cancel
Showing results for 
Search instead for 
Did you mean: 

Getting Message: 589... 'Invalid role string received' after adding a new database to the replication system

Former Member
0 Kudos

Getting Message: 589... 'Invalid role string received' after adding a new database to the replication system, and DSI thread goes down.

I have verified with rs_helpuser that the maint_user is used for the connection to database (mctest) .

I revoked and re-granted replication_role and sa_role to maint_user in mctest.

Dropped and recreated the maint_user login in the ASE, as well as the user in the database. Even made the user an alias to dbo.

To no avail. I'd very much appreciate any help anyone can provide. -Thanks, Andrei

Former Member
0 Kudos

Hi Mark, thank you so much for your pointers. Indeed I found out there was a trigger on the replicate site, which was checking for a role that didn't exist in that server.


I have another question for you, I'd very much appreciate it if you could help me again:

I wanted to get rid of a broken server/database (with disk failures) from my replication system.

The DSI thread was not moving, queue was growing. We were planning to remove that ASE entirely.

I dropped the connection, came back without error.

Yet the DIST and SQT threads went down for the other two databases in the system.

I restarted the RS - and now it's basically not functional: all threads are down.
I'm really worried since this a production server.

Can you think of a method to bring them back up ?


Thanks so much!

Andrei Ursu

Mark_A_Parsons
Contributor
0 Kudos

Good to hear you figured out the first issue and was able to fix it.

As for the latest issue, same as the first issue ... need more info as to why various threads are going down.

For the DIST thread you should try resume distributor <dsname>.<dbname> and watch for any new error messages in the repserver errorlog to indicate why the distributor is going down. (If you're lucky the DIST thread will stay up so nothing else to do.)

If the DIST thread goes down again but not enough info in the errorlog, use admin who,sqm to get the first seg/block of the DIST thread and then sysadmin dump_queue to view the first couple seg/blocks of the DIST thread (this should show the command causing the DIST thread to go down).

Once you've got the command (causing the DIST thread to go down) you can then decide next steps.  For example, you could resume distributor <dsname>.<dbname> skip tran to skip the problematic command and (optionally; as appropriate) manually issue the command in the replicate db (if a ASE command).

NOTE: I'm assuming the down SQT thread is related to the DIST thread, so once the DIST issue is addressed/fixed then the SQT thread should stay up.

NOTE: For future reference, when you have a new issue/discussion, consider starting a new thread otherwise other forum members may skip over this thread not realizing that you've changed topics and are now trying to address a new/different issue.

NOTE: For SAP/Sybase Replication Server related discussions ... you may want to consider posting in the forum where you're likely to get the attention of more Repserver users.

Accepted Solutions (0)

Answers (0)