Date: Sat, 1 Apr 2000 09:11:40 +0930 From: Greg Lehey <grog@lemis.com> To: Soren Schmidt <sos@freebsd.dk> Cc: vallo@matti.ee, freebsd-current@FreeBSD.ORG Subject: Re: Deadlock with vinum raid5 Message-ID: <20000401091140.A51727@freebie.lemis.com> In-Reply-To: <200003311932.VAA70873@freebsd.dk> References: <20000331212544.A59295@myhakas.matti.ee> <200003311932.VAA70873@freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
On Friday, 31 March 2000 at 21:32:22 +0200, Søren Schmidt wrote: > It seems Vallo Kallaste wrote: >> Hi >> >> I have four (4) ATA disks, all same new 20GB IBM 7200rpm models. Intel >> Seattle BX2 mobo, PIIX4 controller. I'm not expecting any performance >> increase or such, simply want to get my hands on. Ata driver, as the >> -current doesn't have wd anymore. >> I have raid5 volume laid over four disks, containing /usr filesystem, >> stripe unit size 256k. Machine deadlocks very reliably while doing >> ordinary buildworld, keyboard is responsive, network also but no disk >> I/O happens. I see syncer and bufdaemon in vrlock state; make, ln, >> syslogd and dhclient in flswait state. I usually don't see such states. >> Sources from which I built the system are todays, before I set up vinum. > > Yup, Greg and I know of this problem, it also happens with the wd > driver and with CAM, so the problem is probably not the driver. > I'm investigating this right now, and it looks like vinum is > botching a struct buf in action somehow. I'm not sure that this is the same problem. Please supply the information I ask for in http://www.lemis.com/vinum/how-to-debug.html. The problem that Søren and I are looking at is usually a panic. We don't really know where it's happening, but we're each sure it's not in *our* code :-) From a Vinum standpoint, it happens between the time that Vinum sends a request to the driver and when the I/O completes, so it's difficult to blame Vinum. On the other hand, we've seen it with SCSI as well, so it's difficult to blame the driver. I'm half guessing that it's something else altogether which is spamming freed data. Vinum mallocs the buffer headers rather than using geteblk(), which could explain why it happens only with Vinum. Greg -- Finger grog@lemis.com for PGP public key See complete headers for address and phone numbers To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20000401091140.A51727>