It is April 15th 2020 and vROps 8.1 just got released yesterday and although I am on 8.0.1 currently I thought that there are some nice features in the new release. These features you can read more about here. I originally intended to be a single part post but then I realiased that since I use Application Remote Collectors, that’s a separate chapter. That will be covered in Part 2.
The steps below show the steps for 8.1, but the same steps also works from 8.1 to 8.1.1, and from 8.1.1 to 8.2.
Before you start the upgrade it is a good idea to check that you have set your root password on all appliances (and that the account is not locked)
You need to download the upgrade package from my.vmware.com and I suggest you also download the upgrade checker. Finally if you have the ARCs deployed (Application Remote Controllers) you might want to pull those down while at it.
Additionally it is recommended that you backup changes to alert definitions, symptom definitions, recommendations and views. These will be reset during the update. Also it is worth noting that you should not modify the default policy (for the same reason, I know just the time you want to read about this right?).
After doing a backup of these, you should stop the cluster from the admin page and then take a snapshot. Please make sure that you do not select to include the memory state. Linux does not tend to like that. In fact if you can I would even suggest to power down the notes. Just to make sure you have a clean state. It makes it easier in case of a recovery scenario.
OK lets start.
First to do the backup, I actually have not had time to do much modification inside yet but to show you where to do the backup.
Go to the Administration tab and then Policies on the left. Highlight the policy you want to export and click the export button.
Next you go to the Alerts tab and on the left select Alert Settings and then export your Alert definitions, Symptom definitions and recommendations. (Similar to policies).
Finally under the Dashboards, go to Virews on the left and export these also.
Next it is time to go to the admin page (https://fqdn/admin). I suggest to first stop HA if you are using it, and then stop the cluster to take snapshots.
Now it is time to take the snapshots. Take snapshots of all appliances. Just in case you have to go back. I powered my appliances down for this. It is not needed but better safe than sorry. If you do not power down the appliances then make sure you do not take a snapshot of the memory state (as mentioned above).
The power down sequence is as follows:
- Application Remote Collectors
- Remote Collectors (Can be done at any time)
- Data nodes
- Master Replica
- Master node
Once done you can start the nodes again. They should be started as follows:
- Master node
- Master Replica node (if you left HA on)
- Data node(s)
- Remote Collectors
- Application Remote Collectors
Log back into the admin interface (https://fqdn/admin) of the master node. Now it is time to run the upgrade checker.
Running the upgrade checker
This step is optional but I do recommend running it. The upgrade checker is a small .pak file that you can download from my.vmware.com. It basically goes through your vROps nodes and checks that all is ok before you go all in on the actual upgrade.
Once you are inside the admin page on the master node go to Software Update and select “Install a Software Update”
Now follow the steps through the wizzard, First Browse and select the update. (It is normally called something APUAT and is not very big). Click on Upload to stage the pak file.
Click on next and accept the EULA.
Then click through the Update information. The important part here is that the package is read only and does not change anything.
Click on next and then on install to install the Software Update.
The installation process starts and you can follow the steps on the screen.
After the installation is finished you should see everything at step 9 of 9 and with status completed.
Now to check the Upgrade you go to Support and select Support Bundles. You should see a bundle for each of the analytics nodes. Click on each bundle and the green download arrow to pull the bundle down.
Once done unzip the file and open up the apuat data folder and then report and click on the index file.
This report has changed a bit since previous versions. In Previous versions you got more of a breakdown of things, now you basically get a report like the one below.
In my case I do not have much to worry about. I know the nodes are in offline state because I put them there, and I don’t need to worry about the validation Product version since I am going from 8.0.1 to 8.1.
So having done this, it is time for the upgrade itself. Of course if you find issues, then you need to rectify these first.
Installing the upgrade
So like before it is back to the Software Update page and selecting Install Software Update. Only now you should select the upgrade itself. Upload the pak file again and then click on next. The steps are almost the same as above but you need to check the “Reset default Content” when doing the upgrade.
Checking the Reset Default content box brings up a warning. Click the Checkbox to continue.
After the EULA agreement page you will be presented with some warnings. Before installing the upgrade please make sure that:
- you have snapshots of the cluster (at least) before you begin the upgrade.
- Have run the Pre-assessment tool (as above). As I mentioned it is optional but understand that you may have to start all over again because the upgrade can break the environment. (Read: Do you have a backup? 🙂 )
- Remember to remove snapshots after the upgrade finishes as it may cause degradation to the system if you leave these in place.
Once you have made sure about these. Click next to go tot he “Install Software Update” Page and click install. Like before it goes through 9 steps on each node.
Now the waiting starts… It might be a good time for a coffee or… lunch. For me, in my environment this took about 1 hour to complete with a one site analytics cluster and one remote collector.
You will be logged out during the upgrade, if you are not logged in again after 5 minutes then manually refresh the browser and login again to follow progress of the installation.
At the end vROps is put back online and you can see the version is 8.1 on the right side here and running in HA mode.
Checking that all is ok afterwards
After you login to vROps again you will be asked to join the CEIP. I leave it up to you whether you do so.
I tend to check if the integrations are working under the Administration tabs, that the environment looks ok. Also are you receiving alerts (You might want to clean up the old cancelled alerts).
Next part is upgrading the Application Remote Controllers if you use that feature. That is in Part 2.