Xref: alice comp.databases.oracle.misc:52367 comp.databases.oracle.server:83090
Path: alice!news-feed.fnsi.net!news.maxwell.syr.edu!newsfeed.stanford.edu!n00zpHeed.netscum.dQ!netscum.int!news-feed.inet.tele.dk!bofh.vszbr.cz!news-proxy.baileynm.com!web.nmti.com!usenet
From: Parichai Kalari <pkalari@abbnm.com>
Newsgroups: comp.databases.oracle.server,comp.databases.oracle.misc
Subject: replication job
Date: Wed, 02 Feb 2000 12:42:55 -0600
Organization: ABB Automation
Lines: 66
Message-ID: <38987AAF.18047254@abbnm.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Mailer: Mozilla 4.6 [en] (WinNT; U)
X-Accept-Language: en

Hi!

Many thanks to all those who responded to my earlier question on
replication.  

I've got another situation with a replication job.  Here are the
details:

Environment used:  Digital Unix 4.0E with Oracle 7.3.4.4.0  The
scheduled job - 'dbms_defer_sys.execute' that pushes the jobs to the
other master site(we have only two master sites in a multi-master
replication env.)got broken and I found that the reason was,  for
some reason it was trying to push as the user 'system'(the first
transaction it is trying to push from deftran has the origin_user as
system) - which is not desired(and the errors were like default logon
not supported).  

Actions taken so far:  1)Unschedule and reschedule the job as repadmin
user and 2)remove and submit the job as repadmin user, but to no avail.
The other transactions that went through earlier and those waiting
behind the first transaction has the origin_user as say 'king'.  When I
tried to specify the userid/pwd (say king/queen)before the dblink in the
destination argument of dbms_defer_sys, it looked like it is trying to
push(has the JQ lock and an entry in dba_jobs_running).  The job didn't
get broken, but the transactions never started to decrease!  I have
tried 'poking' the snp process - actually stopped and started it, but
didn't help.  Tried to push manually with the same result.

The interesting thing though is, when I try to push manually, when I
didn't specify the useid/passwd in front of the dblink like this,

execute
dbms_defer_sys.execute('i2.mastersite2',transaction_count=>300,batch_size=>250); 
 
it threw out some error messages like this:

ERROR at line 1:
ORA-23324: error ORA-00604: error occurred at recursive SQL level
ORA-02063: preceding line from I2.MASTERSITE2, while creating deferror
entry at
"I2.MASTERSITE2" with error -604
ORA-06512: at "SYS.DBMS_DEFER_SYS", line 903
ORA-06512: at line 1

The links work fine though.

But when I specified the userid/pwd like this, 

execute
dbms_defer_sys.execute('king/queen@i2.mastersite2',transaction_count=>300,batch_size=>250);

it said it successfully completed the procedure, but to no effect on
deftran!

The job execution interval is set to 10 seconds(would this have to be
changed?- it works on other systems!).

And the first transaction it is trying to push as system has about
220000 rows affected(as noted from the defcall for that particular tran
id).


Any suggestion?  Comments?  Help?

Thanks
Parichai