*SYSTEM DOWN* NATIVE SAN - ISCSI Cannot login

[vikeswin]

We are running NATIVE SAN 5.8.1975.
Windows 2008 R2 SP1 with cluster services.
2 Node cluster.
Dell R720

We had a power failure last night, the Windows Hosts can back up, howeve the cluster is failing due to CSV, and Qurom missing.
Checked and my ISCSI connections will not connect.
I have removed the all from the favorites, and restart boxes still nothing just a error message "Unable to Login to the target."

1. Question should I remove then the recreate the target names from Starwind, if so general steps to folllow.
2. Any other things to consider besides recreating target names, any other "ISCSI" type troubleshooting.
3. Last should I stop (not destory) cluster services prior to removing anything, does it really matter?

Thanks!

[anton (staff)]

Since you're already both nodes down (did not get how you've managed to do this assuming running UPSes you should have at least graceful shut down w/o even need to resync the content on your both nodes) make sure you do upgrade to V8. Version you're running is too old and recovery sequences are very different for 5.xx and 8.xx branches.

[vikeswin1]

Since you're already both nodes down (did not get how you've managed to do this assuming running UPSes you should have at least graceful shut down w/o even need to resync the content on your both nodes) make sure you do upgrade to V8. Version you're running is too old and recovery sequences are very different for 5.xx and 8.xx branches.

Anton,
Thanks for the reply.

1. So how do I upgrade just install over the current 5.8?
2. What about my current 5.8 license, now do I need to contact Starwind to get a key for 8.x?
3. Can I rollback to my older version, the software still works (Sorry you built a good product!).
4. I cannot believe there is not some "generic" or general iscsi connection troubleshooting tips.
5. Yes we had a power failure, and the machines did not shutdown gracefully.
6. I have a test ISCSI starwind non-ha working, it's my original ones that will not reconnect.

thanks!

[anton (staff)]

1) Seamless upgrade from 5.xx branch to 8.xx branch is not doable (that's why we recommend keep your versions up-to-date). Preserve IMGs and wipe off 5.xx and install 8.xx instead.

2) Contact sales. Old 5.xx used binary keys and new 8.xx uses XML keys are they are not compatible.

3) Yes, you can but you'll have an issues with support.

5) UPS-all-around should solve your issue (in the future).

[vikeswin1]

Okay the system has been restored.

The problem (DOH!), was the nodes needed to sync.
I synced the nodes, and boom i was able to create ISCSI connections, which connected.
Not anyone's fault but mine, but it's been easily over 8 months since I've been in the SW console.
So I forgot about the "by design" data protection of not allowing ISCSI connections without a sync.
I'm thinking I need ISCSI connection in place prior to the sync thus the "DOH".

Below are a couple of suggestions I have come up with, I understand from a SW position.
We rely on the O/S for x,x,x however these are a few outside the O/S things I come up with.

Suggestion box: (granted these could be in v8 not 5.

1. If a system goes out of sync, alerting would be nice. (phone home)
2. If a system is out of sync a message about ISCSI connections will not connect. (Which is by SW design, but in "crisis" mode something easily overlooked)
3. Alerting, or email reporting on health of .img file, or data integrity at risk.
4. A better understanding of SW recovery process from a "hard" shutdown.
5. Last what are others doing to the protect, recover their host Windows systems? (I use Veeam for VMs, fantastic product, considering DPM?)
6. Sorry for the initial typos I know it was hard to read, i'll do better

Cheers!

[anton (staff)]

Good you're good now!

V8 (unlike 5.xx branch) comes with an advanced set of an alerts and notifications did you get a chance to review it?