one of the units of a software RAID 1 configuration is shown as appears to be part of a raid array and the RAID status is active, degrated. I would like to know how I can rejoin the unit to the RAID, all related options are grayed out and I can't find anything in the CLI. Usually under linux I would do mdadm -a and thats all. So I'm a little bit confused here.
More Details: 2 servers, identical hardware configured as iscsi storage with failover/volume replication. Each server has three disks one for system, two in raid1 configuration exported via iscsi in block I/O mode as shared XEN storage. This is all the servers do.
What happend: we've tested a siren which is part of our fire protection system. A few seconds after the siren stopped, we noticed problems concerning our virtual machines and realized that the storage got problems on BOTH servers: md: super_written gets error=-5, uptodate=0 We initiated a failover and restarted the primary server, did failback and restarted the secondary. XEN was now able to access the storage again but the RAID on both servers is now broken as stated above. While rebooting the servers, we got a message that it could be a good idea to do a filesystem check on the raid disks. I thought to better not to do that because I feared even more trouble because of block I/O. Everything else seems to be ok.
There is no hot spare drive, only 2 hdd as raid1 unit md0. The md0 unit is active but degraded because one of the two drives is no raid member anymore but the drive is detected as "part of S/W Raid". I can't change anything in the raid configuration at all, all fields are grayed out.
On a usual linux system, I would use a couple of mdadm commands to bring the drive back and resync, no big deal. But it seems to me that it's not that easy with DSS.
Is it at least possible to shutdown, add a clean drive and add this one as new member to the existing md0 unit? Or do I have to recreate everything from scratch?