Hi, I have a system that's meant as a 'backup disk spool' - it has a 2Gb SCSI drive to boot from, and then 2 x 16Gb IDE drives (IBM-DTTA-351680's).
The system works fine so long as I stick to using the SCSI drives... If I use the IDE's (e.g. backup another machine to them, tar / untar 'usr/src' to them) I start getting problems, e.g. Jan 14 13:01:42 magpie /kernel: wd0: interrupt timeout: Jan 14 13:01:42 magpie /kernel: wd0: status 58<rdy,seekdone,drq> error 0 Jan 15 17:56:52 magpie /kernel: lnc0: Missed packet -- no receive buffer Jan 15 17:57:43 magpie /kernel: lnc0: Missed packet -- no receive buffer The machine also has an NE-2100 embedded network card. We have 4 other identical machines to this, all SCSI based - all work without any problems... Is there anything that can be done? - at a guess I'd say the IDE's/wcd0 are timing out on interrupts, which is then stuffing the network (as the systems been waiting around too long for wcd0's IRQ's?)... The machine is a dual P90 (running SMP) - I get the problems even with a single CPU kernel... It's running Neptune chipset (quite old) - and only has 1 IDE channel. Any pointers? - The drives do appear to work OK, i.e. no corruption, but the _whole_ machine appears to 'stall' when the above occurs = not good... :-( I've tried 3.0-Release and 3.0-Current as of 7th Jan, '99 - Both exhibit identical symptoms... -Kp To Unsubscribe: send mail to majord...@freebsd.org with "unsubscribe freebsd-current" in the body of the message
