Is clustering required for Auto Failover?

[webguyz]

Would like to replicate data on DSS V6 and have it highly available, but not ready to do clustering at the server level. If someone could clear that up for me I would appreciate it.

Thanks!

[symm]

I hope this clears things up for you
I tested autofailover with 2 DSS with autofailover and 1 windows server, it worked
(no clustering on the windows server side)
use the doc on the open-e website to setup the dss autofailover

[webguyz]

Thanks! Just what I needed!

[jisaac]

Just be aware that autofailover does have its limitations.

- the best practice for DSS software updates is to shut down the failover service (which drops all iSCSI connectivity and means you will need to take the server or servers connected to the DSS off-line), update one DSS, reboot, update the other DSS, reboot, then restart the failover services. This means your server will have downtime.

- if you want to add additional storage volumes to the failover set, you have to shut down all hosts accessing the DSS, turn off failover, add the volumes, then turn failover back on and then bring your hosts back up. Again, server downtime.

If your maintenance window permits this kind of server downtime, then go for it. Just be aware that there are times when you'll have to shut down your servers to do maintenance on the SAN.

Hope this helps,
James

[SeanLeyne]

Just be aware that autofailover does have its limitations.

- the best practice for DSS software updates is to shut down the failover service (which drops all iSCSI connectivity and means you will need to take the server or servers connected to the DSS off-line), update one DSS, reboot, update the other DSS, reboot, then restart the failover services. This means your server will have downtime.

I thought the best practice was to take down one node of the failover group, upgrade that node, restart the node, re-establish the failover grouping (allow the failover re-sync to occur) and then take down the other failover node.

Using this process, no impact on the iSCSI initiators (servers access iSCSI) would be seen, since the disk access would follow the failover path.

[webguyz]

Just took a good look at autofailover docs and realized it only supports Block I/O and not File I/O which won't work for us since all our volumes are File I/O. Is there any hope of them supporting autofailover for those of us using File I/O??

Thanks!