"Hansen, Heidi" <[log in to unmask]> wrote:
>In case your shop has installed Powerpatch 5 (or plans to) you might
>want to review this.
>
>=============================================
>
>Dear Valued HP3000 Customer,
>
>Earlier this week, HP discovered a problem with the MPEKX79 patch. The
>MPEKX79 patch was included in 5.5 Powerpatch 5. As a result, we placed
>Powerpatch 5 on temporary production hold, and we have stopped
>distributing the MPEKX79 patch...
...
> A Problem Description:
> =====================
> MPEKX79 was a performance enhancement patch, intended to improve
>the file system performance when a file "cuts back" its EOF (in other
>words, returned unused file space). This file system function is most
>typically used by a DBUTIL ERASE command.
> Intermittently, following an ERASE command in DBUTIL, data is
>not being completely erased from datasets. This results, for details,
>in image error -3, for example:
> TURBOIMAGE ERROR AT $00016710; RETURN STATUS = -3
> DBPUT, MODE 1, ON RESIDENTIAL OF realty
> MPE FILE ERROR 0 RETURNED BY FREADDIR ON ROOT FILE
> For master datasets, dbputs fail with DUPLICATE KEY VALUE IN
>MASTER (error 43).
> Attempts to fix the databases with adager result in the data
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Ahem... "Fix what?"
>which should have been erased being "added to" the data which was put
>into the database follwing the erase.
Ahem (again)... "What kind of fix?" (Adager has A LOT of "fix" functions)
Different "breaks" require different "fixes".
Adager has a large inventory of EXPLICIT "fix" functions (the obvious ones)
as well as a large number of IMPLICIT "fix" functions (some of which, in
fact, are quite subtle).
In this particular case, as HP's Jon Cohen so nicely explains, what we have
is a dataset that has not been completely erased by DBUTIL/erase (i.e., the
"erased" dataset contains some garbage instead of totally empty space).
- One way in which this condition manifests itself is via the TurboIMAGE
error mentioned above. In other words, DBPUT gets fooled by the garbage.
- Another way in which this condition manifests itself is via Adager's
apparent "inability to fix" the database (perhaps via some "obvious fix"
function?).
So, what to do? Once we understand what the problem is (an incorrectly
ERASED dataset), the solution to the problem is simple:
Use Adager's super-fast "ERASE DATASET" function (which does NOT
contain the word "fix" in its name) to "fix" the problem. Adager goes
through every record and actually writes ZEROES to it, explicitly,
without depending on any file-system magic.
Unfortunately, if you trusted that DBUTIL/erase had, indeed, erased every
dataset in your database, and if you then added new entries to datasets
that were veritable Swiss cheeses (full of holes and random bits of
matter), all bets are off. The only way to "recover" from this mess is to
restart from scratch.
Not very palatable, I know. But not even Adager (with its legendary
recovery capabilities) can help you in this horrible case.
Of course, you can (and should) use Adager to examine every path in the
database to discover which datasets have garbaged data. You can then use
other Adager's "fix" functions to ATTEMPT to get out of the mess. You can
try to repair chains and/or to do other therapeutic operations (which, in
fact, could include -- as a radical "amputation" step -- totally erasing
the offending datasets so that you can begin with a clean slate).
Yuck!
The proper solution, of course, is to follow Jon's advice and to get a
patch for DBUTIL/erase. Read on:
> While the vast majority of the problem reports received stem
>from DBUTIL ERASE, there are a few reports of other usages hitting the
>same problem.
>
> What this production hold means to you
> =====================================
> If you have not yet installed patch MPEKX79 or 5.5 Powerpatch 5,
>then please do not install it. If you have installed Powerpatch 5 or
>this patch, please do the following:
>* please make all efforts to avoid entering the DBUTIL ERASE
>command.
>* please make immediate arrangements to obtain the fix (see
>below).
>
>
>What are our plans to lift this Production Hold
>=============================================
>We have created a new patch, MPEKXJ3, which essentially removes the
>MPEKX79 functionality. This patch is currently being tested, and we
>expect to "generally release" (GR) this patch by early next week. At
>that time, it will be available for download from the Electronic Support
>Center. We are fortunate that this problem is easily reproduced-we
>should be able to verify our MPEKXJ3 fix quickly.
>We have incorporated MPEKXJ3 into a new Powerpatch, which will be called
>5.5 Powerpatch 6. Powerpatch 6 is currently under test, and if all goes
>well, it will be available for customer shipments around October 15,
>1998. You will be able to obtain 5.5 Powerpatch 6 via the normal
>channels.
>
>What are our plans to communicate resolution information to customers
>====================================================================
>There are many ways for customers to obtain reactive patches and
>Powerpatches from HP. Fortunately, for the most part, we are able to
>capture information about who has obtained these patches. We are
>drafting a letter that will be sent to those customers. This letter
>will contain the same information that is in this memo. We expect that
>letter to be mailed as soon as possible.
>
>=============================================
>Jon Cohen
>Commercial Systems Division
>mailto:[log in to unmask]
>
>
>=============================================
I want to thank Heidi Hansen for having posted Jon Cohen's message. I want
to thank HP for its prompt response to the challenge. I want to urge
everyone to take immediate action regarding the patches involved.
As "consolation", please remember that these kinds of fundamental bugs are
relatively few in the HP3000 world.
_______________
| |
| |
| r | Alfredo mailto:[log in to unmask]
| e | http://www.adager.com
| g | F. Alfredo Rego +1 208 726-9100
| a | Manager, R & D Labs Fax +1 208 726-2822
| d | Adager Corporation
| A | Sun Valley, Idaho 83353-3000 U.S.A.
| |
|_______________|
|
