On Wed, Jul 19, 2017 at 12:32:52PM -0500, Doug McIntyre wrote:
>On the early morning of 7/24/2017 starting at 12:30AM CDT we will be
>doing OS upgrades on the storage system (znas4) that some of our virtual
>machine hosting is stored on.
Apparently the difference in code revs did something unexpected, and a
few dozen VMs that were hosted on zebi4 rebooted themselves or
shutdown unexpectedly. That hadn't happened previously in testing with
the same mix of VM hosts and VM guest OSs. Zebi4 was running the oldest
version of code, I figure something between the two differences in
code revs is what did it.
I did do a staged upgrade, going to the last one in the code train running,
and then onto the current train and code rev, but that first upgrade
is the one that caused some issues. So, it is good that we are off
of that version, if that old version is the one causing issues and newer
versions don't cause problems like that.
We are complete with this upgrade for this unit. Everything is working
correctly and as expected for the final version of code. I'm seeing
better mean throughput and IOPS as expected, plus bug fixes since the
last code rev we were on.
We have gone through the VM lists and reviewed what should be running
and things are looking good as far as we can see. Let us know if something
isn't right at [log in to unmask]
We'll be monitoring everything as we normally do. If you have any
questions please let us know at [log in to unmask]
Thank you.
--
Doug McIntyre <[log in to unmask]>
~.~ ipHouse ~.~
Network Engineer/Provisioning/Jack of all Trades
