Date: Sun, 14 Jul 2002 15:56:44 +0100 From: ian j hart <ianjhart@ntlworld.com> To: "stable@FreeBSD.ORG" <stable@FreeBSD.ORG> Cc: SXren Schmidt <sos@freebsd.dk> Subject: Re: ATA tags and stuff Message-ID: <3D31912C.210CFEF1@ntlworld.com> References: <3D2F6DC6.A5392FF7@ntlworld.com>
next in thread | previous in thread | raw e-mail | index | archive | help
ian j hart wrote: > > I can reproduce the "tags problem", RELENG_4 built today. > Also noticed some weird secondary IDE stuff. > > Are there any ATA patches I don't know about? > Is it worth trying CURRENT? > > Will do anything (legal) to help solve the problem. > I've borrowed the box for the weekend, so fire away. > > Three IBM-DTLA-307075 rescued from backup duties plus > one new IC35L080AVVA07-0. > > [shown with tags off here] > ad0: 78533MB <IC35L080AVVA07-0> [159560/16/63] at ata0-master UDMA100 > ad1: 73308MB <IBM-DTLA-307075> [148945/16/63] at ata0-slave UDMA100 > > ad2: 73308MB <IBM-DTLA-307075> [148945/16/63] at ata1-master UDMA100 > ad3: 73308MB <IBM-DTLA-307075> [148945/16/63] at ata1-slave UDMA100 > > First tried a LEX BN630LT M/B. Installed 4.6-RELEASE, upgraded > to STABLE. Set up vinum RAID10 volume. While formating > I had systat -vmstat running. Noticed that while the "load" > appeared the same on all four disks, ad0 and ad1 were at ~5% > busy, ad2 and ad3 were at ~50% busy. > > I posted somthing similar before (20010807). I've swapped > the cables around, same result. Tried different cables, same > result. All cables are spec length, don't have any shorter. > > This is a new board - could be a design fault. So I > tried a recent gigabyte board, same result. > > Set tags on. As soon as I load the vinum volume I get > > [snip time /kernel] > ad2: READ command timeout tag=0 serv=0 - resetting > ad2: invalidating queued requests > ad2: timeout waiting for cmd=00 s=d0 e=04 > ad2: flush queue failed > ata1: resetting devices .. ad2: invalidating queued requests > ad2: DMA limited to UDMA33, non-ATA66 cable or device > ad3: invalidating queued requests > done > ad2: no request for tag=0 > ad2: invalidating queued requests > > So far only ad2 and ad3 fail. > > A thought occurs. There could be a master/slave problem > with the firmware in the DTLA drives, which is fixed > in the IC35. I'll swap the IDE cables on the M/B and see what > happens. > > Maybe I'd better sleep first, I'm rambling. > > BTW, this is NOT a vinum problem. I get the same problem > on normal volumes (but only on ad2 ad3). This is just a *neat* > way of loading the drives. Pass the pointy hat. I'd used the IBM feature tool to disable the write cache, but only on two of the three *old* drives. Enabling wc fixes the systat load reading and the tags error. The worse thing is there was a post with all the clues in (on wednesday), and I still didn't get it. -- ian j hart To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3D31912C.210CFEF1>