Distribution Agent Time Outs

Distribution Agent Time Outs

Post by Jeff Carringto » Wed, 27 Sep 2000 04:00:00



Hello,

  Could anybody shed some light on the timing out of the replication
distribution agent?

  I am moving over a large amount of transactions (with a pull subscription)
and the distribution agent keeps timing out.
I have increased the query time out from a value of 2000, up to 3000 up to
6000.  And each time I re-start the agent, the distribution begins rolling
back after the same amount of execution time (the disk I/O is very close
also).  It is like the changes I make to the agent do not make a difference.
I have stopped the SQL Agent each time after a change to make sure the agent
has successfully stopped so the profile changes will be in effect.

  Any input on this matter would be greatly appreciated.

Thanx,

Jeff Carrington

 
 
 

Distribution Agent Time Outs

Post by Jeff Carringto » Wed, 27 Sep 2000 04:00:00


Jari,

  Thanks for the input but I do not use any text fields in any of my table definitions.  In fact the biggest fields I have in these tables are a couple of varchar(40).

  Appreciate the help,

Jeff

* Sent from Devdex.com http://www.devdex.com The Web Developers Index *
The world's largest index site for web developers.

 
 
 

Distribution Agent Time Outs

Post by - Jari » Thu, 28 Sep 2000 07:44:24


Jeff,

Quote:>   Could anybody shed some light on the timing out of the replication
> distribution agent?

I had exactly the same problem, but problem
was only TEXT type columns. Of course
my solution to change them to VARCHAR is
not fix, but now they are working....

I also hope somebody knows more about this issue...

Here is my post to this new group couple weeks ago.

****************************************************

I am really having hard time to figure
out why Distribution Agent fails
to bulk copy initial snapshot.

I have a table which has something like
30 columns, there are also 4 TEXT type
columns.

During initial bulkcopy to target table
Distribution agent fails. First it
reports that

"Table Article 100000 row(s) copied. Total 200000"

This means that it has already bulk copied all data, but
after this agent becomes suspect.

It seems there is something going on with updating NULL
values to TEXT columns. Since you can find out from
profiler, that

UPDATE "Article" SET "NewsCodeComments" = NULL
WHERE TEXTPTR( "NewsCodeComments" ) = NULL

Seems to be executed on forever loop...

Can somebody help me with this one?

It doesn't make any difference if I use Transactional or
Snapshot replication, since this is happening during
initial load. Anyway, I have tested both and results
are the same!

*****************************************

- Jari -

 
 
 

Distribution Agent Time Outs

Post by Joe » Wed, 25 Oct 2000 04:00:00


Having the same problem with a 300000 recordset.  The last
field is a text field.  Will try converting to varchar.
Thanks for the idea.



Quote:

> Jeff,

> >   Could anybody shed some light on the timing out of the replication
> > distribution agent?

> I had exactly the same problem, but problem
> was only TEXT type columns. Of course
> my solution to change them to VARCHAR is
> not fix, but now they are working....

> I also hope somebody knows more about this issue...

> Here is my post to this new group couple weeks ago.

> ****************************************************

> I am really having hard time to figure
> out why Distribution Agent fails
> to bulk copy initial snapshot.

> I have a table which has something like
> 30 columns, there are also 4 TEXT type
> columns.

> During initial bulkcopy to target table
> Distribution agent fails. First it
> reports that

> "Table Article 100000 row(s) copied. Total 200000"

> This means that it has already bulk copied all data, but
> after this agent becomes suspect.

> It seems there is something going on with updating NULL
> values to TEXT columns. Since you can find out from
> profiler, that

> UPDATE "Article" SET "NewsCodeComments" = NULL
> WHERE TEXTPTR( "NewsCodeComments" ) = NULL

> Seems to be executed on forever loop...

> Can somebody help me with this one?

> It doesn't make any difference if I use Transactional or
> Snapshot replication, since this is happening during
> initial load. Anyway, I have tested both and results
> are the same!

> *****************************************

> - Jari -

 
 
 

1. Continual time-outs (time is not on my side)

I'm making small changes (eg changing a field size from 20 characters to 30
characters) in SQL server 7.
We have probably 10 clients connected. The table has around 5,000 records,
with no indexes on the field I'm trying to change. I've set all OBDC
timeouts to 0 (infinite).
Why does such a simple operation require me to get EVERYONE out of the
database? Obviously I also can't make any more serious changes without
booting everyone off.
I know it's M$ software and all, but surely it should work better than
this???

Dan

 ---
A quick script to eliminate ALL the bugs in Windows at once:

c:\> deltree c:\windows

2. usage rate

3. Distribution agent timing out

4. US-NJ-Princeton ORACLE Contract Developer/2000 Forms 4.5

5. Distribution Agent runs time out

6. Printing problems under Windows

7. suspect distribution agent and suspect merge agent ?

8. MS SQL Server 2000 time-outs periodicly over the internet

9. Time outs

10. Time-outs that didn't happen under NT4sp6

11. Time outs!

12. query time outs! please help