Just got back from my holiday - first thing to check is our VCOPS environment of course (and secondly, start writing these kind of posts ). One thing that always attracs my attentions, that 'purple'-alert message in the top-right called the "Admin Alert":
There was a nice question on the VCOPS forum today about having a big screen and rotating through custom dashboards automatically. Custom dashboards is a 'Advanced/Enterprise' feature, and I think you cannot live without them. You can make them switch to other dashboards automatically.
In the release notes for VCOPS 5.7.1 one of the fixed issues for VIN is:
"vCenter Infrastructure Navigator adapter discovers too many resources The vCenter Infrastructure Navigator adapter was configured to discover and calculate resources related to obscure open TCP ports. This configuration increased the load on analytics processing as well as on the storage sub-system that stores this data."
By default your session times out from the UI (both the vsphere ui as the custom ui) after 30 minutes of inactivity. This can be annoying when you're running a big screen display showing your favorite widgets. Quite a lot of articles excists around this so more than enough resources can be found including an official KB from VMware: KB2015135.
However, I've found VMware replaces the web.xml with the default timeout during .PAK updates (like going from 5.6.0 -> 5.7.0 -> 5.7.1). Now I would not recommend replacing the altered previous web.xml since it can hold new settings. But I'm getting tired of getting into vi, edit two files, on all VCOPS servers and repeat that after every update. So here is a small script which does a 'sed' and replaces whatever value in the session-timeout to '-1'. It creates a backup, replaces the value and restarts the UI interface, both vsphere as custom.
The ability to schedule reports in VCOPS is quite cool. Taken from capacity IQ these reports can be run weekly or monthly on a specific time and day of week allowing you to show the report as CSV (to process inside a spreadsheet) or as a readable PDF document. Once SMTP is configured you're able to send the report to multiple users (you, your manager) using comma (,) or semi-column (;) separated e-mail addresses. Also: you can schedule reports on ANY level on ANY object (World, Folder, Datacenter, Cluster, Host, VM, Datastore), and here, the management of all schedules, can become a mess since there is no overview on where schedules are set.
Here is a little query you can run on the UI vm (this is where CapIQ is running).
I'm quite excited. Just found on the VMware website that VCOPS for Horizon View is GA. Together with this release there is also a new version for VCOPS -> 5.7.0.
One of the cool things with this release is the .pak file usage. No longer you'll need an additional Windows instance (or multiple when having multiple PODs) for passing the data to VCOPS. VCOPS is capable of handling the datastream direct from the broker agents and client agents. Installation is a breeze - just install the .pak file on the /admin part on VCOPS, just like doing the update to VCOPS 5.7.0.
VCOPS includes a lot of percentage metrics already. However, if you look at for example CPU Co-stop, System, Idle - they are only available in milliseconds (ms). Now there is a small catch you'll need to take into account which I'll show you in this blog post.
I'm in the process of re-registering my vCenter servers to VCOPS (with another username). Once in a while I have this error message:
Unable to connect to vCenter Server. Details: Failed to get RMI connection. Ensure that vcops service is running and retry the operation. Registration failed for VC : ( vCenter Server address: https://bla-host/sdk, vCenter Server username: bla-user, vCenter Server name: BLA blabla ) RMI interface unavailable