EVA4400 - Read errors

[Joshua Small_2]

Hi,

My background is one of managing a number of MSA2000/P2000 SANs. A recurrent situation there is a HDD that occasionally, and often during a disk scrub, produces a regular log like "recovered read error after x tries".

Usually I get that disk replaced, because it's a safe bet that disk will eventually fail. One particularly bad scenario is letting multiple disks end up in this space, and having each of them go offline next reboot. However, the disk is always marked as "online" until I physically pull it.

I have an EVA4400 in front of me and I'm interested in how a similar situation plays out.

* I can't see a "disk scrub" or equivalent, is there such an equivalent?

* I can see a lot of logging on certain things but nothing that seems to report on single read errors. Disks report as "OK" but I'm concerned that this is a lot like the P2000 "OK". Is there somewhere I should be looking?

[Johan Guldmyr · Accepted solution]

Hi, 

no disk scrubbing on the EVA4400.

If you find the ASC/ASCQ code for the recovered read errors you may be able to find them -  http://www.t10.org/lists/asc-num.htm might help. 

The SCSI events are mentioned if you look closely in the "check condition" events in the controller event log. Look for the field ASC/ASCQ. One relatively common is 11/00 - unrecovered read errors.

However, the EVA usually is quite strict when it comes to these errors and fail the disk out itself.

If you start seeing a lot of these "check conditions" events in the controller event log but no disk being failed out, that might be a good time to ask HP support to check if there are any issues and if something should be done. Looking yourself in the controller event log is not so easy without HP support's tools.