Date: Fri, 3 Sep 1999 23:33:26 -0400 From: David Krinsky <krinsky@bantha.org> To: Soren Schmidt <sos@freebsd.dk> Cc: freebsd-hackers@FreeBSD.ORG Subject: Re: Seagate STT8000A (ATAPI/IDE) on FreeBSD? (fwd) Message-ID: <19990903233326.A346@bantha.org> In-Reply-To: <199909030647.IAA88473@freebsd.dk>; from Soren Schmidt on Fri, Sep 03, 1999 at 08:47:54AM %2B0200 References: <19990903020859.A33016@bantha.org> <199909030647.IAA88473@freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Sep 03, 1999 at 08:47:54AM +0200, Soren Schmidt (sos@freebsd.dk) wrote: > > wst_done: wst0: nonrecovered data error > > I've seen this problem LOTS of times when using the old wd based > atapi subsystem. I've never been able to find out why this is > happening exactly. This was part of the reason I started out > on the new ATA driver (only in -current now). I've never had > this problem using the ATA driver, so I'm pretty sure its the > old driver thats at fault, probably some delicate timing prob. > Using the new driver I do routine backups every night on > a couble of servers, not seen a signle problem yet... Thanks for the help, but I'm afraid I may have to burst your bubble. :-/ Upon your advice I cvsupped to today's -current and built an appropriate kernel with the new ATA driver. I now get one of the two following error scenarios (which one is unpredictable, though usually the former): Sep 3 22:25:17 bantha /kernel: atapi_error: TAPE_WRITE - MEDIUM ERROR skey=3 asc=31 ascq=00 error=00 Sep 3 22:25:17 bantha /kernel: atapi_error: REZERO_UNIT/TAPE_REWIND - MEDIUM ERROR skey=3 asc=31 ascq=00 error=00 or Sep 3 22:34:15 bantha /kernel: atapi_error: TAPE_WRITE - timeout error = 00 Sep 3 22:34:15 bantha /kernel: atapi_transfer: bad command phase Sep 3 22:34:15 bantha /kernel: ad0: status=51 error=04 Sep 3 22:34:15 bantha /kernel: ad_interrupt: hard error Sep 3 22:34:32 bantha /kernel: atapi_interrupt: unknown transfer phase 11 Sep 3 22:34:45 bantha /kernel: atapi_error: TAPE_WEOF - timeout error = 00 Sep 3 22:35:15 bantha /kernel: atapi_error: TAPE_WEOF - timeout error = 00 The former simply returns EIO; the latter appears to sleep uninterruptibly on atprq. The apparent inconsistency in which happens when supports the notion that this is, as you say, a subtle timing issue. In an effort to figure out what was going on, I tried compiling atapi-all.c and atapi-tape.c with ATAPI_DEBUG set, but this only caused the kernel to eventually page fault and panic. :-/ Any ideas? I'm quite happy to help debug this one, if you have any thoughts on where to go with it--I need to get backups working, and I'd much rather not have to buy a new tape drive or switch OSes... Dave. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990903233326.A346>