How to handle an out of sync message when making disk changes

[webguyz]

Hi,

Newbie trial'ing Starwind in a 2 node Compute and Storage separate environment.
Everything was going good until I went into one of nodes Megaraid Storage manager and turned one of the disks in node 1 into a hot spare.

It seemed to hang for a second but then it did it. I looked at Starwind console and it told me node 1 was out of sync (I have 2 targets on it)

I was not sure how to get it back in sync and there appeared to be no traffic on node 1.

I remember the support tech telling me during setup that if I needed to take down a node I just do a normal graceful shutdown and bring it back up and it would force a resync, so I did a normal restart and it did work and resynced!

My 2 nodes are back in sync and everything seems to be working ok, but was wondering if there was a way to force a resync without rebooting the server. Seems that some raid server manager actions tie up the disk access for a few seconds, long enough to get the images out of sync.

Thanks!

[webguyz]

After reading the help and having a bus rescan on my LSI controller forcing my replication off line I discovered the Synchronization button is active (when out of sync) when right clicking on the image file.

Clicked that Synchronization button and it started a fast sync with I assume the partner. After that it was in sync and everything was working ok again. Didn't have to reboot the server which is very handy.

[Oleg(staff)]

Hi webguyz,
Could you please give more clarification about your environment?
Is StarWind VSAN running inside Windows VMs on ESXi environment or you have servers with StarWind VSAN?

[webguyz]

Hi,
Separate Compute and Storage nodes. Each node has 128gb of memory and a LSI 9271 raid controller. Whenever the LSI raid controller makes changes like adding a global hot spare or creating a target it requests a raid bus scan that always take longer then the Starwind default timeout and takes the node I'm working with offline and the targets out of sync. All writing from VMWare hosts stops on all targets on that node. If I wait until the raid function settles down I can then right click on each out of sync image files and the Synchronization button is active. I click on each image files Synchronization button one at a time then a Fast Sync starts and after a few seconds its Synchronized. Everything is back to working after that.

is that the correct procedure to handle these raid controller interruptions?

Thanks.

[Oleg(staff)]

Actually, the best way is to check if everything synchronized and stop StarWind on the node, where you plan to make changes on a RAID level. Because during RAID rebuild this side will work with degraded performance.

[webguyz]

Good idea. What is best way to stop Starwind? Stop the Starwind Virtual San service of the node I'm working on? Or is there a button I missed?

Thanks!

[Boris (staff)]

Stopping the StarWind Virtual SAN service using services.msc or PowerShell will do the trick.