Date: Wed, 20 Jan 1999 19:27:12 +0100 From: Bernd Walter <ticso@cicely.de> To: "Jeroen C. van Gelderen" <gelderen@mediaport.org>, "Kenneth D. Merry" <ken@plutotech.com>, Greg Lehey <grog@lemis.com> Cc: roberto@keltia.freenix.fr, freebsd-scsi@FreeBSD.ORG Subject: Re: Disk dying (Conner CFP and Tagged Queueing Probs) Message-ID: <19990120192712.40364@cicely.de> In-Reply-To: <010101be4467$02243480$0d79eb0a@deskfix.local>; from Jeroen C. van Gelderen on Wed, Jan 20, 1999 at 12:21:21PM %2B0100 References: <010101be4467$02243480$0d79eb0a@deskfix.local>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Jan 20, 1999 at 12:21:21PM +0100, Jeroen C. van Gelderen wrote: > From: Kenneth D. Merry <ken@plutotech.com> > > [...] > >Now we're getting somewhere. It may be that we can crank the tags down to > >8 instead of disabling tagged queueing altogether. > > Probably down to 30 because that's what I've been running with without > problems for about a year... > > I wrote: > > Question is: couldn't it be caused by the drive going bezerk after a > > SYNCHRONIZE CACHE?? I ask this because I don't recall having seen this > error > > before and this system has been running for well over a year... > > You responded: > >I doubt it. The cache sync happens in three places: > > > >- when the da device gets final close > >- when the machine is shut down, and a device hasn't seen final close, its > > cache gets flushed. > >- right after we've done a crash dump to a disk. > > > >The device close can happen at boot time, and when you reboot, but won't > >happen during normal operation. > > I rebooted that machine several times that day. About half of the time I got > syncing problems in addition to the SYNCACHE command error. I suggest you > crank down the tags to 30 (or 8 but I suggest the larger because that helps > narrowing down the problem) and add the SYNC CACHE quirk. My machine does a > daily buildworld so I hope I'll catch the problems... > SYNCCACHE shouldn't be a problem. I have lots of drives rejecting it - it's more an informational thing. It's only critical on some special drives. I personaly own the following: Jan 8 21:44:06 cicely6 /kernel: sd0 at scbus0 target 0 lun 0 Jan 8 21:44:06 cicely6 /kernel: sd0: <FUJITSU M1606S-512 6406> type 0 fixed SCSI 2 Jan 8 21:44:07 cicely6 /kernel: sd0: Direct-Access 1041MB (2131992 512 byte sectors) The drive needs a power down after SYNCCACHE. > > Cheers, > Jeroen > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-scsi" in the body of the message -- B.Walter 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?19990120192712.40364>