mdadm starts resync on every boot
- by Anteru
Since a few days (and I'm positive it started shortly before I updated my server from 13.04-13.10) my mdadm is resyncing on every boot. In the syslog, I get the following output
[ 0.809256] md: linear personality registered for level -1
[ 0.811412] md: multipath personality registered for level -4
[ 0.813153] md: raid0 personality registered for level 0
[ 0.815201] md: raid1 personality registered for level 1
[ 1.101517] md: raid6 personality registered for level 6
[ 1.101520] md: raid5 personality registered for level 5
[ 1.101522] md: raid4 personality registered for level 4
[ 1.106825] md: raid10 personality registered for level 10
[ 1.935882] md: bind<sdc1>
[ 1.943367] md: bind<sdb1>
[ 1.945199] md/raid1:md0: not clean -- starting background reconstruction
[ 1.945204] md/raid1:md0: active with 2 out of 2 mirrors
[ 1.945225] md0: detected capacity change from 0 to 2000396680192
[ 1.945351] md: resync of RAID array md0
[ 1.945357] md: minimum _guaranteed_ speed: 1000 KB/sec/disk.
[ 1.945359] md: using maximum available idle IO bandwidth (but not more than 200000 KB/sec) for resync.
[ 1.945362] md: using 128k window, over a total of 1953512383k.
[ 2.220468] md0: unknown partition table
I'm not sure what's up with that detected capacity change, looking at some old logs, this does have appeared earlier as well without a resync right afterwards. In fact, I let it run yesterday until completion and rebooted, and then it wouldn't resync, but today it does resync again. For instance, yesterday I got:
[ 1.872123] md: bind<sdc1>
[ 1.950946] md: bind<sdb1>
[ 1.952782] md/raid1:md0: active with 2 out of 2 mirrors
[ 1.952807] md0: detected capacity change from 0 to 2000396680192
[ 1.954598] md0: unknown partition table
So it seems to be a problem that the RAID array does not get marked as clean after every shutdown? How can I troubleshoot this? The disks themselves are both fine, SMART tells me no errors, everything ok.