During our patch Tuesday window of late 10/14/2014, and early
10/15/2014, we did a careful test of upgrading another Zebi storage
system around 12:30AM, for Zebi2. This unit handles storage for
various VDC and SV, both managed and customer managed systems.
The previous failure seemed tied to snapshots, and this unit was not
doing any snapshotting, so we felt a controlled test would be safe.
So far to date, one Zebi upgrade went smooth with no issues, one went
horrible for many hours, and now the third one was done.
This time things failed in a simular fashion to the Zebi4 problems
over last weekend but not due to the same reason apparently since
there weren't snapshots in use on the system.
We didn't have the redundant controller upgraded, so once we determined things
weren't running correctly, we failed back to the other controller, and
all our metric measurements returned to normal.
We have noticed a few VMs panic'd due to the increase in latency, and we
are scanning through the lists and fixing them up.
We apologize for this disruption, our vendor thought the problems with
Zebi4 were isolated to that unit, but obviously this isn't the case. We'll
continue to work with them to ensure we won't have any further
distruptions like this due to bad software/firmware upgrades, some of
which are security fixes.
If you still have any problems or questions please let us know at
[log in to unmask], or call us up at 612-337-6340.
Thanks!
--
Doug McIntyre <[log in to unmask]>
-- ipHouse/Goldengate/Bitstream/ProNS --
Network Engineer/Provisioning/Jack of all Trades
