On Thu, Dec 31, 2009 at 12:14 PM, Henning Westerholt <henning.westerholt@1und1.de> wrote:

On Thu, December 31, 2009 10:55 am, alex pappas wrote:
> [..]

Hey Alex,

the restaurant was nice, the beer is indeed good there. :-)

> I just manage to connect with oracle with the following URL :
>
> modparam("sqlops","sqlcon","ca=>oracle://user:pass@alex.gr/cid_name.alex.gr
> ")

Good that you managed to find the correct syntax.

> Now I run into other problems. The same query that from ./kamailio_orasel
> give me result, when I try that from kamailio.cfg I get an error and
> after that the system is getting very high load.

This is probably caused from a segfault which caused a core dump. :-(
After this the system is busy to write the core file with the memory dump
to the disk, which can get really huge depending on the process number and
shared memory size. Do you managed to find some core dump? They should
located in the working directory of the server.

> The error is :
>
> kernel: kamailio[12442]: segfault at 0000000000000010 rip 00002b1b42952764
> rsp 00007fff6b144d30 error 4
>
> Does anyone had the same issue ?

The oracle module is probably one of the DB modules which has got the
least testing, due its dependencies. But i think they are some people
which successful using it. The next step would be to investigate the
backtrace from the core dump.

Another option would be to setup an unixodbc connection on the DB side and
then connect which the db_unixodbc module instead.

Regards,

Henning