distribution sync with backup

hi,
i am not certain what is purpose when the distribution
database is configured with sync with backup. i understand
from the help online that this will not truncate the
publication database log.
however, i can't achieve this.
this is what i did in my testing.
1) configure distribution and publication db with sych
with backup.
2) Backup the distribution db.
3) Add a record in the publication db.
4) Backup the publication db.
5) Make sure that the record is replicated in the
subscriber db.
6) delete the record in the subscriber db.
7) Restore the distribution db from the backup.
8) Backup the publication db.
i expect the missing record to be replicated to the
subscriber db. but it didn't. anything i missed out? or is
my understanding wrong?
thanks for any help.
-
leonghio (leonghio@.yahoo.com.sg)
This option prevents the backup of the distribution database getting ahead
of the backup of the publishing database. You need to add the record then do
the backups - publishing db first then distributor. Next delete the record
on the subscriber then do the restores and it should work ok with the new
record going down.
Regards,
Paul Ibison
|||i tried the following steps but still can't get the
results.
1) make sure both publication and distribution are sync
with backup
2) add a record to publication.
3) log backup publication.
4) full backup distribution.
5) delete the record from subscriber.
6) stop replication agents and restore distribution.
7) start replcation agents.
8) log backup publication again to re-initiate replication.
but record is not replicated again to the subscriber.
what seems to be the problem? am i getting the steps wrong?
thanks!
-
leonghio(leonghio@.yahoo.com.sg)

>--Original Message--
>This option prevents the backup of the distribution
database getting ahead
>of the backup of the publishing database. You need to add
the record then do
>the backups - publishing db first then distributor. Next
delete the record
>on the subscriber then do the restores and it should work
ok with the new
>record going down.
>Regards,
>Paul Ibison
>
>.
>
|||If the record has already gone to the subscriber it won't
be sent again. This is because MSreplication_subscriptions
holds a column called transaction_timestamp which is the
last one sent down. This value can be obtained from
msrepl_commands (the last 8 zeros aren't always seen), so
it would be possible to force the change by updating
MSreplication_subscriptions yourself to a previous
transaction if there is one.
HTH,
Paul Ibison
|||the way the sync with backup setting works is
1) a transaction is entered in the published databases
2) the transaction is written to the transaction log
3) the transaction log is dumped and during the dump the log reader will
then read the transaction from the log and write it to the distribution
database
"leonghio" <anonymous@.discussions.microsoft.com> wrote in message
news:0a5001c4317f$efbf3630$7d02280a@.phx.gbl...
> hi,
> i am not certain what is purpose when the distribution
> database is configured with sync with backup. i understand
> from the help online that this will not truncate the
> publication database log.
> however, i can't achieve this.
> this is what i did in my testing.
> 1) configure distribution and publication db with sych
> with backup.
> 2) Backup the distribution db.
> 3) Add a record in the publication db.
> 4) Backup the publication db.
> 5) Make sure that the record is replicated in the
> subscriber db.
> 6) delete the record in the subscriber db.
> 7) Restore the distribution db from the backup.
> 8) Backup the publication db.
> i expect the missing record to be replicated to the
> subscriber db. but it didn't. anything i missed out? or is
> my understanding wrong?
> thanks for any help.
> -
> leonghio (leonghio@.yahoo.com.sg)

Distributed transaction error, need help please!

Hi, I have configured a linked server, and i have a procedure which makes an UPDATE in a local table using the data in the linked server.

Specifically, I have a function which checks if a given code exists in a linked server's table. The UPDATE changes the value of a column in a local table, if the function returns 1.

I've run the procedure and it gave an error after a few hours cause a simple conversion error inside the function. I solved the error. After this, the procedure did not work more. It gives me the following message:

Server: Msg 7391, Level 16, State 1, Procedure EXISTEONC, Line 16
The operation could not be performed because the OLE DB provider 'MSDASQL'
was unable to begin a distributed transaction.

(EXISTEONC is the function, and in the line 16 there is an OPENQUERY)
Im sure MSDTC is working... i'm lost because i dont know why it worked the first time and not now. Ive also wrote the function again as it was before, but it still doesent works.

Thanks a lot...don't mean to ask a silly question but have you recompiled the stroed procedure ?|||Do you mean opening the procedure and closing it again??
Yes, I did that...|||i'm trying to understand - is the function being called from a stroed proc

wording is not too clear|||i'll try to explain me...

I have a stored proc, it looks like this:

BEGIN TRANS
UPDATE CodeTable SET Found='Yes' WHERE EXISTEONC(ONCCode)=1
COMMIT

the function EXISTEONC takes a code (ONCCode is a field in CodeTable) and searchs it in a table in the linked server. If it was found, it returns 1.

The error appears in line 16 of the function (in the openquery statement to search the code), but only if i execute the Procedure:
If i write this in the Analyzer:
PRINT EXISTEONC('1234')
there is no error, i think, the error appears only if the function is executed inside a transaction ...|||First thing is I would be 100% sure that DTC is running ok

If this is OK go backa and drop and recreate everything i.e. stroed proc and the function

If this is still not working then it may be some sort of corruption in the data access components on your server thats messed up the DLL's responsible for OLE DB.

Not sure how to approach this - Reinstall the MDAC?

Any ideas folks ??|||I'm 100% sure DTC is running (i stopped it, and restarted it about 500000 times).
I compiled the function and the procedure again...
With SQL server I can see the data in the linked server. Remember: if I execute the function outside the UPDATE, for example:
PRINT dbo.EXISTEONC('1111')
it works, so, the connection is working and the data is not corrupted.

but, the *&#%#@. procedure still doesent work...

I forgot to say this: the server database is a Sybase Adaptiver Server Anywhere 6.0, and I connect to it through an ODBC... i access the ODBC in Sql Server 2000.