_DISTRIBUTED_LOCK_TIMEOUT parameter

Thread view:

Enable EMail Alerts

Start New Thread

Thread rating:

purohitatul@gmail.com - 31 Aug 2007 19:34 GMT

Hi All,

Oracle version: 9.2.0.8
We are facing a problem with one of our jobs that runs on say database
db_a and writes on db_b.
Job, at times, aborts giving error "ORA-02049: timeout distributed
transaction waiting for lock".
One of the recommendations that came up was to increase
DISTRIBUTED_LOCK_TIMEOUT to 7 minutes instead of default 60 secs.

A had a couple of questions on this change and it would be great if
anyone can provide pointers:

1. By changing this parameter, is the impact limited to operations
performed over dblink or if anyone has seen anything different.

2. Any other way that this error can be fixed.

Further whats confusing is no other query is locking the table when
this job runs and this job doesn't lock itself out.

Thanks,
Atul

Reply to this Message

Mark D Powell - 01 Sep 2007 00:05 GMT

On Aug 31, 2:34 pm, purohita...@gmail.com wrote:

> Hi All,
>
[quoted text clipped - 19 lines]
> Thanks,
> Atul

A better solution would probably be to
1- find out what application is holding the lock and see if you can
modify/correct it to commit its changes in a more timely manner and
2- modify the current application to capture the error and try again
up to X times. This way you do not force every application to wait 7
minutes before timing out.

IMHO -- Mark D Powell --

Reply to this Message