I have a PS6000 that shows a warning of degraded. When I look at the Group manager, all disks show online and status healthy. I logged into the CLI and ran the diskview -j to see the status of the disks.
which drive do you think would be causing the degraded status? Could it be drive 9 ? GUI shows all are working and spares are fine. Could it be a false warning?
There's no command to force a spare into the raidset. Before you attempted to move it out of the pool will will involve a lot of IO activity on that member, do you have a backup?
If so, or when you do. You could during a quiet IO period try to fail the controller over and see if the passive controller when it picks up the drives will see the need to rebuild.
If you do decide to migrate the array out of the pool, and if you have critical volumes that will fit on ONE member you can BIND volumes to individual members. it's slower but would prioritize your volumes vs. moving the member to another pool will not allow that.
Also do not DELETE that member. Create a new pool, and move that member to that new pool. Once it's there, THEN you can delete it. When you add it back into the group, please don't use RAID5 anymore. Use RAID6.
What is the status of the controller batteries? It could be drive 9, but it doesn’t seem like it is that many errors. Is the drive and array firmware up to date?
There should be 16 drives, 0 to 15. Only 15 drives show in the RAIDset. The "f" indicates one drive has failed. By process of elimination, it's drive 2 which now shows as a spare has failed out. What version of firmware is that array running? I suspect it's older. The other spare should have kicked in, unless it too at some point failed and has a history of failure tag on it.
I would pull drive 2 and 4. Wait 3 minutes and put drive 4 back in. See if a rebuild starts.
Thanks Don and Josh for the responses. The battery shows good. We are not running the latest on the firmware.
The last operation we did was to replace the failed drive 4. Drive 4 had failed a short time ago. We put in a refurbished drive and added it to the array as a spare. Does it change the situation and steps recommended.
Yes that firmware was released Sept 2015. I would suggest removing the two spares 2 and 4. Wait two minutes and then reinstall drive FOUR (4). See if a rebuild starts. Then I would replace drive 2 with another qualified EQL drive.
I would also not use this array heavily and make sure you have a backup. Another drive failure and the array will be offline.
I read another post that said that adding a foreign drive to the SAN will only allow to add as a spare. They had to force the drive back into the raidset. Do you know how to this?
dwilliam62
6 Operator
•
1516 Posts
•
6374 Points
2382
1
Posted September 9th, 2021 02:00
There's no command to force a spare into the raidset. Before you attempted to move it out of the pool will will involve a lot of IO activity on that member, do you have a backup?
If so, or when you do. You could during a quiet IO period try to fail the controller over and see if the passive controller when it picks up the drives will see the need to rebuild.
If you do decide to migrate the array out of the pool, and if you have critical volumes that will fit on ONE member you can BIND volumes to individual members. it's slower but would prioritize your volumes vs. moving the member to another pool will not allow that.
Also do not DELETE that member. Create a new pool, and move that member to that new pool. Once it's there, THEN you can delete it. When you add it back into the group, please don't use RAID5 anymore. Use RAID6.
Regards,
Don