Maintenance mode

[tractng]

Hello,

We are have a planned network core maintenance.  We are not sure if the 5 hosts and the SAN will be affected so we are planning to shutdown the VMs and place the hosts into maintenance mode. We run NFS and iSCSI in our environment.  I have a few questions.

1) Do we disable DRS (currently set at Partially automated) prior to placing the hosts into Maintenance mode?

2) Do we manually shutdown down (right click and shutdown) the 3 instances of vCLS?  Last time we had done something similar, the vCLS kept on powering back on.  The last host had vCLS which prevented us from placing in Maintenance mode.  When we finally shutdown vCenter, then was able to manually powered the vCLS down.  Just want to know if this is done correctly.

We are on version 7 of vCenter/ESXi.

Thanks.

TT

[a_p_]

Do not disable DRS, as this will affect configured option, e.g. delete resource pools! Instead, set it to "Manual" to temporarily deactivate it.

The steps to shutdown the whole vSphere environment should work as you mentioned. However, there's another option to disable/delete the vCLS VMs as the last step before you shutdown the vCenter Server. See https://www.yellow-bricks.com/2020/10/27/demo-time-how-to-delete-the-vcls-vms/

André

[tractng]

As always, thank you for your feedback.

The sequence I envisioned:

<Prior to maintenance>

1) Place DRS in manual mode

2) shutdown all the VM (including vCenter, vCLS at the end)

3) Put hosts on Maintenance mode

<After maintenance>

1) Verify network connectivity

2) Take hosts off Maintenance mode

3) Power on vCenter (let it fully boot up)

4) Power rest of VMs on

5) Set DRS back to Partially automated

[a_p_]

I added the shutdown, and power on sequences, and moved DRS up one step for the restart.

<Prior to maintenance>

1) Place DRS in manual mode

2) shutdown all the VM (including vCenter, vCLS at the end)

3) Put hosts on Maintenance mode

4) Shut down ESXi hosts

5) Shutdown storage systems

<After maintenance>

1) Power on storage systems

2) Power on ESXi Hosts

3) Verify network connectivity

4) Take hosts off Maintenance mode

5) Power on vCenter (let it fully boot up)

6) Set DRS back to Partially automate

7) Power rest of VMs on

[tractng]

Interesting that you mentioned storage systems.  Last year we had a maintenance system similar to this one, some of the advice was to leave the storage systems online (i left it online) since the network connectivity would come back up on its own.

I am so reluctant to power any hardware especially the SAN down when they have been running many years.

TT

[a_p_]

Sorry my bad, I somehow missed that it's just a network maintenance.

André

[tractng]

Thank you!

[tractng]

hi a_p_,

I have a questions regarding #6.  Why would I want it to be set back to this setting (Partial automated) prior to starting the VMs?  When we turn on the VMs across the 5 hosts randomly, would DRS shuffle the VMs around too much?  Or maybe since its partial, the shuffling of the VMs to different hosts won't occur.

6) Set DRS back to Partially automate

[a_p_]

It actually depends on what you want to achieve.

With DRS in "Manual" mode, you'd have to acknowledge the Power On Recommendation for each VM. If this is what you want, i.e. power on VMs on selected hosts, then set DRS to "Partially Automated" as the last step.

However, for VMs that should/must run on predefined hosts or on the same host as other VMs, I'd suggest that you configure DRS rules, and set DRS to "Fully Automated".

André

[depping]

---

@a_p_ wrote:

I added the shutdown, and power on sequences, and moved DRS up one step for the restart.

<Prior to maintenance>

1) Place DRS in manual mode

2) shutdown all the VM (including vCenter, vCLS at the end)

3) Put hosts on Maintenance mode

4) Shut down ESXi hosts

5) Shutdown storage systems

<After maintenance>

1) Power on storage systems

2) Power on ESXi Hosts

3) Verify network connectivity

4) Take hosts off Maintenance mode

5) Power on vCenter (let it fully boot up)

6) Set DRS back to Partially automate

7) Power rest of VMs on

---

Also, if you are using retreat mode for the vCLS VMs, you will need to disable it again so that the vCLS VMs are recreated.

[tractng]

Our maintenance schedule went well.

Thanks for your help!

[Tinto1970]

---

@tractng wrote:

I am so reluctant to power any hardware especially the SAN down when they have been running many years.

a good advice is to be reluctant to let the SAN run for many years without an update 😉

---