TrueNAS and ESXi best practices 2022 ?

[Harry33]

Good day everyone. I am using TrueNAS Scale with VCenter 7. The ESXi hosts are version 7 U2D. The ESXi hosts and the TrueNAS server are connected with 10GbE.

I have read this thread - https://www.truenas.com/community/t...ractices-for-vmware-nfs-datastore-2018.71086/

I am looking for some feedback on best practices for using TrueNAS Scale with ESXi/VCenter. I am currently using the Release version of TrueNAS Scale, with NFS3, NFS4, and ISCSI datastores in VCenter.

Before performing an upgrade of TrueNAS and rebooting, what should be done on the VCenter side ? Should all VMs be powered down ? ESXi hosts put into maintenance mode, or ESXi hosts powered down ? Datastores unmounted in VCenter and re-mounted after the TrueNAS system comes back up ?

Any ideas on NFS versus ISCSI ? I'm not concerned with the last few percent of performance. Stability is my main goal.

 

[jgreco]

Stability is my main goal.

Yikes. Good lord, then, why are you running Scale? Core is intended for this purpose.

Before performing an upgrade of TrueNAS and rebooting, what should be done on the VCenter side ?

The usual options are to;

1) Migrate all workloads off the datastore(s), or
2) Suspend (hibernate) all the workloads, or
3) Shut down all the workloads.

The usual VMware stuff for maintenance options. Has nothing to do with TrueNAS really.

 

[Harry33]

Thank you for the reply jgreco. I guess I should have mentioned that this is a test environment :)

I applied the Release update to TrueNAS Scale today. After rebooting the TrueNAS Scale system, the NFS datastores showed as inaccessible in VCenter. The ISCSI datastore still worked. I restarted the ESXi hosts, and restarted the NFS service on Scale, and then the NFS datastores became available again.

For anyone working with Scale/Core and VCenter 7, what is your plan for TrueNAS updates ?

I am open to using either Scale or Core. Scale was the first product that I tried. I will try Core later on.

 

[HoneyBadger]

Welcome to the forums.

In a test environment, you can afford to be a little more fast and loose with the upgrades. Shut down all the VMs you can, hot-migrate the ones you can't (vCenter comes to mind here as a top candidate that you want to keep running) to another datastore (a second TrueNAS machine, perhaps?) and then apply the update.

But if you're "testing" with a plan to "move a solution into production" then my answer changes significantly, and also starts with "ask yourself if you really want to be the throat that gets choked in a SAN-down scenario, and if not then call iXsystems"