In Part I of this post, I covered two of the four steps that I think are involved in doing a VMware Capacity Planner exercise. In this post, I will cover the other two and hopefully, you’ll find the tips found in these posts of some use. So, here goes:
Once the collectors can access everything and scans are completing in time, we’re ready for the next stage:
- Clean-up of Virtual Machines: While doing initial scans, any virtual machines in the environment can be discovered as well. That’s OK at the time of initial scans as gathering inventory for those machines could be useful for other purposes. However, as this is generally a consolidation estimation exercise, presence of those machines can seriously skew the estimates made by Capacity Planner. This is because it assumes the inventory to be physical and a “slice” of a CPU consumed in case of a virtual machine, will skew the calculations. For that reason, I filter them out at this stage and remove them from the collectors. An easy way to filter them out is to look for “Chassis” information and if the machine is from “VMware” then you could safely delete those machines.
- Uplink to Capacity Planner Dashboard: If successful with Inventory and Performance scans and with VMs removed, the collector is ready to have the scheduler set to “Active”. Before that step, the standard process should be followed to add the collector(s) to https://optimize.vmware.com Once done, I normally do a manual “Data Synchronization” first. This ensures that the link-up is successful. The first sync can also have a substantial amount of data so love and attention is critical. We want to ensure that all uploads are good and afterwards, the collector is empty. Check that the “Outbox” folder is empty under the install directory on the file system. The data uploaded can take a while to appear on the various “Dashboards” (up to a day!) so don’t panic if you can’t see anything there immediately after the uploads. If the collector is visible under “Company -> Collector Tools -> Configure Collectors” and data upload is successful, then it will appear sooner or later. To make extra sure, go to “Dashboards -> Collection Dashboard” and select either the “Inventory Uploads” or “Performance Uploads” tabs, to see progress. Please note: The times shown on the website are in PST.
- Scheduled Performance Scans: After a successful manual data sync, ensure that “Jobs” only has “Performance” and “Data Synchronization” tasks (both hourly by default) are enabled. Inventory doesn’t have to be run again, unless there is a change in target systems. Once everything is in place, the scheduler can be set to “Active” by unticking the “Suspend Scheduler” tick box.
The scans are kept running for two weeks for a “Consolidation Estimate (CE)” or four for a “Consolidation Assessment (CA)” engagement. During that time, systems can change in terms of access and/or could be decommissioned so a daily check on the collection activity is necessary.
The last step in the process is reporting. There isn’t much to say here as it’s a standard process and works reasonably well. I created my companies using version 3.0 of the product but 2.8 can also be used. Just be aware that menu items and their locations change based on the version.
- Test Reporting: After some data is collected on the dashboard site and has been processed i.e. after a week or so, it’s always a good idea to start creating the reports in a “test” capacity. If there are going to be issues, you’re better off finding out sooner rather than later. Issues can be anything from the rights to create reports to lack of certain data. The latter will be shown in the “Assessment Report” as detailed system information will be in the sheets later in the document. The main benefit is that once you’ve done this work, all you need to do at the end is to generate the final report.
I am glad I do this because that way, I found out quite early that creating a project in Version 3.0, did not allow me to add target machines to the assessment. I was able to figure out a workaround for it and the process is documented here but the point is that starting on the report early has many benefits so do it!
- Final Reporting: Once the amount of time has passed which is suitable for your assessment and if you’ve already done the homework by creating test reports, all that needs doing afterwards is to run the reports with the current information and deliver the reports to the clients (after double-checking, of course!)
Reports are quite easy to generate. They are PDFs so pretty much set text and format, apart from some of the fields that you can change. In a way, it’s a good thing as they come with all the usual disclaimers and next steps etc. However, there are times when one would like to have some flexibility to edit or at least eliminate unnecessary sections. In my case, I would’ve liked to be able to eliminate “Section 3” which deals with “Software Profiling”, which is more of a VDI assessment thing and I was only after Server Consolidation reporting.
Once the assessment is done and reports are generated, don’t forget to decommission the environment. That basically involves:
- Deactivation of Notifications/Subscriptions,
- Removal of collectors from the dashboard,
- If applicable, “hide” the company from the Dashboard,
- Uninstall of the data collector software from the collectors, and
- Shut down and decommissioning of the collector machines.
It’s important to do these steps while on the case as it’s easily forgotten once the reports are generated. There might be cases where the collection can be kept running for more confidence in data in which case, you don’t need to do these steps.
I think that covers pretty much everything I had in mind to document. I hope this will benefit someone (or even me!) in the future. As always, please do let me know if you have any comments and we might be able to improve the process even further!