EVA4400 will not complete Leveling, caps at 94%

[Lithium123]

Hi

Hope someone can help me. I've got an EVA4400 and unfortunately has 1 drive go into attention state. The space available unfortunately has been maxed out and I've just upgrading to a new SAN to handle our capacity needs next week.

Since the capacity is maxed I could not ungroup the attentioned drive to replace. While it took me a while to find away to increase the available space on the SAN, 2 other drives had issues. 1 failed and another became unidentifiable in a matter of 1 evening.

The failed drive has ungrouped on it's own and I've since replaced it. Fortunately no data has been lost. I've been monitoring the leveling for almost 30 hours now and it's been hitting 94% and going back down to 69-85% to only hit 94% again and drop. Have I run out of space to complete the leveling? Does the unidentifiable drive still count in the array?

Any help will be much appreciated.

Thanks

[hvhari]

Hi,

Levelling is a internal process in EVA which distributes the data evenly to all disks in a diskgroup. This is a low priority task and can take few days on an EVA which has good host i/os always happening. You need not worry about this as it is not going to affect the data redundancy. Levelling will be triggered with every disk change (Eg. you have replaced a failed disk, or added a new disk to the diskgroup).

Please ensure that there are no other disks showing any failure symptoms. HP's Guided Troubleshooting can help you in checking disk issues. Please refer below link.

www.hp.com/support/hpgt

[Johan Guldmyr]

Hey, if you take a 'controller configuration dump' you can check each drive's 'occupancy'. If this is at 0 then the disk is probably not part of the disk group.

[Lithium123]

Hello, thanks for the reply, Johan. There are 2 drive in the 1st enclosure as unidentified, in bays 2 and 4. They don't appear in the group anymore however I've been told that because the 2 drives read as populated, they are actually still part of the group, eventhough not displayed within the group. I'm not sure who to believe but seeing that I've got 1 failed drive and 2 unidentified, I should have lost data but I haven't.

I tried the controller configuration dump but because they aren't in the group, I can't check the occupancy.

My question is, will I cause any harm pulling out these 2 drives to replace? Will I end up losing data?

Really appreciate the help.

[Johan Guldmyr]

Maybe the leveling is taking long time because a reconstruct is going on? In that case it could be very bad to pull out a disk. Levelling taking long time is not really an issue, or are you having any issues with the array?

Check with HP support?

[Lithium123]

The reconstruction had completed day before yesterday, and the leveling has been running for over 48 hours now. It's a weekend so there's nearly no load on the EVA.

HP recommended I pull both drives out and replace, but I've a consultant said that's insane as I will lose data, but just leave the drives in and let the leveling complete.

[Johan Guldmyr]

I have seen leveling take a week. ( I believe it was with 1TB FATA drives)

[Steven Clementi]

What is your Disk allocation level at for the disk group in question?

What about the protection level for that group?

If there is enough free disk space in the group, weird things tend to happen.  You should try to keep the allocation below 90%... some even recommend below 85%.