command have timed out.

<snip etc.>

I mailed the list a while ago about the ncr53c8xx SCSI driver hanging the
system with SCSI timeouts a while ago on a Motorola PowerStack with the
2.4.x kernels.  I am running xntpd at boot time.  I can (will) try
disabling it and seeing if a 2.4.x kernel stays up for a longer period of
time (typically less than 6 hours previously).

Jeff Rugen


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/