Hmmm.... Seti@Home and a database import are probably *very* different
workloads from the perspective of the PERC controller. The database
load is going to be *very* I/O intensive (and would have a very good
chance of tickling bugs in a device driver) whereas I would expect
Seti@Home to be CPU-intensive, doing little if any I/O. (Unless
SETI@Home was making your system page hard...) How much I/O do you
figure Seti@Home was driving?
Were you experiencing "restarts", or "panics"? Most UNIXen can be
configured to dump an image of the kernel at the time of a panic and
preserve it after a restart. I *presume* Linux has the same ability,
although like other UNIXen, it may not be default behaviour. In any
event, if you *can* capture a kernel dump, it *could* go a long way to
diagnosing and fixing the problem. Providing you can find somebody who
*wants* to, that is. :-(
Oh well, thanks for the report...
Janine Sisk wrote:
In our case the system was spontaneously restarting, while under moderately heavy load like running Seti@Home or loading a multi-GB Oracle export. As far as I can recall it never restarted while just sitting there idle.
-- //www.freelists.org/webpage/oracle-l