Hello,
On 8/29/12 1:23 PM, Øyvind Kolbu wrote:
On 2012-08-29 at 07:38, Daniel-Constantin Mierla
wrote:
seeing the sql commands would help, also the log
messages with debug=3
Now with "db_check_update" enabled. I've
deleted the row for username 2549619
in the voip_location table from the secondary server between then re-registration.
I think the problem is that the check is done using only one of the "read
table"
connections, and not done for each of the "write table" connections.
is
usrloc module configured in db only mode?
Read operation cannot be done in 'parallel' fashion of the cluster, it
is only serial or round robin, stopping at first successful query (even
when there is no data records in the result of the query).
If usrloc keeps also the records in memory, there is no select done at
runtime, only updates, which are write operations.
db_check_update is only for SQL UPDATE command, in MySQL you can see the
number of affected rows and if it is 0, then the module does an insert
(another write operation).
So, the question is why do you think the check is done on one 'read
table', or you don't refer to the check of affected rows for UPDATE?
Cheers,
Daniel
--
Daniel-Constantin Mierla -
http://www.asipto.com
http://twitter.com/#!/miconda -
http://www.linkedin.com/in/miconda
Kamailio Advanced Training, Berlin, Nov 5-8, 2012 -
http://asipto.com/u/kat