Logical standby issues

 
Need some help or opinions regarding Oracle logical standby database.

Environment:

Logical standby, version 10, release 1, patch 4. 
Primary and standby are tight coupled (maximum availability mode). 

These are the main issues I encountered so far: 


1) Deletion of consumed archive logs both on the standby and the primary site: 
how to detect what logs CAN be deleted: i.e. no one, standby or primary, is 
going to request this archive log in the future as far as standby is 
concerned(backups are not considered here)? 

Recently primary failed to spawn new sessions recently. I suspect this is due 
to FAL not founding archive log. Error reads:  

Wed Oct 26 21:42:30 2005
FAL[server, ARC1]: Begin FAL noexpedite archive (branch 559249864 thread 1 
sequence 14126 dest 
(DESCRIPTION=(ADDRESS_LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=172.17.16.52)(PORT=1522)))(CONNECT_DATA=(SERVICE_NAME=VSSECLOG_XPT.NEYDARLINAN.IS)(INSTANCE_NAME=VSSECLOG)(SERVER=dedicated))))
ARC1: Standby redo logfile selected for thread 1 sequence 14126 for destination 
LOG_ARCHIVE_DEST_2
FAL[server, ARC1]: Complete FAL noexpedite archive (thread 1 sequence 14126 
destination 
(DESCRIPTION=(ADDRESS_LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=172.17.16.52)(PORT=1522)))(CONNECT_DATA=(SERVICE_NAME=VSSECLOG_XPT.NEYDARLINAN.IS)(INSTANCE_NAME=VSSECLOG)(SERVER=dedicated))))
FAL[server, ARC1]: Begin FAL archive (branch 559249864 thread 1 sequence 14116 
dest 
(DESCRIPTION=(ADDRESS_LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=172.17.16.52)(PORT=1522)))(CONNECT_DATA=(SERVICE_NAME=VSSECLOG_XPT.NEYDARLINAN.IS)(INSTANCE_NAME=VSSECLOG)(SERVER=dedicated))))
Wed Oct 26 21:42:31 2005
Errors in file c:\oracle\product\10.1.0\admin\vs\bdump\vs_arc1_472.trc:
ORA-00308: cannot open archived log 
'D:\ORACLE\ORADATA\VS\ARCHIVE\ARC14116_0559249864.001'
ORA-27041: unable to open file
OSD-04002: unable to open file
O/S-Error: (OS 2) The system cannot find the file specified.
 
I suspect I am doing something wrong deleting primary/standby logs. Should I 
always execute dbms_logstdby.purge_session and rely on DBA_LOGMNR_PURGED view?

2) Time from time standby breaks but this is more due to the schema patches. I 
found that Oracle recycle bin can be a problem(which feature can be disabled 
anyway), operations involving snapshots(snapshots are not supported, that's 
true. But things like "create table as select from snapshot" would fail on 
standby as well.)  I'd like to discuss those things and try to set some policy 
to prevent such problems in advance so that standby just runs w/o admin 
intervention. 

3) DB Guard leakes PGA memory on the primary. It leaked it for a fact(by OS 
reports), it was not V$-view bug. Had to disable it. Not really happy about 
that.


5) Primary is not happy about unexpected standby restarts. FAL generaly fails 
to reconnect causing DB pauses for 4-5 seconds every 15 or so minutes. I found 
some recommendations(kind of "try this, if not try that") on the metalink. It 
recommended more or less to defer/enable standby log destination(at primry db) 
but what is the real solution anyway?


6) To sum up: something's a bit shaky in my environment at least. Should I give 
up with maximum availability mode and reduce the database to maximum 
performance mode? I am interested to hear from others.

Thank you in advance, 
Laimis

Other related posts: