"Install Updates" Process - Terrifying!

[Paul Hutchings]

Is it just me, or is the "Install Updates" process on a P4000 management group pretty damned terrifying?

Our 9.0 system was due some updates so I did them this morning, only the second time I think I've done so.

To be fair, it all seems to have gone just fine, but it's a little unnerving to be hit with approx 80 emails, one for each warning and alarm generated by a node going up and down just as would happen if you pulled the power or "something bad" happened.

Just an observation and maybe a little feedback if anyone from HP is reading that some kind of "logic" to minimize the notifications during the update would do wonders for my heart rate :-)

[Dirk Trilsbeek]

yes it is. Not only is it a "fire and pray it works"-process, starting such an upgrade using a management client and relying on that client to stay alive during the whole process is somewhat dangerous. This is not a 1000 â ¬ NAS with low impact on a companys infrastructure, this is an enterprise grade SAN some companies (like us) have their whole VM infrastructure based on.

Mix in that possible problem with changing MAC addresses and disabled clusters (completely ridiculous) due to the then invalid license and you know why we didn't touch the SANiQ 9 update yet.

[Bart_Heungens]

This is quite strange since I did already several upgrades on 50+ systems from 8.0 to 8.1 to 8.5 to 9.0 and 9.0.1 without any issues. The mechanism in CMC 9.0 makes life even much easier than before...

Off course the client must be stable and running during the upgrade but that is for every firmware update that you are doing.

And if you get too much emails, I know that I am upgrading so I can expect that nodes are rebooring. This is a proof that alerting is working... Too me at least...

[Paul Hutchings]

I think the upgrade process is obviously as simple/convenient as it gets i.e. hit "Go" and it just goes off and does it.

I take the point about the alerts showing that the alerts are working, but equally you do get a bit "boy who cried wolf" in that whilst you're being bombarded with "routine" alerts caused by the upgrade, there's no easy way to see any alerts which are NOT caused by the upgrade.

[Bart_Heungens]

Offcourse you are right what you say... However if I take a look at patch management on server level which requires a reboot, or the upgrade of a full blade enclosure, SIM also generates quite some alerts.

In that case, you can stop monitoring inside SIM for the upcoming hour or so...

Maybe this should be something to come in the upcoming releases of P4000?

[Paul Hutchings]

Be nice to know what is on the roadmap for P4000, nice product but always seems a little lost/forgotten in a company the size of HP :)

[Jitun]

Its a work in Progress, 9.0 was quite different from previous versions.
I would expect things to be ironed out in the future versions.