ora-25408 can not safely replay call

[Abhishek Sharma]

Hi Team,

             We have integrated our Django project with sqlalchemy ORM.

Django: 1.4.2

SQLALCHEMY-0.8

Recently we started getting the following error while making the service call and it leads to internal server error:

ora-25408 can not safely replay call

Is there any way in sqlalchemy to handle this issue at DB connection level while making DB connection.

Thanks

Abhishek Sharma

[Mike Bayer]

SQLAlchemy can re-connect after an exception that is marked as a "disconnect" exception, if this is one, however there is not a mechanism built in that does this automatically without raising an exception; this is because that would require a replay of the entire transaction.      There is a patch for a full-blown "transaction replay" extension that I wrote for a client some years ago but it's never been merged, that is viewable at https://bitbucket.org/zzzeek/sqlalchemy/issues/3104/transaction-replay-extension if you'd like to see how that works.

if the ora-25408 error is not being trapped as a "disconnect" exception, you can customize fully what happens when this exception is intercepted using the handle_error event: http://docs.sqlalchemy.org/en/rel_1_0/core/events.html#sqlalchemy.events.ConnectionEvents.handle_error.  We use this at Openstack to fully customize what happens when various exceptions are intercepted.  We use our own retry decorators (e.g. @retry_on_failure, things like that) in order to replay methods fully when certain exceptions are intercepted.

Thanks

Abhishek Sharma

--
You received this message because you are subscribed to the Google Groups "sqlalchemy" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sqlalchemy+...@googlegroups.com.
To post to this group, send email to sqlal...@googlegroups.com.
Visit this group at http://groups.google.com/group/sqlalchemy.
For more options, visit https://groups.google.com/d/optout.

[Abhishek Sharma]

Team we have configured pool_recycle=1200 as part of engine creation because we have 30 minutes timeout as network level to close the idle connection.

We are still seeing ORA-25408 with the following logs:

Traceback (most recent call last):

[Wed Sep 30 03:03:48 2015] [error]   File "/opt/mdui/.env/lib/python2.7/site-packages/sqlalchemy/pool.py", line 250, in _close_connection

[Wed Sep 30 03:03:48 2015] [error]     self._dialect.do_close(connection)

[Wed Sep 30 03:03:48 2015] [error]   File "/opt/mdui/.env/lib/python2.7/site-packages/sqlalchemy/engine/default.py", line 418, in do_close

[Wed Sep 30 03:03:48 2015] [error]     dbapi_connection.close()

[Wed Sep 30 03:03:48 2015] [error] DatabaseError: ORA-25408: can not safely replay call

Can someone help us to resolve this issue.

[Abhishek Sharma]

Team can someone help us to resolve this issue.

I am not sure whether due to this issue my apache process got hanged and need to restart to get response.

Please help us to troubleshoot this issue.

[Mike Bayer]

I'm not familiar with the ORA-25408 error or what causes it.  You might want to check on the cx_oracle mailing list.