EVA 6100 - Controller detected only one port of a FC device

[Bob Firek]

I received the following error message on CV this morning.
____________________________________
Full Description:
An HSV200 controller has detected only one port of one or more Fibre Channel devices. These Fibre Channel devices have entered the 'Single Port on Fibre' state, which should be corrected as soon as possible. The Storage System, the connected Fibre Channel port and the missing Fibre Channel port are reported in the FRU section. A summary of the affected devices is reported in the Evidence section.

Failed Component Type:
BE Port

Full Corrective Action Statement:
This event is probably a symptom of another problem. Check for failed Fibre Channel ports and multiple instances of this event. Also check for patterns, such as multiple occurrences of this event pertaining to physical disk drives on the same rack or loop. If this is an isolated occurrence of this event, ungroup the physical disk drive and remove it from the system.
_____________________________________


From what I can determain this is an isolated occurrence. I want to remove this drive so I can replace it. Last week this same drive had an issue in that it lost its connection. I decided to ungroup the drive but since this is an eight drive Disk Group I get the message:" Operation failed. Command rejected - Diske in the group would go below the minimum required number."

Is there any way to safely ungroup this drive so I can replace it or is this error an indication of something othe than the hard drive.

Any help would be appreciated.
Thanks,

Bob

[IBaltay]

Hi,
you can review the EVA controller log via the Command view EVA to see what disk is affected and then you can online (without the downtime)ungroup this disk and after the ungroup you can online replace it...

[Víctor Cespón]

You cannot ungroup a disk from a 8-disk disk group. You'll need to get a new disk, add it to the disk group and then remove the failed one.

[Bob Firek]

VCESPON,

Your suggestion make sense and is a good one. the only problem is that I currently have the array full. Could I ungroup a drive from the other diskgroup that I have that has more than 8 drives then add that drive to the 8 disk drive group, ungroup the bad drive, replace it and then assign that drive to the other disk group that I borrowed it from. what do you think?

Also someone here suggested that we just pull the drive out, let the disk group level out and then replace the drive. That worries me a bit but the argument is that is is build for a drive failure. I just can't help but think that somewhere is the slot electronic something might get confused. what do you think.

[Bob Firek]

IBaltay,

That would be okay but as VCESPON pointed out, I can not ungroup a disk in an eight disk drive group. What do you think about just pulling the drive out?

[IBaltay]

the idea of ungroup the disk from the DG in which you have more than 8 disks seems to be good, then you add it to you affected DG(which will have temporarily 9 disks) to ungroup the failing one...

[Víctor Cespón · Accepted solution]

We'll you have your EVA on a not recommended configuration. An 8-disk group and all slots filled. This leads to the inconvenients you're seeing now.

Of course you can just pull the drive, the EVA will take it as a disk failure and let you replace it. But pulling a drive that is online is not supported.

Very likely that nothing happens, we have done worst things on EVAs on the lab and trainings, but not a suggested procedure for a production system.

In this case you have to ungroup a disk from another disk group to have an slot where to add a new disk. Then add a disk there, ungroup the failing disk and replace it by the first disk we ungroupped.

Once the operation is done the disks will not be on the same positions as before, but that's not important.

[Bob Firek]

vcespon,

You are correct. An 8 disk group is a pain. As luck would have it we are in the process of ordering drives and controller to populate our EVA so this problem will go away soon I hope.

I understand that just pulling a drive from an otherwise working array is not supported or really even wise but in general should not cause any problems in your experinece, correct?

Also I your experience, this issue with my drive, is there any chance that it may be a problem with the slot or have you seen this type of error before with flakey hard drives? The operationa state in CV is One loop connection lost. I've checked everything on the EVA that I could think of, even the logs on the loop switches and as far as I can tell evreything points to the hard drive.

Thanks for your advice.

Bob

[Víctor Cespón]

For the "One loop connection lost" situation you can even try to upgrade the firmware on the disk. That will perform a reset of the disk and sometimes that's enough for it to recover connection to the loop.

This situation does not mean that the drive itself is faulty, many times just reseating the disk solves the problem (pull and reconnect before 1 minute).

If the connection keeps failing, then replace the disk.

[Bob Firek]

vcespon,

Got ya. Acutally this drive totally disconnected last week. The disk group releveled and had automatally ungrouped the drive. I went ahead and reseated it and it came back on line. I let drive spin for several hours with no issues so I regrouped it. Now exactly seven days later I have this error. I just emailed our vendor asking for a warrenty replacement. Hopefully he will get one here soon. Thanks a lot for your advice.

Bob