ONTAP Discussions

DR question

deygaurab
2,594 Views

Hi     I have metrocluster setup between 2 sites on a FAS6210. I have a query related to in the event of a DR. If first site goes down,How can my host in the second site  access the mirrored LUN (from FAS6210) in the DR site. It is clear that the mirrored lun on the second site will be used in the even of a site disaster. But how will the host in the second site access the lun? any zoning activation required?

Regards

1 ACCEPTED SOLUTION

aborzenkov
2,594 Views

MetroCluster is standard HA pair; the only difference is ability to do disaster recovery. It means, hosts on every site must have SAN connection to both controllers, just like with normal HA pair, and use multipath software. As you said, disaster recovery procedure effectively does takeover; so hosts will continue to have access to LUNs, the same as would happen after normal takeover.

The primary difference is, normal takeover happens automatically and completes in small amount of time, so it is transparent. Disaster recovery usually is performed manually, after disaster has been declared. So most likely all hosts accessing LUNs on failed site had already timed out IOs and need to be restarted to ensure sane behavior.

View solution in original post

4 REPLIES 4

aborzenkov
2,594 Views

Detailed step by step procedure for disaster recovery is documented in High-Availability Configuration Guide for your Data ONTAP version (it was called Active/Active Configuration Guide in earlier versions). Did you read it? Do you have an specific question about this procedure?

deygaurab
2,594 Views

My question is specfic to the steps required to be done once a site fails. I understand , a manual cf takeover is required. but is that enough for the host on the DR site to see the mirrored LUN ?

Regards,

aborzenkov
2,595 Views

MetroCluster is standard HA pair; the only difference is ability to do disaster recovery. It means, hosts on every site must have SAN connection to both controllers, just like with normal HA pair, and use multipath software. As you said, disaster recovery procedure effectively does takeover; so hosts will continue to have access to LUNs, the same as would happen after normal takeover.

The primary difference is, normal takeover happens automatically and completes in small amount of time, so it is transparent. Disaster recovery usually is performed manually, after disaster has been declared. So most likely all hosts accessing LUNs on failed site had already timed out IOs and need to be restarted to ensure sane behavior.

deygaurab
2,594 Views

Thanks Aborzenkov..

Woudl you have the Ontap A/A configuration guide as you mentioned earlier?

Regards

Rahul

Public