Sorry Ed but this doesn’t wash. While
many of the questions on this list deal with things the questioners don’t
know that doesn’t excuse the product from not working the way it is
supposed to. That is to say if the product is supposed to vault in a certain
way and you follow the documentation and it still doesn’t do it the way
it is written up and you are told to wait for the next MP or release that is a
problem in QA rather than with the end user. The discussion in this thread
wasn’t about the day to day “how do I….?” but rather
with the defects seen in 6.0. Why did it take them until MP4 until they got
one that seemed worth the risk of installing? Why has MP5 which should be the
BEST to date seem to have introduced so many new issues?
From:
veritas-bu-bounces AT mailman.eng.auburn DOT edu
[mailto:veritas-bu-bounces AT mailman.eng.auburn DOT edu] On Behalf Of Ed Wilts
Sent: Friday, August 17, 2007
11:42 PM
To:
veritas-bu AT mailman.eng.auburn DOT edu
Subject: Re: [Veritas-bu]
Netbackup 6.5 - Anyone actually using it?
> Their QA really dropped the ball with this release.
If by “this”
you mean 6.0, you may possibly be right. With 6.5, that has yet to be
proven.
I feel rather sorry
for the NetBackup QA group though – they have a *huge* challenge testing each release
because there are so many variations possible in customer configurations.
Dozens of client operating systems and versions, varying sizes from small to
20+ media servers and combinations like you wouldn’t believe – a
large mix of master and media servers plus all of the clients and a boatload of
options. SSO was originally for 2 or 3 media servers and not too long
after that you’ve got customers with 25-100 media servers. Some
customers have environments that meet best practices but many don’t.
Some customers do stuff that isn’t supported – a lot don’t
even know how to manage a DNS at all. Some customers have huge databases
and some have a billion files (we’re well over a billion here). It’s
impossible to QA all of the combinations and long gone are the days where a
vendor could test every possible combination before certifying. Some
write images to disk, some to tape, some to multiple simultaneous tape
libraries, and some to disk that emulates tape. Some people throw encryption
devices in the middle and some throw miles and miles of network between their
media servers and their tape drives. Sometimes they exceed the specs
doing so (been there, done that) and then wonder with they’re getting
SCSI errors over fibre.
If they had to QA
every possible combination that they support, it would take so long that you’d
never see another new release. Sometimes they just have to code it, hope
it works, and fix the edge cases where it breaks. When you think long
enough about how hard of a problem it must be, you’ll realize it’s
a miracle it’s as good as it is. After all, if it was really that
bad of a product, why are we all still running it? If testing
is so easy and so cheap, why don’t we all have test environments of our own?
After all, we know what kind of workload we’re going to be putting
through it a lot more so than Symantec does so it should be easier for us.
> Maybe one day I'll run into the NB product manager.
There’s a lot more than 1
product manager – the product is that big. I drive by the developer’s
offices every day to work so I’ve met a few of them and a few of their
support folks and met more than 1 product manager. Generally, they’re
good people and do the best they can with the resources they have. Perhaps
beating them up is just too easy on the list.
…/Ed
--
Ed Wilts, RHCE, BCFP,
BCSD
Mounds View, MN, USA
mailto:ewilts AT ewilts DOT orrg