2009/11/09

Gap is not resolved in dataguard. WHY ?


One of my customers had an environment like that:
2 node rac primary site and one active dataguard in standby site. 11gr1 (11.1.0.7 )version. The problem was gaps in standby site. When everthing is working fine log writer process send change vectors to standby and active dataguard updates tables immediately. But when network failes or manually a gap is created by shutting down standby site gap is never resolved. I looked around and finally realized that the problem was because of insufficient archiver processes.
in that case the value was 2.

Now read this information:
LOG_ARCHIVE_MAX_PROCESSES init parameter has a range 1 to 30.
Gaps are resolved using archiver processes in primary site. When a gap is detected ARCn process sents archived log to standby.But one process is always dedicated to local arhiving and never talks to standby site.

A maximum of thirty ARCn processes can be enabled if there are a large number of archive logs that need to be transferred. Twenty-nine of these ARCn processes may ship to remote locations;
one ARCn process is always dedicated to local archival.



As we have 2 threads in rac then 2 archiver processes are dediceated for local archiving. This makes the sense : no archiver process is ready to send standby site.
all are dediceted.
increase the value greater then 2 and problem is solved. I set it to 8 and gap is resolved.





No comments:

Post a Comment