eva8100 leveling process

[susiri]

We have a EVA8100, had a loop failure in two enclosures which showing one loop fail in all the disks. We have done the code load for one disk (Eg Disk 023), after code load activity the Migration process was shows status of not migrating ( see the attached).

As a second step we have done ungroup, reseat, group, disk 023 but disk is still showing leveling process as 7%. How  long will it take to finish this process ?

If this task fail, what will be the next step ?

Highly appreciate your suggestion.

[Johan Guldmyr]

Hi, why did you upgrade firmware on anything when there was a loop issue?

Next step (or in my opinion, the first step) is to fix the loop issue.

[hvhari]

You should be concerned about the loop issues and get it fixed. Some times this can be quite time consuming. Not sure code load was part of your efforts to fix the loop issues.

Coming to "Levelling" - this is an low priority internal activitity within EVA. It can even take couple of days depending on the load on EVA (ie. I/O from hosts). However you need not worry on this, as it has no impact on the data availability.

Coming to "Migrating" - this is a process when a disk is being marked as "proactive failure" and the data in the disk is "moved" to other disks within the disk group. You should not make any HW changes (is replace/remove/reseat etc.) or do any Firmware upgrade etc., while this process is going on, as it can impact data availability.

[Johan Guldmyr]

>>>Coming to "Migrating" - this is a process when a disk is being marked as "proactive failure" and the data in the disk is "moved" to other disks within the disk group. You should not make any HW changes (is replace/remove/reseat etc.) or do any Firmware upgrade etc., while this process is going on, as it can impact data availability.

I'd like to add that it could make you lose your data, especially for vraid5 and if your hw change causes the same RSS to lose contact with a disk. Does it say 'ungrouping' when you manually migrate the data off a disk?

[susiri · Accepted solution]

We have sort this issue by restarting loop switch.

Thanks lot for all the valuable comments