Msg 2525 on DBCC NEWALLOC of empty 2Gb/1Gb SQL6.5SP3 database

Msg 2525 on DBCC NEWALLOC of empty 2Gb/1Gb SQL6.5SP3 database

Post by MPoppers » Wed, 31 Dec 1997 04:00:00



After spending a few days loading data into a database (one
2048Mb data segment, one 1024Mb log segment, each exclusively on
its own device) on a SQLv6.5 (SP3) Server, I got around to
setting up some maintenance tasks w/in SQL Executive, one of
which, among other options, runs SQLMAINT with the -CkAl option.
The allocation error I get (which also appears as a "Msg 2525"
after the "INDID=2" line for 'sysobjects' when I run DBCC
NEWALLOC on either it *or* on a totally empty, "fake" db of the
same spec.s) is "Table Corrupt: Object id wrong; tables: alloc
page 0  extent id=208  l page#=208 objid in ext=14 (name =
sysreferences) objid in page=14 (name = sysreferences)objid in
sysindexes=1 (name = sysobjects)".  Any info. which y'all might
have on whether or not this error is significant would be
appreciated.

Given that I had previously created a 1000Mb/512Mb database on
this server and that -CkAl ran clean on it, is it possible that
one data and/or log device cannot be more than a certain size and
that I exceeded one or both max.s?  I intend to create another
"fake" db with two 1000Mb data segments (create one 1000Mb, then
extend to 2000Mb) and/or with two 1000Mb data devices in order to
see what happens with a subsequent -CkAl, and I guess I'll apply
SP4 iff neither I nor y'all have come up with a workaround.  TIA!

--

% Peritus Clavis %  425 Park Avenue #12-07 =**= New York, NY
10022
% Machinae Vivit %  tel: 212/836-8299 =***= fax:
212/836-6499,-8689
%%%%%%%%%%%%%%%%%%       e-mail: see message header, and tweak!

 
 
 

Msg 2525 on DBCC NEWALLOC of empty 2Gb/1Gb SQL6.5SP3 database

Post by MPoppers » Wed, 31 Dec 1997 04:00:00



> ...Given that I had previously created a 1000Mb/512Mb database on
> this server and that -CkAl ran clean on it....

Incorrect: I had never created maintenance tasks for that db.  I just
ran DBCC NEWALLOC on it and got the same 2525 error, so I assume I'd
get the same error (minus the error#!) if I did a SQLMAINT -CkAl on
it.  This, of course, only enhances the problem from my perspective
(esp. as we want to use db.s which will eventually be a lot bigger
than 2Gb between data&indices), so your thoughts are [still]
earnestly requested [while I find and download SP4!].  TIA!

By the way, since I've seen mention of ???Gb-size databases in use
under SQLv6.5-pre-SP4, am I safe in assuming that something about the
environment/config. may be behind the problem?  We're hosting the
Server on a Compaq Proliant 7000, using a SmartArray card to control
four pairs of 9Gb disks.  The first disk [pair] is partitioned into a
2Gb FAT device, a 2Gb NTFS device for the OS (WinNTv4SP3) and other
software (e.g. SQL Server), and another NTFS device of size 4Gb (used
as a work area); the second and third disk [pair]s together form a
18Gb NTFS device used, primarily, for data devices; and the fourth
disk [pair] is partitioned into a 9Gb NTFS device used, primarily,
for log devices.
--

% Peritus Clavis %  425 Park Avenue #12-07 =**= New York, NY  10022
% Machinae Vivit %  tel: 212/836-8299 =***= fax: 212/836-6499,-8689
%%%%%%%%%%%%%%%%%%       e-mail: see message header, and tweak!

 
 
 

Msg 2525 on DBCC NEWALLOC of empty 2Gb/1Gb SQL6.5SP3 database

Post by MPoppers » Tue, 06 Jan 1998 04:00:00



> ...am I safe in assuming that something about the
> environment/config. may be behind the problem?....

Actually, I'm safe in assuming that something configuration-wise was done
to the server, as a -CkAl on the "master" db, which last succeeded 11 Dec
1997, after which two other people (one of whom I'll be quizzing shortly
about all this!) did a reinstall of SQL Server (and SP3) on the machine
in question (see prior post for some details about it), which (of course)
blew away "msdb"/the maintenance tasks.  I just recreated the "master"
daily-maintenance task (SQLMAINT -CkAl...), as well as DBCC NEWALLOC,
show the same type of "Table corrupt" errors.

One of the main reasons for the reinstall was to change the sort order to
'nocase' (ID=42) from whatever the default was (perhaps 'careless_34',
ID=49?) when I installed SQL Server.  However, we have another 'nocase'
server, and I don't see any errors when I DBCC NEWALLOC its "master" db.
Can anyone think of some other server-wide configuration change that
would be behind the "Table corrupt" [spurious?] error(s)?  TIA.

--

% Peritus Clavis %  425 Park Avenue #12-07 =**= New York, NY  10022
% Machinae Vivit %  tel: 212/836-8299 =***= fax: 212/836-6499,-8689
%%%%%%%%%%%%%%%%%%       e-mail: see message header, and tweak!

 
 
 

Msg 2525 on DBCC NEWALLOC of empty 2Gb/1Gb SQL6.5SP3 database

Post by MPoppers » Tue, 06 Jan 1998 04:00:00




> > ...am I safe in assuming that something about the
> > environment/config. may be behind the problem?....

> Actually, I'm safe in assuming that something configuration-wise was done
> to the server, as a -CkAl on the "master" db, which last succeeded 11 Dec
> 1997, after which two other people (one of whom I'll be quizzing shortly
> about all this!) did a reinstall of SQL Server (and SP3)....

> One of the main reasons for the reinstall was to change the sort order to
> 'nocase' (ID=42) from whatever the default was (perhaps 'careless_34',
> ID=49?) when I installed SQL Server.

Actually, as per the Event Log, I determined that the old sort order was
ID=40.

Quote:>                                                             However, we
> have another 'nocase'
> server, and I don't see any errors when I DBCC NEWALLOC its "master" db.
> Can anyone think of some other server-wide configuration change that
> would be behind the "Table corrupt" [spurious?] error(s)?  TIA.

I should add that a SP3 Hotfix was applied (increasing the server version to
6.50.276) and that the changed parameters at the time (as per the quizzed
co-worker's documentation) were
        free buffers: increased from 3276 to 14336
        LE threshold maximum: increased from 200 to 1500
        LE threshold minimum: increased from 20 to 150
        LE threshold percent: increased from 0 to 10
        memory: increased from 65536 to 286720
        network packet size: decreased from 4096 to 512
Some time between then and last week (call it the end of Dec.), I doubled
"open objects" from 400 to 800.

I'll be installing SP4 now on the very doubtful chance that the Hotfix was
screwy.  If that doesn't work, and not having heard from y'all, I'll give
Microsoft a call.

--

% Peritus Clavis %  425 Park Avenue #12-07 =**= New York, NY  10022
% Machinae Vivit %  tel: 212/836-8299 =***= fax: 212/836-6499,-8689
%%%%%%%%%%%%%%%%%%       e-mail: see message header, and tweak!

 
 
 

Msg 2525 on DBCC NEWALLOC of empty 2Gb/1Gb SQL6.5SP3 database

Post by MPoppers » Wed, 07 Jan 1998 04:00:00



> ...
> I'll be installing SP4 now on the very doubtful chance that the Hotfix was
> screwy.  If that doesn't work, and not having heard from y'all, I'll give
> Microsoft a call.

It didn't work, so I opened up a case w/ Microsoft.  Sure enough, the sort-order
change (which, naturally, forced a rebuild of "master"'s indices) triggered the
index-related problem.  This bug is still around in SP3, but supposedly fixed in
SP4 (even though there's not a whisper about it in either SP4's FIXLIST.TXT or
in any publicly-available knowledge base that I checked...which is why my
per-incident charge will be refunded!).

Thanks to y'all for responding...not.

--

% Peritus Clavis %  425 Park Avenue #12-07 =**= New York, NY  10022
% Machinae Vivit %  tel: 212/836-8299 =***= fax: 212/836-6499,-8689
%%%%%%%%%%%%%%%%%%       e-mail: see message header, and tweak!

 
 
 

Msg 2525 on DBCC NEWALLOC of empty 2Gb/1Gb SQL6.5SP3 database

Post by David Stol » Sat, 10 Jan 1998 04:00:00


I just had a similar experience, fortunately only the pubs and model
database came back with 2525s, so I just restored them.

I had a server crash (of unknown origin).  The master DB seemed to have
been corrupted, all the server config options were way out of whack (max
users=64K, etc).  Anyway, so I used setup to rebuild the master, recovered
it along with msdb - and all seemed to be OK.

That night's DBCCs uncovered the 2525s.




> > ...
> > I'll be installing SP4 now on the very doubtful chance that the Hotfix
was
> > screwy.  If that doesn't work, and not having heard from y'all, I'll
give
> > Microsoft a call.

> It didn't work, so I opened up a case w/ Microsoft.  Sure enough, the
sort-order
> change (which, naturally, forced a rebuild of "master"'s indices)
triggered the
> index-related problem.  This bug is still around in SP3, but supposedly
fixed in
> SP4 (even though there's not a whisper about it in either SP4's
FIXLIST.TXT or
> in any publicly-available knowledge base that I checked...which is why my
> per-incident charge will be refunded!).

> Thanks to y'all for responding...not.

> --

> % Peritus Clavis %  425 Park Avenue #12-07 =**= New York, NY  10022
> % Machinae Vivit %  tel: 212/836-8299 =***= fax: 212/836-6499,-8689
> %%%%%%%%%%%%%%%%%%       e-mail: see message header, and tweak!

 
 
 

1. Help! msg 2525 from DBCC fix_al

recieved the following from DBCC fix_al (dbname)
Msg 2525, Level 16, State 1
Table Corrupt: Object id wrong; tables: alloc page 63488 extent id=635041
page#=63504 objid in ext= -1532532493 (name = -1532532493) objid in
page=1532532493 (name = Individual)objid in sysindexes=1532532493 (name =
Individual)
MS PSS suggests restoring a known good backup and transferring the tables
from DB "A" (has new data) to DB "B" (restored DB with old data). Transfer
Manager has not work 100%.
Then I "BCPed" the table data out with the intent of dropping the table,,,,,
could not drop the table.
There has to be way to pull this off!
maybe scandisk for SQL? :)
Richard.

2. Upgrading to SQL Server 2000

3. CFV: comp.databases.ms-sqlserver

4. Help! msg 2525 DBCC fix_al

5. Progress MFG/PRO JOB

6. Binary Sort Order and Table Corruption (msg 2525)

7. SET NULL / SET NOT NULL

8. error msg 2525

9. Continuing Problem: Msg 2525 Table Corrupt Object Id Wrong

10. Binary Sort Order and System Table Corruption (msg 2525/605)

11. Continuing Problem: Msg 2525 Table Corrupt Object Id Wrong

12. TABLE CORRUPT, Msg 2525