2016-01-06 03:36 AM
we are seeing an issue with our Snap Mirror jobs which run from our Production NetApp over to another NetApp at our DR site. This issue is occuring on both controllers.
Currently we have a point to point 100Mbps link between our HQ and DR site, this line has been moved to another MPLS provider. The lines details and WAN networking equipment are still the same, just the provider has changed.
What we have found is since the changeover to the new WAN ptp link, all snapmirror jobs which run to the DR site are failing with the message ‘transfer’failed’. When watching the job start at the intended time, its begin to perform a transfer, is able to transfer a certain amount of data but then eventually fails with the message transfer failed appearing on the screen. If we try to rerun the snap mirror the same thing happens again.
Fortunately for the moment we still have the old MPLS link available and have the ability to revert back to it, as soon as we do this, the snap mirrors work without issue.
We have done the following to understand and try to address the issue:
I have taken a look at the snapmirror.log file on the destination filer and since we have made the change today I can just see details that the transfer failed but no additional detail on what caused it.
Below is an exmaple:
dst Wed Jan 6 10:03:54 GMT CENSLFASPROD01:NETAPP_prf1_esx_w2k8_sas_04 CENWBFASDR01:NETAPP_prf1_esx_w2k8_sas_04 Start
dst Wed Jan 6 10:10:01 GMT CENSLFASPROD01:NETAPP_prf1_esx_w2k8_sas_04 CENWBFASDR01:NETAPP_prf1_esx_w2k8_sas_04 Abort (replication transfer failed to complete)
Does anyone have any clues what might be the issue? Has anyone else suffered similar issues?
2016-01-08 09:12 AM
Did you have a look at the snapmirror.log on the source to see if there were any related messages there that might be more useful?
You could also try breaking and then resyncing one of the relationships over the new WAN link to see if that would kick start things.