Start a Conversation

This post is more than 5 years old

Solved!

Go to Solution

1556

January 26th, 2007 13:00

Replacing a failed HBA

Hi all, I was wondering if anyone here has a down and dirty procedure for replacement of a failed HBA on a Solaris server? Our SAN is multi fabric with using Brocade switches and CX600 storage arrays. We have a Solaris server HBA that failed and since it is a GEPACS server I want to do the replacement and rezone as quickly as possible to get the app back online.
I'm guessing that prior to replacement I should deregister the old HBA in Navi, Then once the new HBA is installed rezone with new WWN. That done I can update the storage array it is connected to check for connectivity and away we go. Did I leave anything out? Thanks mack

4 Operator

 • 

2.1K Posts

January 29th, 2007 08:00

So just for a complete picture, if the version of FLARE you are running on the array is older you may need to plan for the alternative to using the reconnect button. I would suggest checking ahead of time if you have it available if you are running anything older than FLARE 19.

If you don't have the reconnect button you have to plan for an outage to the storage on the server. The only way to get all the initiators properly connected to the storage group is to remove the host from the storage group and put it back in (after registering the initiators of course). This will get all the initiators connected properly. We used to have to do this, but I can't remember if the change came in FLARE 16 or 19. Hopefully you won't have to do this as the reconnect is much "nicer".

Good luck.

4 Operator

 • 

2.1K Posts

January 29th, 2007 08:00

Sorry Mack, I tried to post this earlier today but Powerlink seemed to be having "issues".

The only thing I would add may or may not have been a part of your "check for connectivity" at the end of the process.

Often a new HBA will be registered on the array but not properly associated with the storage group for the host. On newer version of FLARE (I'm not certain when the change was introduced) you can right click on the host in NaviSphere and look at Connectivity Status there (instead of for the entire array). It will; be very clear on this screen if the new HBA initiators are connected to the proper storage group. If not, they will be showing attached to "~management". If this happens, use the reconnect button on this dialog to correct it.

Once you have the initiators all pointing to the right storage group you just need to check PowerPath and make sure it starts using the new paths and the old paths are removed. I'm a bit weak on the actual commands needed for this on Unix hosts, but if you don't know them, someone else should be able to provide that. If it were a Win2k host I would use "powermt config" to make sure the new paths are recognized and "powermt check" to get rid of the old dead paths.

Hope this helps.

21 Posts

January 29th, 2007 08:00

I beleive that option was made available to FLARE 19 which is what I am running.
I do have the reconnect option. That should make things much easier that way. But I will still need to rezone the new hba prior to reconnecting the new HBA to the storage array since the new WWN will change.

21 Posts

January 29th, 2007 08:00

Thanks Allen, good additional info, I figured I had it covered pretty well but wanted to make sure I wasn't missing something important. This is only the second time in 3 years I've had to do this. And this is going to be under a time constraint after hours so I want to make sure it goes as smoothly as possible. I believe the power path commands are pretty much the same for unix based servers as well. I will make sure I have those commands available as well since I can't touch the servers as they belong to GE. I'll have to give them to the GE CE to run. Thanks again...mack

21 Posts

January 29th, 2007 08:00

I beleive that option was made available to FLARE 19 which is what I am running.
I do have the reconnect option. That should make things much easier that way. But I will still need to rezone the new hba prior to reconnecting the new HBA to the storage array since the new WWN will change.

21 Posts

January 29th, 2007 08:00

Lots of looks but no responses....hmmmm...that ain't good...
No Events found!

Top