Roger,
"I could go on and on, so I will just say that reliable "dot zero" releases
are the exception rather than the rule, here in the real world."
I understand what you're saying and largely agree. But in TSM's case that
would be the x.x.0.0 release. And history has shown those to be pretty
much uniformly bad.
It's the reported poor quality of the maintenance releases intended to fix
THOSE problems - the x.x.x.0 releases - that are the real shame.
Thanks for the insights.
Tab
Roger Deschner <rogerd AT UIC DOT EDU>
Sent by: "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
10/30/2002 09:48 AM
Please respond to "ADSM: Dist Stor Manager"
To: ADSM-L AT VM.MARIST DOT EDU
cc:
Subject: Re: TSM reliability (was: tape missing under q libv
(HELP))
Tab, I NEVER trust versions of ANY software that end in zero, from ANY
vendor. Software is like fine wine. When it is brand-new, it has to be
aged a bit to develop its full quality. I do not trust software that has
not accumulated a few patches. Windows 98SE was so much more stable than
the original Windows 98. I'm starting to trust XP now that Service Pack
1 is out. I could go on and on, so I will just say that reliable "dot
zero" releases are the exception rather than the rule, here in the real
world.
I once worked for a software vendor. It never failed that we were
informed of a major bug within 30 minutes of irrevokably shipping a new
version, no matter how much testing effort we had put into it. Now as a
software consumer, I think back on those days as I put the brand-new
version that just arrived on my top shelf to await the first patches.
ITSM is no better or worse than any other complex software in this
regard.
Roger Deschner University of Illinois at Chicago rogerd AT uic DOT edu
============ "In theory, theory and practice are the same, =============
========= but in practice, theory and practice are different." =========
On Tue, 29 Oct 2002, Tab Trepagnier wrote:
>"4.2.2.0 is a really bad release to be running but I do not think it is
>your
>problem. 4.2.2.12 or 4.2.2.13 are pretty good."
>
>One reason why I'm still running 4.1.5.0 is that pretty much every
>"x.x.x.0" release of TSM since has had such problems that several forum
>participants have warned about its use.
>
>I've been using this software since ADSM V2, and as it "matures" it just
>gets scarier.
>
>I understand that a lot of new functionality has been added, and that it
is
>a complex product. I also understand that novice users - like I was in
the
>days of ADSM 2 and 3.1 - can hang themselves with the system.
>
>But 3.1 had problems that corrupted data until the M5 version. I don't
>think IBM *ever* got 3.7 really working right. TSM 4.1 seems to be OK,
but
>the horror stories of 4.2 probably comprise 1/4 of this forum's content.
>And now tales of 5.1 blowing up appear on this forum seemingly every
other
>day.
>
>In IBM's VRML notation, the expectation is that any x.x.x.0 is a
production
>release meaning that it has passed some sufficient level of testing to be
>"certified". Where the "L" field is non-zero, it is a patch - basically
a
>temporary hack.
>
>In the next month or so I am going to be upgrading my 4.1 system to
5.1(?).
>Even if I immediately upgrade to the latest maintenance, patch, hack,
etc.,
>will I be placing my data at risk from internal corruption simply due to
>poor QA? My little company has almost 1/2 million dollars invested in
TSM
>hardware and software. I'm sure I can get a high-maintenance backup
system
>that corrupts data for a lot less money than that.
>
>Venting...
>
>Tab Trepagnier
>TSM Administrator
>Laitram Corporation
>
|
