Replication and automatic failover

[tsv]

If you are using DSS V6 on 2 SAN's is it possible to have automatic fail over if the primary unit dies or do you have to manually move it after the failure is detected?

[symm]

it's automatic failover to the secondary node

please read the article in this link

Open-E DSS V6 Synchronous Volume Replication with Failover over a LAN


http://www.open-e.com/library/how-to-resources/

You will need to log in

[sweet_man]

If the connection is lost, the replication will resume as soon as the connection is back. Once the connection is restored, only the synchronization of the data that has not yet been transferred will be carried out. This will happen due to the ECC check of the transmitted data (tied to the metadata). As long as the replication remains active and there are no error-messages, the copy will be consistent.



If the source server fails, you need to promote the destination server to become your source and clear the metadata.



Afterwards, enable the target on the destination server so the initiator will connect.



A good practice would be to add the destination IP address on the initiator side in early stages of setting up the replication.

[hugor99]

Hi,

I'm not seeing how this is "Automatic" failover. When I hard-power-off my primary open-e box, the secondary box gets marked as "active", but the virtual IPs never come up.

I have to log into the secondary open-e box, mark the volume as "source" and stop/start the replication service. Then my vmware boxes can see the virtual ips.

Is there some way to make this more automatic, i.e. have the virtual ips come up on the secondary box when the primary box fails?

Cheers

[salmon]

hugor99 you definitly configured it wrong. It looks like you didn't add replication tasks to failover.

[hugor99]

hugor99 you definitly configured it wrong. It looks like you didn't add replication tasks to failover.

They are added. When I do a manual failover, it works fine. The VMs using the datastore on the volume still work, I can ping the virtual IPs, it's all working fine.

But when I just power off the primary machine, I can no longer ping the virtual IPs, and all my VMs that use the datastore freeze until I fix it...

[hugor99]

OK, I'm not sure what the difference was, but I reconfigured everything going back through the ppt slides from the website (which I did before, but maybe I missed something) and it is now working automatically.