[opensuse] Setting up raid1 device not working
Hi all, Sometime in the last few days, swap has stopped working (it was on a raid1 device /dev/md3). Three other raid1 devices seem to be working normally. I do not know what happened, but as it was swap nothing was lost. I am trying to reset this up, but am running into dificulties. As a starting point, I deleted the entry for /dev/md3 in /etc/mdadm.conf and rebooted (I could not find any other way to stop /dev/md3). cat /proc/mdstat showed only the three working raid1 devices. Next, using the YaST Partitioner I create a new raid1 device and add /dev/sba6 and /dev/sdb6 (the same partitions used orignally) to the raid, format the raid device as swap and set the mount point to swap (this is what I did at initial installationas well). This now fails with an obscure error message: ------------ Failure occured during follwing action: Formatting software RAID /dev/md3 (2.0 GB) with swap System error code was: -3004 ------------ At this point cat /proc/mdstat shows for the mew raid device md3 : active (auto-read-only) raid1 sda6[1] sdb6[0] 2104468 blocks super 1.0 [2/2] [UU] resync=PENDING bitmap: 9/9 pages [36KB], 128KB chunk Looks like it needs to be resynced(?) but I don't understand. Isn't this a new raid device, and wouldn't the formatting as swap initialize everything to a correct state? At the same time, the output from /var/log/messages at this time is (sorry for the wrapping) Jul 29 11:34:00 raboud-office2 kernel: Uniform Multi-Platform E-IDE driver Jul 29 11:34:00 raboud-office2 kernel: ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx Jul 29 11:34:00 raboud-office2 kernel: st: Version 20080221, fixed bufsize 32768, s/g segs 256 Jul 29 11:34:00 raboud-office2 kernel: Driver 'st' needs updating - please use bus_type methodsroot 4221 0.0 0.0 0 0 ? S< 11:34 0:00 [md3_raid1] Jul 29 11:34:41 raboud-office2 kernel: md: bind<sdb6> Jul 29 11:34:41 raboud-office2 kernel: md: bind<sda6> Jul 29 11:34:41 raboud-office2 kernel: md: md3: raid array is not clean -- starting background reconstruction Jul 29 11:34:41 raboud-office2 kernel: raid1: raid set md3 active with 2 out of 2 mirrors Jul 29 11:34:41 raboud-office2 kernel: md3: bitmap initialized from disk: read 1/1 pages, set 16442 bits Jul 29 11:34:41 raboud-office2 kernel: created bitmap (9 pages) for device md3 It is now over two hours later and nothing has changed. There is a process running that seems related but not promising (no memory, no cpu time). root 4221 0.0 0.0 0 0 ? S< 11:34 0:00 [md3_raid1] I am lost at this point. I would try to do this outside of Yast (just in case it is a bug), but I don't know the right steps. Unfortunately I am not a raid expert. I would appreciate any insight on this. -- Don -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Wed, 30 Jul 2008 05:42:25 Don Raboud wrote:
Hi all,
Sometime in the last few days, swap has stopped working (it was on a raid1 device /dev/md3). Three other raid1 devices seem to be working normally. I do not know what happened, but as it was swap nothing was lost. I am trying to reset this up, but am running into dificulties.
Don, Please run mdadm --detail /dev/md3 (must be run as root) and post it the results here. Regards, -- =================================================== Rodney Baker VK5ZTV rodney.baker@iinet.net.au =================================================== As Zeus said to Narcissus, "Watch yourself."
On Tuesday 29 July 2008 03:17:08 pm Rodney Baker wrote:
Please run mdadm --detail /dev/md3 (must be run as root) and post it the results here.
[Rodney - sorry for the previous pm]. /dev/md3: Version : 01.00.03 Creation Time : Tue Jul 29 11:34:41 2008 Raid Level : raid1 Used Dev Size : 2104468 (2.01 GiB 2.15 GB) Raid Devices : 2 Total Devices : 2 Preferred Minor : 3 Persistence : Superblock is persistent Intent Bitmap : Internal Update Time : Tue Jul 29 11:34:41 2008 State : active, Not Started Active Devices : 2 Working Devices : 2 Failed Devices : 0 Spare Devices : 0 Name : 3 UUID : 7c165c89:72824d9e:34b0f802:89a3cf1a Events : 0 Number Major Minor RaidDevice State 0 8 22 0 active sync /dev/sdb6 1 8 6 1 active sync /dev/sda6 -- Don -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Wed, 30 Jul 2008 07:02:45 Don Raboud wrote:
On Tuesday 29 July 2008 03:17:08 pm Rodney Baker wrote:
Please run mdadm --detail /dev/md3 (must be run as root) and post it the results here.
[Rodney - sorry for the previous pm].
No probs.
/dev/md3: Version : 01.00.03 Creation Time : Tue Jul 29 11:34:41 2008 Raid Level : raid1 Used Dev Size : 2104468 (2.01 GiB 2.15 GB) Raid Devices : 2 Total Devices : 2 Preferred Minor : 3 Persistence : Superblock is persistent
Intent Bitmap : Internal
Update Time : Tue Jul 29 11:34:41 2008 State : active, Not Started Active Devices : 2 Working Devices : 2 Failed Devices : 0 Spare Devices : 0
Name : 3 UUID : 7c165c89:72824d9e:34b0f802:89a3cf1a Events : 0
Number Major Minor RaidDevice State 0 8 22 0 active sync /dev/sdb6 1 8 6 1 active sync /dev/sda6
-- Don
It looks to me like the array has been created OK but has not been started. Try running mdadm --run /dev/md3, then mkswap /dev/md3. If it successfully creates a swap filesystem on the array after starting it, then you should be able to do swapon /dev/md3. There are others on the list who have much more experience than I with software raid, so they may have a better idea but that is what I'd be trying for starters. Regards, -- =================================================== Rodney Baker VK5ZTV rodney.baker@iinet.net.au =================================================== We gave you an atomic bomb, what do you want, mermaids? -- I. I. Rabi to the Atomic Energy Commission
On Wednesday 30 July 2008 05:33:22 am Rodney Baker wrote:
It looks to me like the array has been created OK but has not been started. Try running mdadm --run /dev/md3, then mkswap /dev/md3. If it successfully creates a swap filesystem on the array after starting it, then you should be able to do swapon /dev/md3.
First of all, cat /proc/mdstat still shows a resync pending (nearly 20 hours have elapsed). md3 : active (auto-read-only) raid1 sda6[1] sdb6[0] 2104468 blocks super 1.0 [2/2] [UU] resync=PENDING bitmap: 9/9 pages [36KB], 128KB chunk Then # mdadm --run /dev/md3 mdadm: failed to run array /dev/md3: Device or resource busy However, now (10 seconds later) cat /proc/mdstat shows md3 : active raid1 sda6[1] sdb6[0] 2104468 blocks super 1.0 [2/2] [UU] bitmap: 0/9 pages [0KB], 128KB chunk No more resync pending. OK.... Try to make a swap # mkswap /dev/md3 mkswap: error: swap area needs to be at least 40kB Usage: mkswap [-c] [-v0|-v1] [-pPAGESZ] [-L label] /dev/name [blocks] Doesn't work because it is not running. # mdadm --detail /dev/md3 [...] State : active, Not Started [...] However I can't 'run' it (same as starting?) # mdadm --run /dev/md3 mdadm: failed to run array /dev/md3: Device or resource busy Stopping has no effect either. # mdadm --stop /dev/md3 mdadm: fail to stop array /dev/md3: Device or resource busy What is it busy doing? lsof shows nothing using /dev/md3 /dev/sda6 /dev/sdb6
There are others on the list who have much more experience than I with software raid, so they may have a better idea but that is what I'd be trying for starters.
Thanks for your help so far. -- Don -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (3)
-
Don Raboud
-
Don Raboud
-
Rodney Baker