topic Re: HSG80 in Disk Enclosures

HSG80

James R. Wentling — Mon, 08 Dec 2003 09:22:08 GMT

When replacing a controller in a dual controller configuration, we get the following error when we attemp to "set failover copy=this" 'cannot set failover with connection...unit offset out of range'. What paramater is this referencing? how do I correct?

Re: HSG80

Kevin Harlan — Mon, 08 Dec 2003 16:52:47 GMT

I am not an HSG80 expert, but I might be able to spot something. I would want to see the output of the following commands:

show this
show units
show connections

It's been a while since I worked with the HSG80, so I don't remember exactly whether offsets are a property of the units, or the unit names, or the connections, etc.

Re: HSG80

James R. Wentling — Mon, 08 Dec 2003 17:37:43 GMT

This has to do with the connections. Here is the history. The new controller came in formated for ACS V8.7. We run 8.6-8, I had to downgrade by reformating the nvram to be able to accept the 8.6 program card. This message is the result after the reformat and replacemnt card inserted and the set failover coppy=this was entered on the runniong controller.

Re: HSG80

Mike Naime — Mon, 08 Dec 2003 20:26:51 GMT

Are you sure that you did not mean to use
SET MULTIBUS COPY = THIS
This is short for setting Multi-bus failover mode.

Understanding that when replacing the old controller, you must first SET NOFAIL to take the controllers out of a failover mode. you are trying to set the controllers back to your original failover mode.

Did you try the help command? (?)
Example: SET FAILOVER ?

Re: HSG80

James R. Wentling — Tue, 09 Dec 2003 07:53:37 GMT

we run in transparent failover mode, and yes we set nofailover prior to replacement. I reaserched the set failover command in the Cli refernece manual, I did not see a qualifier that mentioned this offset value.

Re: HSG80

James R. Wentling — Tue, 16 Dec 2003 09:08:24 GMT

We finally called in HP for support. The fix was a bad ECB cable that caused the cache always see a low battery status, thus not allowing cache to be mirrored or controller redundant mode. I guess just a wierd message to indicate this failure.