vSAN maintenance

[dgreebe]

Hi All

I have a question which I cannot find the answer to;

We have 5 vSAN (hybride) hosts, with a total storage of 84TB. 2 hosts are running 6U3, the others are 6U2.

I have to update 3 hosts, so I placed 1 host in maintenance with "Ensure Availability" and within 5 minutes my host was in maintenance. After that I was trying to update the firmware of my SSD's, but that didn't work (no update has been applied) so I started with the host update.. I think it took about 30 minutes untill it was completed.

After I've put my host back in the cluster, about 3.5TB was starting to resync. I can imagine why, because certain objects on the host (that was in maintenance), are not in sync with the other part of the objects (FTT=1) but I don't no why that much.

We have a particulare small environment but what about an environment with 10 VSAN hosts and arround 150TB of storage ?

Doing updates on that will cost allot of time.

The maintenance-window also took more than 60 minutes, so I also checked if the automated-rebuild was triggered but couln't find anything in my events on the hosts or on the vSAN Cluster... Does someone knows where I can see that this event/condition has been triggered ? I cannot look into the vsantraces-files.

Can someone help me understand the situation ?

- Why was 3.5TB needed to resync (which took 8+ hours)

- Where is the event shown that my vSAN automatically rebuilding because my host was more then 60 minutes in maintenance.

- Based on my experience above, why should I use "ensure availability"?

in my opinion, doing maintenance by setting the host in maintenance with Full data migration is almost the same advantage, because also than allot of data has to be moved....

Hope somebody can clear things up for me...

[batuhandemirdal]

Vsan The construction of the world is long as you say yes.

[dgreebe]

Sorry, but what do you mean with that ?

[SureshKumarMuth]

We are not sure what was the component status before upgrade, if you are doing the upgrade/update on all hosts continuously without checking the resync statue , it may lead to this situation. again for 30 min 3.5TB data resync is bit high and in my guessing, I think there was pending resync before you started with this host update.

You have RVC command to check the component status and resync status and it is recommended to run and check the status often while updating multiple hosts.

- Why was 3.5TB needed to resync (which took 8+ hours)

Reason is given above

- Where is the event shown that my vSAN automatically rebuilding because my host was more then 60 minutes in maintenance.

vSAN will automatically rebuild, it is the behavior of the product so no event will be generated for that, you may run commands to check the status

- Based on my experience above, why should I use "ensure availability"?

Ensure accessibility will check if there is minimum level of component availability to run the vms fine. this process will check all the component state and create a missing component before entering the host into maintenance mode.

In case of full data migration it will take more time based on the amount of data residing on the host which is to be placed in maintenance and based on the storage policies.

https://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/products/vsan/vmware-ruby-vsphere-...

Tips and tricks on Ruby vSphere Console (RVC) for managing a Virtual SAN environment - Support Insid...