[OpenSER-Devel] db_mysql commands out of sync, presence

[Henning Westerholt]

On Fri, March 28, 2008 2:14 pm, Juha Heinanen wrote:
> i'm getting these also in usrloc db_mode = 1:
>
> Mar 28 14:50:24 tutpro /usr/sbin/openser[11688]: LOG: Handling PUBLISH to
> <sip:jh at tutpro.com>
> Mar 28 14:50:24 tutpro /usr/sbin/openser[11688]:
> ERROR:db_mysql:db_mysql_submit_query: driver error on query: Commands out
> of sync; you can't run this command now
> Mar 28 14:50:24 tutpro /usr/sbin/openser[11688]: ERROR:core:db_do_query:
> error while submitting query
> Mar 28 14:50:24 tutpro /usr/sbin/openser[11688]:
> ERROR:presence:update_presentity: unsuccessful sql query
> Mar 28 14:50:24 tutpro /usr/sbin/openser[11688]:
> ERROR:presence:handle_publish: when updating presentity
> Mar 28 14:50:24 tutpro /usr/sbin/openser[11688]: ERROR: PUBLISH to
> <sip:jh at tutpro.com> failed
>
> i would really appreciate if db experts could help in solving this
> problem.  what does Commands out of sync sync mean?  when does it
> happen?

Hi Juha,

sorry for the late reply, caused from the VON and additional issues..

This was discussed already some time ago:
http://lists.openser.org/pipermail/devel/2008-February/012238.html

Additional information is also present in the MySQL documentation:
http://dev.mysql.com/doc/refman/5.1/en/commands-out-of-sync.html

> although in the above the errors resulted from publish, i have
> seen the same happening also without any request processing as i have
> reported earlier.  i have not changed anything in the way my openser is
> running.  these errors just started to appear at some point this year
> when i upgraded to a newer trunk.

Anca commited a fix related to a bug in the presence code, that caused
this problems. She also mentioned that an additional fix is necessary. I
don't know if this is already fixed..

> if the cause for the errors cannot be found, perhaps it would would then
> be better to go back to the working db implementation that existed at
> the end of last year.

I don't think that the error really is located in the db implementation.
We've done quite a lot of work in the db interface and the additional
modules that uses it, i don't think its a good idea to revert this all.

I actually do some tests, but not at all for presence. Anybody else
observed problems similar to this?

Cheers,

Henning