I was wondering if people are using this for managing EMC sans. I
know a few companies which dropped the product after realizing that it
was impossible to provision multiple luns easily and they still needed
to buy EMC storage scope to look at the performance of the SAN. They
also had serious problems with stability (2005 had critical system
patches coming out once per month) and it was, well, dog slow.

But with HP behind them now, they could have cleaned up their act a
bit.

On 25 Jul 2006 10:23:13 -0700, gregwallach (AT) yahoo (DOT) com wrote:

Only minimally. We have been looking at them for an XP SAN (HP
branded HDS) but the cost if hard to justify for what it does.

With a good perl coder you can get all the same information from the
point tools' cli.

Managing SAN's is still a dark art, and probably will be for another
2-3 years.

~F

We'd found that the our cost per lun was actually 25% higher using
their product, provisioning time was much longer, and the customer
support was, well, horrible.

It would take a week to get a reply, then the reply would be only
asking for a database dump, which would then take 6 weeks to diagnose
(most of which ended in more critical patches). It was shocking.
Our total up time for the application was less than 30%, by the time
the patch would come, everyone was using native tools anyway.
Faeandar wrote:

gregwallach (AT) yahoo (DOT) com wrote:

I'm going to hurt myself laughing.

The storage reporting still doesn't support Windows clusters. My guess
is that HP doesn't expect any SAN users to actually cluster their
Windows systems. We had the same problem with HP Storage Area Manager
but they eventually hacked in some code that sometimes gets it right.

HP promised us a fix last July (2005). Then September. Then January
2006. The last date I heard was Februrary, 2007.

HP, in my opinion, doesn't understand storage reporting requirements at
all and has not shown any ability to deliver on code development to
support even obvious functionality.

.../Ed

Ouch, yes I remember that too. I just had really really hoped they
had fixed it. We had to provision to each member of the cluster
seperately and a seperate step for each HBA in each machine. So a
dual fabric cluster architecture would require you to go in 4 seperate
submissions, each taking several steps. And each LUN would require a
seperate process. So for 20 luns, you have to single step and submit
80 different provisioning jobs, each one taking a LOT of time to
accomplish. Take that vs doing it through clariion native tools where
you can create 20 luns in one action (you simply put "20" luns in the
number of luns you want to create) and then assign them all to a
storage group in a second action. 10 minutes vs several hours.

That was the reason we dropped it. We realized even if we got it
working, it would be soo expensive to maintain it would bankrupt our
budget and we'd have to hire twice the staff because provisioning would
take so long. Oh well, pretty interface.

Greg

Ed Wilts wrote: