| 1 |
On Donnerstag 19 Februar 2009, Arnau Bria wrote: |
| 2 |
> Hi all, |
| 3 |
> |
| 4 |
> after a reboot (my machine was up for more than 6 months) I saw my raid |
| 5 |
> with only one of two devices used, and I had to read the second |
| 6 |
> partioition: |
| 7 |
> |
| 8 |
> #mdadm --manage /dev/md1 --add /dev/hda1 |
| 9 |
> mdadm: re-added /dev/hda1 |
| 10 |
> |
| 11 |
> # cat /proc/mdstat |
| 12 |
> Personalities : [raid1] [multipath] [faulty] |
| 13 |
> md1 : active raid1 hda1[2] hdc1[1] |
| 14 |
> 29302464 blocks [2/1] [_U] |
| 15 |
> [>....................] recovery = 2.8% (846976/29302464) |
| 16 |
> finish=19.9min speed=23731K/sec |
| 17 |
> |
| 18 |
> |
| 19 |
> Looking at dmes I saw this message: |
| 20 |
> |
| 21 |
> |
| 22 |
> raid1: raid set md2 active with 1 out of 2 mirrors |
| 23 |
> md: considering hdc1 ... |
| 24 |
> md: adding hdc1 ... |
| 25 |
> md: adding hda1 ... |
| 26 |
> md: created md1 |
| 27 |
> md: bind<hda1> |
| 28 |
> md: bind<hdc1> |
| 29 |
> md: running: <hdc1><hda1> |
| 30 |
> md: kicking non-fresh hda1 from array! |
| 31 |
> md: unbind<hda1> |
| 32 |
> md: export_rdev(hda1) |
| 33 |
> raid1: raid set md1 active with 1 out of 2 mirrors |
| 34 |
> md: ... autorun DONE. |
| 35 |
> |
| 36 |
> so, seems that hda1 was out of sync... is this the meaning? |
| 37 |
> How could it happen? cause mdadm did not send any email about my raid |
| 38 |
> status.... |
| 39 |
|
| 40 |
maybe your last shutdown was not really clean - and the disk had not enough |
| 41 |
time to flush its cache completly. Possible explaination: hd with big cache, |
| 42 |
ext3, a bit too fast to turn off at shutdown. |
Archives