2012-04-25 05:26 PM
We've been testing SRM 5 failover with a client the past week or so and came across a strange issue this past Monday.
We have PowerShell scripts running at the start of the SRM failover that breaks the SnapMirror relationship for a selected group of about 15 LUNs.
These LUNs are being replicated via SnapMirror every 15 minutes and are also updated every 12 hours with schedueld full DB backup jobs and every 1 hour with transaction log backup jobs via SMSQL.
We successfully failed over to the DR site, SRM worked (yay) as well as the PowerShell scripts. We brought up a VM...LUNs attached...able to run an SMSQL backup to verify...looking good.
About an hour into the exercise, the SQL LUNs flip to read-only on the VM. Sure enough, I look at System Manager and see the SnapMirror relationship transferring data, even though I previously confirmed that all necessary SnapMirror relationships were correctly in a "broken-off" state.
My question is - would there be something that would replicate data via SnapMirror even if the relationships were broken off? We ran a test today to see if SMSQL would transfer despite the relationship status, but the test worked as expected...backup ran fine but displayed the proper error for failing to update SnapMirror relationship due to status being broken-off.
Any help is greatly appreciated.
2012-04-26 06:45 AM
Thanks for the responses, guys.
I suppose what I'm trying to figure out is if something would kick off the re-sync command automatically, like SnapManager or something else.
I have a case open with NetApp Support and am awaiting escalation. In the meantime, I will try additional troubleshooting to see if I can replicate the issue.
Just a very strange case indeed.
Thanks again - this is why I love the NetApp community
2012-04-26 06:50 AM
Ha ha Trey,
I don't think SnapMgr will kick off re-sync automatically.
Please alert the community the final finding from our lovely NetApp support gurus.