Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

dbcc tablealloc with fix

801 views
Skip to first unread message

denise

unread,
Mar 22, 2004, 5:44:34 PM3/22/04
to
We run dbcc nightly, not in single server mode and have a
lot of 2521 and 2546 errors. I'd like to run tablealloc with
fix on some of the tables. I need some directions on how to
do this, and how to tell if they tables are system tables.
Examples are object=z_audit_header, i_request_q,
c_labelchute_queue. Is it better to run it in single user
mode on all tables? Thanks very much for any help.

Bret Halford

unread,
Mar 22, 2004, 10:29:39 PM3/22/04
to
System tables all start with "sys" and have an object id < 100.

If you can, it is best to run with the database in single user
mode, but if you can't put the db in single user mode, it is
still better to run dbccs than not run them.

-bret

denise

unread,
Mar 23, 2004, 10:41:52 AM3/23/04
to

So do I just enter the command: dbcc tablealloc
(z_audit_header, full, fix) ? Will it take very long, or
affect users? Thanks again, I'm obviously new to this.

mikhail

unread,
Mar 23, 2004, 11:25:44 AM3/23/04
to

yes, the command is dbcc tablealloc(<table>, full, fix), if
you need to run it for a system table, you'll have to put
the database in single user mode. the running time will
depend on the table size, obviously, the larger table is,
the longer it will take to check it. as for the users, they
will be affected since dbcc will put a shared intent lock on
the table preventing users from changing it.

misha

denise

unread,
Mar 23, 2004, 2:09:46 PM3/23/04
to

Thanks for the help. I was hoping to run this with minimal
impact on users. We are up 24x7. Is it correct then that
since I run dbcc regularly not in single user mode these
"false alarm" errors will always be found?

mikhail

unread,
Mar 24, 2004, 7:45:43 AM3/24/04
to

denise, we get over 100 spurios errors every night due to
users' processing when we run dbcc. the only thing done for
these errors is to check if they occur on the same table and
page 2-3 nights in a row.

misha

denise

unread,
Mar 24, 2004, 1:46:43 PM3/24/04
to
This helps a lot, thanks!

Kambo

unread,
Mar 26, 2004, 11:58:12 AM3/26/04
to
We run the checkverify just after checkstorage if it reports any
faults. A tablealloc full fix can take a while. I ran one on a
sysobjects table that was reporting a fault and it's elapsed time was
35 minutes for a table with 4000 entries.

David L. Cherin

unread,
May 5, 2004, 2:37:15 AM5/5/04
to
Have a look at our white paper on Data Recovery & Repair.

http://www.whitesands.com/Products/drt_white.pdf

<denise> wrote in message news:405f6c4e.7e...@sybase.com...

0 new messages