Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I have an FAS270C. For months, I've been running it in a split-head manner (that is, with each head serving data totally independently, and without any clustering even being enabled) in order to facilitate moving some data around. I finally got everything situated, moved all the data to one of the heads, and was trying to get clustering set back up.

Now when I try to install OnTap onto the "new" head, it cannot see any of the disks in the head shelf. (That is, the shelf into which the heads are inserted.) I've booted into maintenance mode, and it shows me that the 0b adapter, which should be the adapter that that shelf and its disks should be presented on, is in "OFFLINE (physical)" state. If I try to enable it with either "storage enable adapter 0b" or "fcadmin online 0b", it waits for about 30 seconds and then says:

[fci.initialization.failed:error]: Initialization failed on Fibre Channel adapter 0b. [fci.adapter.online.failed:error]: Fibre Channel adapter 0b failed to come online.

There is currently nothing attached to its external 0b port. I've tried it with and without an SFP plugged into it, and with and without its internal termination switch on.

The currently active head can see those disks, and can see that two of them are assigned to the other head. Before I started reconfiguring, the "new" head could see disks on that shelf. They may even be the same disks that OnTap was installed on previously.

Does anyone have any idea how to proceed?

share|improve this question
add comment

2 Answers 2

If you can get away with some downtime for both heads, the first thing I'd try would be to swap the slots the controller cards are in, to isolate if it's an issue with the brain card itself, or with the slot in the shelf. (You say there's no external connection to 0b; is this a system with only the head shelf and no external shelves on 0c as well?)

share|improve this answer
    
Right. For the time being, there are no external connections to the "bad" filer at all. Checking the slot location is a good idea I didn't think of. If worse comes to worst, I can schedule some more downtime, but I'm actually trying to get this cluster set up to avoid some other downtime. Typical, huh? –  wfaulk Apr 15 '10 at 14:14
add comment
up vote 0 down vote accepted

Turns out that I had to have an SFP ("mini-GBIC") module plugged into the 0b port and with a physical termination adapter connected to that. The internal termination switch doesn't seem to have any effect.

I also updated the "partner-sysid" environment variable in the head's CFE ("BIOS") to reference the other head's serial number. It was undefined. I don't think this had any relevance, though, as I think that's only referenced for clustering, and, again, I hadn't gotten that set up yet.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.