1 person found this helpful
Upgraded from 6.4 to 6.6.
There's not much to state about the upgrade procedure. It took about two hours to do both the OS and application update combined, which went without a hitch.
My main concern was whether the Horizon pak would carry through and work, which it did.
The new HTML5 interface is absolutely awesome - a vast improvement on previous versions. The layout is far simpler to navigate and makes a lot more sense. Even the Horizon dashboards seemed to jump out as a lot more useful, even though the widgets haven't changed.
The OOB dashboards are so much cleaner and useful than previously, and the intergrated management packs, like VSAN, is a very nice touch.
It gets a huge thumbs up from me at the moment.
1 person found this helpful
I'm sharing my customer experience here in Brazil. They were running vRealize Operations Manager 6.4 with three nodes: two data nodes with HA enabled (one Master and one Master Replica) and one data collector. Their environment has about 2.000 VMs being monitored by vROPs, and each data node has 780GB.
First we checked all appliances for free space. Then we took cluster offline, powered-off all appliances, took a snapshot, then we started the appliances and the cluster. Bear in mind that you can be surprised by a "File System Check". Please, add to your planning one-two extra hours.
In our first try, the upgrade of operating system failed. We were using Internet Explorer, the upload of the OS-VA PAK file finished successfully, but we got stuck in web browser about one hour with no progress. So, we went to Google Chrome and canceled the installation. We checked all appliances, and we found that the PAK file was distributed just to one data node.
We started a new installation using Google Chrome, and the VA-OS PAK was distributed to all nodes, and the upgrade took about 40 minutes to finish with success !
Then we started the vRealize Operations Manager upgrade. The VA PAK file was uploaded successfully, and the whole process took about one hour to finish with success.
So, when everything finished we deleted all snapshots and checked all appliances for free space.
2 people found this helpful
We upgraded from 6.5 to 6.6 in our test environment first to validate all solutions were playing nicely with each other. After a few days with no major issues, we were confident to roll this out to our Production cluster. Following best practices, we took snapshots being sure to uncheck the virtual appliance memory and quiesce the guest operating system options as well as backup all customized content.
The update was a breeze and there were no hitches with the installation procedure. It was a very straight forward procedure! I was also very pleased to find that everything including views, reports, dashboards, super metrics, and policies were still left intact. The only things we had to redo were to re-permission groups to allow viewing of the Home Page and re-share dashboards since all groups inherited a lot of the unassigned dashboards which we didn't want for some of our users. So my advice is to take note of how you share your dashboard prior to upgrading.
Our main motivation for upgrading was to take advantage of the newest dashboards and HTML5 UI which are really sleek and very easy to navigate! Honestly, I get quite frustrated at any big changes to the UI, but it took me only a few hours of poking around to discover how the new layout was reorganized. My manager thinks the same and really likes the new "Workload Balance" overview on the Home page, but I thought it could have been combined with the dashboards. Honestly, I'm not a huge fan of the new Home Page as it doesn't seem to allow much customization and if you choose to disallow the permission altogether, the user is faced with the error, "Missing Privileges: Home/View Home Page" when they logon since it's the first landing page. Also, there is no obvious way to hide the "Recommended Actions" which could be confusing to users who don't have permissions to certain objects in the environment. What I am excited about is the PDRS (available on 6.5 and onward) which we are taking advantage of starting today in our development environment! Hopefully we can avoid contention, better balance our clusters, and take the guess work out of learning the varying workloads we support.
Further update: Having used 6.6 at a couple of sites over the past week now, I've only found one area that I don't like - sharing dashboards.
This seems very buggy at the moment and is a bit of a nightmare to customise. Pre-6.6 this worked very well - remove dashboards from the everyone group and then share them with particular groups only knowing that they would not be able to see anything else. With v6.6 this doesn't seem to work very well. Firstly the dragging and dropping mechanism fails quite often when adding to groups, and the interface seems to stop working often regardless of what browser is used. It also fails to properly list the dashboards and is very random in what is displayed.
I've had to create some custom access profiles at each site, and at both the issues were consistent. Even worse, when logging in with a test account, we could usually find a way to view dashboards that we had specifically not shared with a group, or even disabled outright.
My other gripe in regards to this level of customisation comes with creating custom roles. We have certain groups that we don't want to see the Home tab, who we would prefer to default straight to the dashboard page at logon; however, it doesn't seem like that is possible. If you deny permissions to the home page, the user simply gets a permissions error when they logon. It would be grand if there was somewhere where we can set the default start page based on user group to get around this. If that already exists, I haven't found it yet.
I'd also like the ability to re-order the dashboard groups too.
Otherwise, everything else so far is a huge improvement on previous versions.