topic Re: Replication error. in Operating System - HP-UX

Replication error.

Boniface Chiwenda — Sat, 03 Jul 2004 04:27:39 GMT

Hi everybody
I am trying to replicate some data using a cron
setting.
In the first place the replication itself is working perfectly well when I use dbstudio. The replication script also works pretty well when I execute it using sqlplus. How I what to execute the same script using a cron on a unix server and
this is what Iam getting.

begin sys.dbms_refresh.refresh('SNAPADMIN.GEN'); end;

*
ERROR at line 1:
ORA-12008: error in snapshot refresh path
ORA-01115: IO error reading block from file 11 (block # 144420)
ORA-01110: data file 11: '/home/asy/rdbms/ASYDB/tra03ASY_DB.dbf'
ORA-27092: skgfofi: size of file exceeds file size limit of the process
Additional information: 262143
Additional information: 368641
ORA-06512: at "SYS.DBMS_SNAPSHOT", line 617
ORA-06512: at "SYS.DBMS_SNAPSHOT", line 674
ORA-06512: at "SYS.DBMS_IREFRESH", line 577
ORA-06512: at "SYS.DBMS_REFRESH", line 211
ORA-06512: at line 1

has any one got a solution to this problem.

Regards
Bonnie.

Re: Replication error.

Yogeeraj_1 — Sat, 03 Jul 2004 04:57:43 GMT

hi bonnie,

Oracle 8.1.7 performs more stringent checking on O/S limits when opening the
database.

Can you check the filesize limit with ulimit command? You may have to change it to unlimited.

Also, confirm your OS and Database versions.

best regards
Yogeeraj

Re: Replication error.

Boniface Chiwenda — Sat, 03 Jul 2004 05:48:49 GMT

Thanks Yogeeraj
What i have as the file size is 4194303.
Let me give it a shot.
Thanks again for your help.

Best regards

Bonnie.

Re: Replication error.

R. Allan Hicks — Tue, 06 Jul 2004 09:02:44 GMT

Yogeeria is correct about the ulimits.

One of the things you may want to make sure of is who is the user in the cron job. Since you can run it from sqlplus, and not from cron, a better question might be what is different?

Are you running as a different user in cron than when you run the script in SQL. That is the user's limits that you want to adjust.

If cron looks should start to look like a dead end, what about submitting it as a timed job under enterprise manager? Since you are probably using advanced replication, the job system is already active.