Failover service won't start

[andrewa]

Since upgrading to V6 back in December I have been having problems getting the fail-over service to start, despite going through the setup of volumes, targets, networking etc 2 or 3 times now.

I keep getting a "Configuration of both nodes finished with errors." message each time I try and start the service on the primary.

I am using 6.0up14.8101.4221 32bit with File I/O iSCSI volumes that have been initialised.
The replication tasks are running Ok and consistent but don't know what has changed or what I am doing wrong.

Is there anywhere I can look to see what may be causing this or can anyone offer advice on how to resolve?

Thanks

Andrew

[andrewa]

Really starting to need some help here and so hoping one of Open-E's developers or similar will step in to help, as I am guessing that no other users have experienced this or been able to fix at least.

I have now upgraded to ver6-update4362 hoping that there was a bug in last update but this has made no difference
.
This time instead of restoring config from backups I reconfigured both nodes from scratch on both, recreating iSCSI targets etc ensuring the same LUN no, SCSI ID and target name but when I attempt to start the failover service on the primary, it gets busy for a while talking to the secondary and looks like it's going to work and then comes back with the same unhelpful "Configuration of both nodes finished with errors." message. At least when I stop replication and attempt to start failover it provides some useful information that the replication tasks are not running.

Is it
- Something to do with my File I/O volumes (using VMware), as these are the same as I have used since earlier version 5 installations?
- Network configuration?
- Volume metadata?

I have gone through the video and the step-by-step guides but can't find any major gotchas that I've over-looked but I must be missing something! We used the failover without any problems before upgrading to v6 so something must have changed that is affecting it???


Any help would be appreciated.

Thanks in advance


Andrew

[pfeifferl]

Hello,

maybe it's the same issue we have ...

Each time we try to start the failover service there comes a configuration error. I encountered that the replication was reconfigured by system from source - destination to source - source. I must correct the replication, then start the replicatrion again and the failover service can be started.

This happens each time we shutdown the systems. I think it's a bug ... (open-e support?)

[andrewa]

Thanks for you reply but my replication has been working fine with source and destination settings remaining as they should.

I think I may have discivered what the problem is but not sure if this is a bug or a known issue.

It would appear that after upgrading to V6 you had to change the iSCSI daemon option from the default SCST to IETD in order to see LUNs/datastores previously created with Open-E v5 and formatted as datastores by Vmware, the iSCSI failover service will not start!

Even when volume replication is consistent, iSCSI target names and SCSI IDs set to be identical, the fail-over service will not start if you are running with the IETD iSCSI option.

You need to switch to SCST and create new or re-format existing volumes in order for VMware to see the LUNs/Datastores, both of which require a large maintenance window and moving of VM data files in order to fix.

Not sure if there is some way of re-signaturing the disks rather than re-formatting but hopefully someone will be able to assist on this one. I vaguely remember seeing something somewhere about it but can't find it now!


Thanks

[webguyz]

I was getting ready to redo my DSS server and get ready for Citrix Xenserver which means setting my ISCSI to IET from the SCST I currently have. From what your telling me is if I do this I will not be able to do autofailover since the volumes are IET, is that correct? Ouch!

Can anyone from Open-E verify that?? If thats true then I won't be able to use Xenserver with DSS 6 because I need to ultimately need have autofailover to a second DSS.

[webguyz]

Latest versions of DSS V6 now support SCST so there is not need to switch to IET for the ISCSI daemon when using Xenserver.