HP EVA Disk ungrouping

[S R ARUN KUMAR]

Hi,

We recently had a proactive disk replacement alert in EVA 8000 and hence we ungrouped the impacted disk.(As part of procedure to replace the disk)

Immediately after ungrouping the disk, leveling started and it took 2 days to complete fully and Once it reaches 100% again leveling started in the DG.(2nd leveling is also too slow, assuming it will take another two days to complete)

Can any one explain why two leveling is happening? Is it ok to replace the disk?

NOTE : No other simultaneous disk failures in this EVA.

[Torsten.]

Leveling is "balancing" the disks, No problem to replace the disk, the leveling will be suspended.

[Johan Guldmyr]

Yeup, replacing while leveling is not a problem. But, do try to avoid replacing disks when reconstruction is going on.

[SUBHAJIT KHANBARMAN_1]

Hi Arun,

The amount of time it takes to level as a result of removing/adding disks in a diskgroup is directly proportional to the number of disks in a diskgroup and of course the size of the disks.


Example
-------------
1TB FATA drive removal/addition in a 240 disk DG will take days.
On a 120 disk DG in an EVA8000 with 450GB FATA drives - it averages array 4 days.

If we lets say expand with 5 drives, leveling takes about a week. And new space is not available before 1-2 days

space sometimes would be increasingly available as leveling progresses.


I would suggest you to send me the Controller Event Log and Configuration log of your EVA....I will look into it and suggest you if anything wrong.

[krusty]

One other item that affect leveling is load. If the array is really busy, ii services I/O from the hosts first - leveling is a "background" process. This is also a reason it can take so long.

Cheers,

Curt