Prior Updating
ESXi hosts to 6.0 Update 2
1
1.
Enable Anti-Affinity rules per requirements.
2.
Ensure Clusters have HA enabled
with single host loss.
3.
Ensure DRS is enabled and fully automated
with enough resources to balance Anti-Affinity rules.
4.
Create Baseline via Update Manager
with Critical and Important Patches, including the Full Update 2 Patch.
1. Create
a new Host Upgrade Baseline
2. Click
Next
3. Select
the extension baselines to include the baseline group
4. Click
Finish
Updating ESXi hosts to 6.0 Update 2
1)
Taken care of in the Prior to
Update, but do check resources are still sufficient to support a host down and
the anti-affinity rules.
2)
Place a host into MX mode
allowing DRS to migrate all VMs off of the host to be updated. Templates and powered
off VMs can stay on the host. *However, they migrate quickly and should have no
impact on the upgrade.
3)
Many Options to update ESXi.
For this process we are utilizing Update Manager.
4)
Disable any alerts / monitoring
to avoid any false alarms.
5)
Select the object (Host) click
Update Manager Tab
6)
Click Attach
7)
In the Attach Baseline or Group
Window, select the baseline that was previously created.
8)
Click Attach.
9)
Right Click the host and “Scan
for Updates”
10)
Select the type of Upgrade à Patches and Extensions
11)
Click Scan
12)
ReSelect the Host, click Update
Manager and Click Remediate.
13)
Select Upgrade Baselines,
Accept the terms and click Next
14)
Select Immediately to begin the
process
15)
Select the power state of VMs
(Since the Host was already placed in MX mode and no VMs live on the host, this
does not matter, but please choose Do Not Change VM power State.
16)
Select Retry entering MX mode in case of failure
17)
Click Next
18)
You should not get to the
cluster screen since we are doing a single host at once, so click Finish.
19)
Once the host remediation is
completed, Sanity check the host:
a.
Change DRS to manual (Only
first upgraded host, afterwards sanity checks should be quick – Suggestion).
b.
Place the host out of MX mode.
c.
Move one VM to the newly
updated host.
d.
Have Infosys verify network
connectivity and authentication (If required).
e.
Have a customer to check the
VM, once the VM is checked, turn DRS back to fully automated to allow the
cluster to become evenly balanced.
i.
If any issue is found, follow
the backout plan:
ii.
Migrate the VM back to another
host.
iii.
Place host back into MX mode.
Back out plan
If any issue arises after applying the patches to
the first ESXi host then stop any further remediation of patches until the
problem is rectified. We will engage VMware support immediately.
If no resolution is found within the window, we
will schedule for a re-install of the Host back to ESXi 6.0 (Same build number
previously). As the vCenter is not a
part of this specific part of the upgrade no rollback plan is required for it.
No comments:
Post a Comment