Just a quick post about something I experienced at a client, with ESXi 6.0 hosts, today:
If you have trouble performing VMware snapshots, and see a msg.snapshot.error-QUIESCINGERROR error, check the host time settings and NTP.
In this case, snapshots of VMs located on other hosts in the cluster were fine, but once a VM was moved to the new host, snapshot operations failed after an hour or so.
It turns out a new host in the cluster was not properly set up to use NTP, and time drift between the host and the vCenter caused the snapshot failures. Correcting the time on the host and configuring NTP resolved the issue.
Always remember: If the problem isn’t DNS, it almost certainly is NTP.
What really stands out after having time to digest it, is how VMware and VMworld felt energetic again. The keynotes were good, especially on day 1. That keynote is probably the best VMware keynote I’ve ever seen. Everything VMware has been talking about for years, perhaps without actually being able to get the message clearly across to everyone, seems to click neatly into place now. There is a vision now, a vision you can actually relate to, and believe in. Even the tagline be_tomorrow, makes more sense now.
I don’t know if anyone else has noticed, but it feels like something has changed internally in VMware in the last year and a half or so. There seems to be a new drive, a clearer focus. To be frank, it feels fun again, something it really hasn’t for the last couple of years.
As per usual, my biggest take-away from attending VMworld is networking and talking to real life people — The same people I “talk” to virtually all the time. I even met quite a few new people this year, and that’s always awesome!
My Session highlights:
I attended a few sessions too, and would like to highlight two of them:
Both these sessions were awesome. If you work as an architect and haven’t had a look at VMware Validated Designs yet, drop what you’re doing and go have a look. Right now.
VMware Cloud on AWS was a little light on details (naturally, since it’s not even released/available yet) but for now this one gave a really good overview of what is is, and perhaps more crucially what it isn’t.
As a VMUG Leader I attended the VMUG Leader Lunch, which had an awesome Q&A session with Pat Gelsinger and Joe Baguley —That session should have been recorded too.
I met up with Ed and Chris, all three hosts of vSoup were finally in the same city at the same time, for the first time since 2011! We recorded a quick vSoup Podcast, and even got Emad Younis as a surprise guest. That recording is still unreleased, hopefully we can get the audio cleaned up and it published pretty soon.
VMworld 2016 has left me happy. Happy with the direction VMware is going, happy with the event and really happy I wore that shirt for the vRockstar party. As a side note, my FitBit logged 108,427 steps while I was in Barcelona, not to bad for under a weeks worth of conference.
Now, can someone tell me where VMworld Europe 2017 will be held?
For some time now I’ve been advocating the usage of VCSA instead of the traditional Microsoft Windows based vCenter. It has feature parity with the Windows version now, it’s easier to deploy, gets right-sized out of the box and eliminates the need for an external Microsoft SQL server.
One of the questions I often face when talking about the appliance, is how do we handle backups? Most customers are comfortable with backup up Windows servers and Microsoft SQL, but quite a few have reservations when it comes to the integrated vPostgres database that the VCSA employs. One common misconception is that a VCSA backup is only crash-consistent. Thankfully vPostgres takes care of this on it’s own, by using what it calls Continuous Archiving and Point-in-Time Recovery (PITR).
I essence, vPostgres writes everything to a log file, in case of a system crash. Since this is done continuously, every transaction that should hit the DB is recorded and can be replayed if required. From the Postgres documentation:
“We do not need a perfectly consistent file system backup as the starting point. Any internal inconsistency in the backup will be corrected by log replay (this is not significantly different from what happens during crash recovery). So we do not need a file system snapshot capability, just tar or a similar archiving tool.”
With regards to the VCSA, this means that your image level backups will be consistent, and there isn’t really a need to dump and export the vPostgres DB and then archive that. Yet another reason to switch to the appliance today!