MSA Storage
cancel
Showing results for 
Search instead for 
Did you mean: 

HSG80 external cache battery failure

 
dentistcorn
Regular Advisor

HSG80 external cache battery failure

Today we had a production outage due to both Cache Batteries failing in the same controller enclosure. The batteries had just been replaced Saturday morning with brand new out of the box ECBs.
When the ECBs failed, all the LUNs on the controller pair failed. Has anyone seen this before? I presumed the ECBs were to allow the controller to commit data in cache to disk in the event of a power outage. Why would all the LUNs fail?
The controller is at 8.6-8. I checked the HBA settings and everything is good. We actually had HP bless the config back in February and haven't touched since.
Any assistance would be greatly appreciated.
2 REPLIES 2
BR757138
Occasional Advisor

Re: HSG80 external cache battery failure

Can you specify the actions you did. Be as specific as possible.


Regards
Erwin van Londen
HP Master ASE SAN Architect
http://www.hitachi-storage.com
Uwe Zessin
Honored Contributor

Re: HSG80 external cache battery failure

The cache batteries will only preserve the contents of the cache memory until main power is back - they do not power the controllers and disk drives to allow a cache flush (if that is what you mean by "the ECBs were to allow the controller to commit data in cache to disk in the event of a power outage").

If a battery fails the controller will stop access to the unit because it needs a protected writeback cache for operating mirrorsets and RAIDsets to prevent the 'write-hole'.

You could set the controllers to UPS=DATACENTER_WIDE, which tells them to no longer check the battery status, but:

any loss of power will cause loss of data! Worse, you could damage your filesystem(s), because the writeback-cache undermines any careful write-ordering of the file system to keep it's integrity!!
.