VRAID1 offline when a Hard Disk fail. Why?

[ngo son]

I have a EVA 4000 disk array with configuration:
- 1 enclosure with full 14 disk 300GB.
- all 14 disk go to 1 disk group DG1, protection level = 0.
- inside DG1, I create 4 RAID1 vdisk (vd1, vd2, vd3, vd4) that utilize nearly 100% capacity.

Last weekend, one of my physical disk failed and then I saw that:
- vd1 and vd2 still online and visible.
- vd3 vd4 go offline.
- the free space of DG1 = 0.

When 2 click to vd3, vd4, it said sth like that I have two options: 1 is turn on the vdisk and the other is delete it.

I tried to turn the vdisk on but then the screen said that I still have the hardware failure and I have to repair it first.

Now I am waiting the replacement disk, but I want to know what happened to my EVA.

Do anyone have any idea?

Thanks for reading.

[Anthony Martin_1]

The configuration that you describe is the cause of your problem. In this configuration (protection level = 0), whenever a disk drive fails you will suffer an outage of some sort.

Here is the link to the Best Practices White paper
http://h20195.www2.hp.com/v2/GetPDF.aspx/4AA2-0914ENW.pdf

You need to change DG1 to have a Protection Level (Sparing) of 1.

Anthony

[Víctor Cespón]

Even if you have "Disk drive failure protection" set to 0, a RAID 1 is a RAID 1. There should not be any data loss.

All vdisks are spread over all disks on the same manner, so all vdisks should have the same status.

You need to get the log file of that EVA checked by someone.

[Patrick Terlisten]

Hello,

a single disk failure shouldn't result in a failing VRAID 1 vdisk. You should log a call at HP for technical log analysis. Maybe another disks got a problem.

The protection level has nothing to do with that. It's only a "space reservation".

Regards,
Patrick

[ngo son]

Thank you,
I have send all the log to HP staff and waiting their answers.

Regards.

[DogBytes]

Only one enclosure involved here... perhaps where the problem lies. Please let us know what HP says.

[McCready]

This is one of those config's that I never hope to have...

I think that it is a combination of
"protection level = 0" and "utilize nearly 100% capacity." Seems like there is no space left for the EVA to mirror the data that was on the failed drive, so it is now "hanging" until the failed drive is replaced.

Fix: Always have at least twice the size of the largest drive you have free in the disk group with protection level 0, or simply set it to protection level 1, which is what it functionally does. I would still leave at least another drive's worth of space available as a general rule, which might make rebalancing an RSS easier to do, depending on how the drives were added in the first place (the RSS group may not be big enough to continue without getting a replacement drive in.)