Date: Tue, 13 Apr 1999 19:22:33 -0600 (MDT) From: "Kenneth D. Merry" <ken@plutotech.com> To: asami@FreeBSD.ORG (Satoshi - Ports Wraith - Asami) Cc: scsi@FreeBSD.ORG Subject: Re: timed out while idle? Message-ID: <199904140122.TAA06901@panzer.plutotech.com> In-Reply-To: <199904140059.RAA49792@silvia.hip.berkeley.edu> from Satoshi - Ports Wraith - Asami at "Apr 13, 1999 5:59:32 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
Satoshi - Ports Wraith - Asami wrote... > * From: "Kenneth D. Merry" <ken@plutotech.com> > > * > Wow, 60 seconds? That's indeed a pretty good lunch. :) > * > * Yep. > > Do you think it's a bad idea to reduce it to something like 15 > seconds? Not that it matters much but the disk reads pretty much > stall during that time, and if it's not going to catch any false > positives, I'd like them to recover faster. Is there some way to find > out if there's something that came back after 15 seconds but before > 60? The reason the timeout is as high is it is now is because of some devices (zip drives?) that spin down their media and then spin back up when they get a command. It takes them a little bit to spin up, so that's why the timeout is as long as it is. As far as figuring out whether something came back between 15 and 60 seconds, probably the only way to do that would be to timestamp each CCB as it is sent down to the controller. It can be done, but is it really worth it? > * I don't think the timeouts in and of themselves will cause panics, unless > * maybe the drive never recovers. > > How do I find out if the drive recovered? If it is still responding. In most cases that I've seen a BDR will cause the drive to wake up and start functioning again. > * I dunno what's going on there. It could be indirectly caused by the > * timeout, but I really don't know how that could happen. > > Next time I'll try to catch it with a debug kernel. That'll be more helpful. At least you'll be able to (hopefully) see exactly which line caused the panic. Ken -- Kenneth Merry ken@plutotech.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199904140122.TAA06901>