THIS IS AN AUTOMATED MESSAGE, DO NOT REPLY.
The following task has a new comment added:
FS#122 - Kamailio 3.1.3 crash
User who did this - Jasmin Schnatterbeck (jasmin)
----------
hmm... is it possible to run commands with the EXEC module out of the tm:local-request event route?
Do you have any idea how to identify the dialog, if the $dlg variables are not set?
Thanks!
Jasmin
----------
More information can be found at the following URL:
http://sip-router.org/tracker/index.php?do=details&task_id=122#comment190
You are receiving this message because you have requested it from the Flyspray bugtracking system. If you did not expect this message or don't want to receive mails in future, you can change your notification settings at the URL shown above.
THIS IS AN AUTOMATED MESSAGE, DO NOT REPLY.
The following task has a new comment added:
FS#122 - Kamailio 3.1.3 crash
User who did this - marius zbihlei (marius.zbihlei)
----------
after checking better with on_send route that is not an option ... It is not run for local generated replies
----------
More information can be found at the following URL:
http://sip-router.org/tracker/index.php?do=details&task_id=122#comment189
You are receiving this message because you have requested it from the Flyspray bugtracking system. If you did not expect this message or don't want to receive mails in future, you can change your notification settings at the URL shown above.
THIS IS AN AUTOMATED MESSAGE, DO NOT REPLY.
The following task has a new comment added:
FS#122 - Kamailio 3.1.3 crash
User who did this - marius zbihlei (marius.zbihlei)
----------
ha, another problem. Because the datagram process is started/forked from the PROC_MAIN, the child_init functions of other modules are not actually called at all :(. So you don't have access to the sql_ops commands from the mi_datagram process. The problem is that the event_route[tm:local-request] is called from the context of the fifo process (which is wrong).
You can you on_send routes to achieve the same result (but careful because they are issued for every request, so no need to do in in normal route)
Marius
----------
More information can be found at the following URL:
http://sip-router.org/tracker/index.php?do=details&task_id=122#comment188
You are receiving this message because you have requested it from the Flyspray bugtracking system. If you did not expect this message or don't want to receive mails in future, you can change your notification settings at the URL shown above.
THIS IS AN AUTOMATED MESSAGE, DO NOT REPLY.
The following task has a new comment added:
FS#122 - Kamailio 3.1.3 crash
User who did this - Jasmin Schnatterbeck (jasmin)
----------
here gdb output...
----------
More information can be found at the following URL:
http://sip-router.org/tracker/index.php?do=details&task_id=122#comment187
You are receiving this message because you have requested it from the Flyspray bugtracking system. If you did not expect this message or don't want to receive mails in future, you can change your notification settings at the URL shown above.
THIS IS AN AUTOMATED MESSAGE, DO NOT REPLY.
The following task has a new comment added:
FS#122 - Kamailio 3.1.3 crash
User who did this - Jasmin Schnatterbeck (jasmin)
----------
ok, I put mi_datagram at the end of the list.... (as I think, that does mean, that mi_datagram is loaded after sql ops, not sqlops after mi_datagram? (as you have stated))
mi_datagram at the end of the list: I tested again and triggered a timeout call, but the problem persists.
I attached new gdb output..
Jasmin
----------
More information can be found at the following URL:
http://sip-router.org/tracker/index.php?do=details&task_id=122#comment186
You are receiving this message because you have requested it from the Flyspray bugtracking system. If you did not expect this message or don't want to receive mails in future, you can change your notification settings at the URL shown above.
THIS IS AN AUTOMATED MESSAGE, DO NOT REPLY.
The following task has a new comment added:
FS#122 - Kamailio 3.1.3 crash
User who did this - marius zbihlei (marius.zbihlei)
----------
Quote "an you confirm that sqlops is loaded in the cfg after the mi_datagram module? "
Fixing the bug is as simple as putting mi_datagram at the end of the load list. This is a quick dirty fix, as a more generic fix should be use.
Marius
----------
More information can be found at the following URL:
http://sip-router.org/tracker/index.php?do=details&task_id=122#comment185
You are receiving this message because you have requested it from the Flyspray bugtracking system. If you did not expect this message or don't want to receive mails in future, you can change your notification settings at the URL shown above.
THIS IS AN AUTOMATED MESSAGE, DO NOT REPLY.
The following task has a new comment added:
FS#122 - Kamailio 3.1.3 crash
User who did this - Jasmin Schnatterbeck (jasmin)
----------
thanks for the explanation!
I didnt't see your msg, when I posted last time...sry
here is the load order of the modules, sqlops is already loaded after mi_datagram:
loadmodule "db_mysql.so"
loadmodule "mi_fifo.so"
loadmodule "mi_datagram.so"
loadmodule "kex.so"
loadmodule "tm.so"
loadmodule "tmx.so"
loadmodule "sl.so"
loadmodule "rr.so"
loadmodule "pv.so"
loadmodule "maxfwd.so"
loadmodule "usrloc.so"
loadmodule "registrar.so"
loadmodule "textops.so"
loadmodule "siputils.so"
loadmodule "xlog.so"
loadmodule "sanity.so"
loadmodule "ctl.so"
loadmodule "auth.so"
loadmodule "auth_db.so"
loadmodule "domain.so"
loadmodule "presence.so"
loadmodule "presence_mwi.so"
loadmodule "presence_dialoginfo.so"
loadmodule "presence_xml.so"
loadmodule "nathelper.so"
loadmodule "nat_traversal.so"
loadmodule "mediaproxy.so"
loadmodule "avpops.so"
loadmodule "uac.so"
loadmodule "uac_redirect.so"
loadmodule "enum.so"
loadmodule "exec.so"
loadmodule "path.so"
loadmodule "permissions.so"
loadmodule "utils.so"
loadmodule "sqlops.so"
loadmodule "cfgutils.so"
loadmodule "dialog.so"
loadmodule "diversion.so"
loadmodule "lcr.so"
Jasmin
----------
More information can be found at the following URL:
http://sip-router.org/tracker/index.php?do=details&task_id=122#comment184
You are receiving this message because you have requested it from the Flyspray bugtracking system. If you did not expect this message or don't want to receive mails in future, you can change your notification settings at the URL shown above.
THIS IS AN AUTOMATED MESSAGE, DO NOT REPLY.
The following task has a new comment added:
FS#122 - Kamailio 3.1.3 crash
User who did this - marius zbihlei (marius.zbihlei)
----------
I have explained above. Fifo process doesn't run mod_child_init for modules that are loaded after the mi_datagram process. So a quick fix would be to change the order you load the module in the cfg. Albeit this is not a fix, it should get rid of the crashes.
Marisu
----------
More information can be found at the following URL:
http://sip-router.org/tracker/index.php?do=details&task_id=122#comment183
You are receiving this message because you have requested it from the Flyspray bugtracking system. If you did not expect this message or don't want to receive mails in future, you can change your notification settings at the URL shown above.
THIS IS AN AUTOMATED MESSAGE, DO NOT REPLY.
The following task has a new comment added:
FS#122 - Kamailio 3.1.3 crash
User who did this - Jasmin Schnatterbeck (jasmin)
----------
oh, that's it :-)
I triggered an expired call (stopped audio streams only), mediaproxy sent request to end the dialog via mi, kamailio crashed... but why?
Thanks Jasmin
----------
More information can be found at the following URL:
http://sip-router.org/tracker/index.php?do=details&task_id=122#comment182
You are receiving this message because you have requested it from the Flyspray bugtracking system. If you did not expect this message or don't want to receive mails in future, you can change your notification settings at the URL shown above.
THIS IS AN AUTOMATED MESSAGE, DO NOT REPLY.
The following task has a new comment added:
FS#122 - Kamailio 3.1.3 crash
User who did this - marius zbihlei (marius.zbihlei)
----------
hello
Just had another look at the trace. It escaped me the first time, but the crash happens in the MI fifo process. the fifo process receives a dlg_terminate_dlg command and sends a BYE, that somehow causes, in the context of that process, the SQL query to get executed(this is normal albeit a little weid). But because how the process is initializated (I will explain in a minute), the dbh structure (which is per child initialized) is still NULL in that process.
Now the problem with the initialization. The module are initialized (mod_child_init actually so this is per process module initialization) in the order they are loaded. But because the FIFO process forks and then blocks on accepting datagrams, modules loaded after it are not mod_child_inited (the way the sr_module inits them is via some recursive calls). This wouldn't cause such a big fuss if the sql query is not executed.
Can you confirm that sqlops is loaded in the cfg after the mi_datagram module? Also I ask other core developers to look at this problem and comment on it
Marius
----------
More information can be found at the following URL:
http://sip-router.org/tracker/index.php?do=details&task_id=122#comment181
You are receiving this message because you have requested it from the Flyspray bugtracking system. If you did not expect this message or don't want to receive mails in future, you can change your notification settings at the URL shown above.
