Date: Fri, 12 Jan 2001 17:31:25 -0800 From: Joseph Filla <jfilla@mindmaker.com> To: Greg Lehey <grog@lemis.com> Cc: freebsd-questions@FreeBSD.ORG, freebsd-stable@FreeBSD.ORG Subject: Re: vinum create causes my 4.2-Stable SMP machine to reboot Message-ID: <3A5FAFED.6F305826@mindmaker.com> References: <3A5F9BFB.9C24C85F@mindmaker.com> <20010113110111.B66238@wantadilla.lemis.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Greg Lehey wrote: > > [Format recovered--see http://www.lemis.com/email/email-format.html] > > Please don't wrap log output. I've turned off wrapping of lines. Netscape 4.76 won't allow me to set outgoing message wrap to 0 characters. I've set it to 999. > > On Friday, 12 January 2001 at 16:06:19 -0800, Joseph Filla wrote: > > System information: > > 4.2-STABLE FreeBSD 4.2-STABLE #3: Wed Dec 6 17:46:52 PST 2000 > > SMP > > 2 Promise ATA66 controllers > > 4 IDE drives: > > ad4 > > ad6 > > ad8 > > ad10 > > All dangerously dedicated > > > > > > For the last few months, I have had vinum running in a raid 10 > > configuration: > > > > su-2.03# vinum list > > 4 drives: > > D a State: up Device /dev/ad4s1e Avail: 0/11500 MB (0%) > > D b State: up Device /dev/ad6s1e Avail: 0/11500 MB (0%) > > D c State: up Device /dev/ad8s1f Avail: 0/11500 MB (0%) > > D d State: up Device /dev/ad10s1f Avail: 0/11500 MB (0%) > > > > 4 volumes: > > V swap State: up Plexes: 2 Size: 649 MB > > V var State: up Plexes: 2 Size: 1999 MB > > V usr State: up Plexes: 2 Size: 1999 MB > > V home State: up Plexes: 2 Size: 17 GB > > > > <snip> > > > > After a few months of use on this box, I decided to use the remaining > > free space on the drives by creating another raid 10 volume. Through > > Sysinstall, I created a partition on each of the four drives, and > > changed the disklabels on all 4 new partitions to type vinum. > > > > I created another vinum config file vinum_home2.conf: > > drive e device /dev/ad4s1f > > drive f device /dev/ad6s1f > > drive g device /dev/ad8s1g > > drive h device /dev/ad10s1g > > > > volume home2 > > plex org striped 257k > > sd length 7705m drive e > > sd length 7705m drive f > > plex org striped 257k > > sd length 7705m drive g > > sd length 7705m drive h > > > > I know putting another volume on another partition in the same slice > > isn't optimal but I don't have any other drives. > > You really need to make that a single partition. But that shouldn't > be the problem. > > > I ran 'vinum create /etc/vinum_home2.conf' and the machine rebooted. > > Spontaneously, with no messages? I was on an ssh connection to the server both times. The first time, I got a few lines back but the second time nothing. By the time I walked over to the server, it had rebooted. > > > vinum_history doesn't show the the create command at all. > > No, it wouldn't. You had a crash. > > > The system rebooted into single user mode and I was eventually able > > to get back to a running system. vinum -l showed the new volume > > home2, the new plexes and the new subdisks but after running init > > and start, I could never get the subdisks past faulty. One of the > > striped plexes was always in state crashed. > > Where's the list? In the frustration of the moment, I didn't save a copy of the list. My bad. After reading some postings to -questions, I looked in /dev/vinum and there were no devices in drive, plex, sd etc for my new volumes. So I ran vinum -makedev and tried to init and start the new volumes. From /var/log/messages: Jan 9 17:04:07 hera /kernel: vinum: drive e is up Jan 9 17:04:07 hera /kernel: vinum: drive f is up Jan 9 17:04:07 hera /kernel: vinum: drive g is up Jan 9 17:04:07 hera /kernel: vinum: drive h is up Jan 9 17:04:07 hera /kernel: vinum: removing 80 blocks of partial stripe at the end o f home2.p0 Running makedev didn't solve my problem. I rebooted and here is /var/log/messages: Jan 9 17:57:10 hera /kernel: vinum: loaded Jan 9 17:57:10 hera /kernel: vinum: reading configuration from /dev/ad10s1f Jan 9 17:57:10 hera /kernel: vinum: home2.p0.s0 is crashed Jan 9 17:57:10 hera /kernel: vinum: home2.p0 is faulty Jan 9 17:57:10 hera /kernel: vinum: home2 is down Jan 9 17:57:10 hera /kernel: vinum: home2.p0.s1 is crashed Jan 9 17:57:10 hera /kernel: vinum: home2.p1.s0 is up Jan 9 17:57:10 hera /kernel: vinum: home2.p1 is up Jan 9 17:57:10 hera /kernel: vinum: home2 is up Jan 9 17:57:10 hera /kernel: vinum: home2.p1.s0 is up Jan 9 17:57:10 hera /kernel: vinum: home2.p1.s1 is up Jan 9 17:57:10 hera /kernel: vinum: updating configuration from /dev/ad8s1f Jan 9 17:57:10 hera /kernel: vinum: updating configuration from /dev/ad4s1e Jan 9 17:57:10 hera /kernel: vinum: updating configuration from /dev/ad6s1e Jan 9 17:57:10 hera /kernel: vinum: updating configuration from /dev/ad6s1f Jan 9 17:57:10 hera /kernel: vinum_scandisk: /dev/ad6s1f is down Jan 9 17:57:10 hera /kernel: vinum: Can't read device /dev/ad6s1f, error 5 Jan 9 17:57:10 hera /kernel: vinum: updating configuration from /dev/ad8s1g Jan 9 17:57:10 hera /kernel: vinum_scandisk: /dev/ad8s1g is down Jan 9 17:57:10 hera /kernel: vinum: Can't read device /dev/ad8s1g, error 5 Jan 9 17:57:10 hera /kernel: vinum: updating configuration from /dev/ad4s1f Jan 9 17:57:10 hera /kernel: vinum_scandisk: /dev/ad4s1f is down Jan 9 17:57:10 hera /kernel: vinum: Can't read device /dev/ad4s1f, error 5 Jan 9 17:57:10 hera /kernel: vinum: updating configuration from /dev/ad10s1g Jan 9 17:57:10 hera /kernel: vinum_scandisk: /dev/ad10s1g is down Jan 9 17:57:10 hera /kernel: vinum: Can't read device /dev/ad10s1g, error 5 Jan 9 17:57:10 hera /kernel: vinum: couldn't read configuration Similar results as I wrestled with it for a day before eventually doing a rm -f on it all > > > I eventually had to detach and rm all remnants of home2 and got my > > system back to the original state. Thinking I may have done > > something incorrectly, I retried the vinum create config and the > > system rebooted again. This time vinum_history recorded the > > following: > > > > 12 Jan 2001 14:05:27.960421 *** vinum started *** > > 12 Jan 2001 14:05:27.961106 create vinum_home2.conf > > The ten seconds here are hardly enough time for a reboot. Let's > intersperse these messages: > > > 12 Jan 2001 14:05:37.293978 *** vinum started *** > > 12 Jan 2001 14:12:43.507023 *** vinum started *** > > 12 Jan 2001 14:12:44.694607 l > > ... > > > > This time the system came backup correctly, however this time, running > > vinum -l show me nothing regarding my desired volume home2. Here is the > > relevant stuff from /var/log/messages: > > Jan 10 16:03:34 hera /kernel: vinum: loaded > > Jan 10 16:03:34 hera /kernel: vinum: reading configuration from /dev/ad8s1f > > Jan 10 16:03:34 hera /kernel: vinum: updating configuration from /dev/ad10s1f > > Jan 10 16:03:34 hera /kernel: vinum: updating configuration from /dev/ad4s1e > > Jan 10 16:03:34 hera /kernel: vinum: updating configuration from /dev/ad6s1e > > 12 Jan 2001 14:05:27.960421 *** vinum started *** > > 12 Jan 2001 14:05:27.961106 create vinum_home2.conf > > 12 Jan 2001 14:05:37.293978 *** vinum started *** > > Jan 12 14:05:37 hera /kernel: vinum: drive e is up > > Jan 12 14:05:37 hera /kernel: vinum: drive f is up > > Jan 12 14:05:37 hera /kernel: vinum: drive g is up > > Jan 12 14:05:37 hera /kernel: vinum: drive h is up > > Jan 12 14:10:37 hera /kernel: vinum: loaded > > Jan 12 14:10:37 hera /kernel: vinum: reading configuration from /dev/ad8s1f > > Jan 12 14:10:37 hera /kernel: vinum: updating configuration from /dev/ad10s1f > > Jan 12 14:10:37 hera /kernel: vinum: updating configuration from /dev/ad6s1e > > Jan 12 14:10:37 hera /kernel: vinum: updating configuration from /dev/ad4s1e > > Jan 12 14:10:37 hera /kernel: vinum: updating configuration from /dev/ad6s1f > > Jan 12 14:10:37 hera /kernel: vinum_scandisk: /dev/ad6s1f is down > > Jan 12 14:10:37 hera /kernel: vinum: Can't read device /dev/ad6s1f, error 5 > > Jan 12 14:10:37 hera /kernel: vinum: updating configuration from /dev/ad8s1g > > Jan 12 14:10:37 hera /kernel: vinum_scandisk: /dev/ad8s1g is down > > Jan 12 14:10:37 hera /kernel: vinum: Can't read device /dev/ad8s1g, error 5 > > Jan 12 14:10:37 hera /kernel: vinum: updating configuration from /dev/ad4s1f > > Jan 12 14:10:37 hera /kernel: vinum_scandisk: /dev/ad4s1f is down > > Jan 12 14:10:37 hera /kernel: vinum: Can't read device /dev/ad4s1f, error 5 > > Jan 12 14:10:37 hera /kernel: vinum: updating configuration from /dev/ad10s1g > > Jan 12 14:10:37 hera /kernel: vinum_scandisk: /dev/ad10s1g is down > > Jan 12 14:10:37 hera /kernel: vinum: Can't read device /dev/ad10s1g, error 5 > > Jan 12 14:10:37 hera /kernel: vinum: couldn't read configuration > > > > Neither reboots created any dumps in /var/crash > > Do you have dumps enabled? I do now... I can always run the create command again and catch it. > > The best I can guess here is that your partition layout is funny in > some way. Let's have a look at the disklabel output for each > spindle. Just the end, like this: > > 8 partitions: > # size offset fstype [fsize bsize bps/cpg] > c: 4124640 0 unused 0 0 # (Cyl. 0 - 1006*) > h: 4124640 0 vinum # (Cyl. 0 - 1006*) disklabel ad4: 8 partitions: # size offset fstype [fsize bsize bps/cpg] a: 253952 0 4.2BSD 0 0 0 # (Cyl. 0 - 403*) c: 40088160 0 unused 0 0 # (Cyl. 0 - 63631) e: 23552000 253952 vinum # (Cyl. 403*- 37787*) f: 16282208 23805952 vinum # (Cyl. 37787*- 63631*) disklabel ad6: 8 partitions: # size offset fstype [fsize bsize bps/cpg] b: 253952 0 swap # (Cyl. 0 - 403*) c: 40088160 0 unused 0 0 # (Cyl. 0 - 63631) e: 23552000 253952 vinum # (Cyl. 403*- 37787*) f: 16282208 23805952 vinum # (Cyl. 37787*- 63631*) disklabel ad8: 8 partitions: # size offset fstype [fsize bsize bps/cpg] c: 40088160 0 unused 0 0 # (Cyl. 0 - 63631) e: 251904 0 4.2BSD 0 0 0 # (Cyl. 0 - 399*) f: 23552000 251904 vinum # (Cyl. 399*- 37783*) g: 16284256 23803904 vinum # (Cyl. 37783*- 63631*) disklabel ad10: 8 partitions: # size offset fstype [fsize bsize bps/cpg] c: 40088160 0 unused 0 0 # (Cyl. 0 - 63631) e: 253952 0 4.2BSD 0 0 0 # (Cyl. 0 - 403*) f: 23552000 253952 vinum # (Cyl. 403*- 37787*) g: 16282208 23805952 vinum # (Cyl. 37787*- 63631*) Note how ad8s1g has a different offset and size from the other related partitions. I noticed that before running create and I tried to make my lengths accordingly. I didn't intentionally do it this way. > > Greg > -- > When replying to this message, please copy the original recipients. > If you don't, I may ignore the reply. > For more information, see http://www.lemis.com/questions.html > 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-questions" in the body of the message -- Joe Filla Systems Administrator Mindmaker, Inc. (408) 467-0468 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?3A5FAFED.6F305826>