Topic: Raidrebuild restating

[donald]

Hi

I have a SME 7.3 (upgraded from 6.0.1). When i upgraded it i took one disk out of the raid so i had a backup if the upgrade failed. The upgrade went fine and i added a disk som i could build up the raid again.

The rebuild starts but it never finishes, it just restarts at som point.

[root@has ~]# cat /proc/mdstat
Personalities : [raid1]
md2 : active raid1 hdd2[1] hda3[0]
      262016 blocks [2/2] [UU]

md1 : active raid1 hdd3[2] hda2[0]
      79678272 blocks [2/1] [U_]
      [==>..................]  recovery = 10.2% (8156032/79678272) finish=468.9min speed=2539K/sec
md0 : active raid1 hdd1[1] hda1[0]
      102208 blocks [2/2] [UU]

unused devices: <none>

I dont know what to do to debug it. Any ideas ?

[Stefano]

hi...

open a shell during rebuilding process and give:

Code: [Select]

tail -f /var/log/messages | grep hd

it'll show you any error..

HTH
Ciao
Stefano

[pfloor]

Donald,

The rebuild will restart if you reboot before the rebuild is complete.

"finish=468.9min" (that's almost 8 hours). Are you rebooting the server before it finishes the rebuild? 

[donald]

I guess one of the drives is FUBAR

Apr 25 20:47:48 has kernel: hda: irq timeout: status=0xd0 { Busy }
Apr 25 20:47:48 has kernel: hda: irq timeout: status=0xd0 { Busy }
Apr 25 20:47:48 has kernel: hda: irq timeout: status=0xd0 { Busy }
Apr 25 20:47:48 has kernel: end_request: I/O error, dev hda, sector 39002208
Apr 25 20:47:48 has kernel: hda: status error: status=0x58 { DriveReady SeekComplete DataRequest }
Apr 25 20:47:48 has sshd(pam_unix)[3811]: session opened for user root by (uid=0)
Apr 25 20:47:48 has kernel: hda: drive not ready for command
Apr 25 20:47:48 has kernel: hda: irq timeout: status=0xd0 { Busy }
Apr 25 20:47:48 has kernel: hda: irq timeout: status=0xd0 { Busy }
Apr 25 20:47:49 has sshd(pam_unix)[3811]: session closed for user root
Apr 25 20:47:50 has kernel: hda: irq timeout: status=0xd0 { Busy }
Apr 25 20:47:50 has kernel: end_request: I/O error, dev hda, sector 39002216
Apr 25 20:47:50 has kernel: hda: status error: status=0x58 { DriveReady SeekComplete DataRequest }
Apr 25 20:47:50 has kernel: hda: drive not ready for command
Apr 25 20:47:50 has kernel: hda: irq timeout: status=0xd0 { Busy }
Apr 25 20:47:50 has kernel: raid1: hda: unrecoverable I/O read error for block 38797568
Apr 25 20:47:50 has kernel: hda: read_intr: status=0x51 { DriveReady SeekComplete Error }
Apr 25 20:47:50 has kernel: hda: read_intr: error=0x40 { UncorrectableError }, LBAsect=39002365, sector=39002365
Apr 25 20:47:50 has kernel: end_request: I/O error, dev hda, sector 39002365

Thanks for your help

[pfloor]

I guess one of the drives is FUBAR

Namely hda and it could be the ribbon (or even power) cable to that drive.  Have you tried a new (high quality) 80 wire ribbon cable?

Also, you have a strange setup:

hda <Primary drive on first IDE bus
hdd <Secondary drive on second IDE bus

Why did you set it up this way?  The preferred setup is hda and hdc.

[donald]

I will try to replace the cable.

I drive setup is strange because a non-it-guy replaced the server. Its about 100 miles away so