Hi list,

this is a serious thing we just cope with. Has anyone observed this
with an "own" smsbox implementation that does a lot of
connect/re-connects via lifetime?!

Stipe

[EMAIL PROTECTED]
-------------------------------------------------------------------
Wapme Systems AG

Vogelsanger Weg 80
40470 D�sseldorf

Tel: +49-211-74845-0
Fax: +49-211-74845-299

E-Mail: [EMAIL PROTECTED]
Internet: http://www.wapme-systems.de
-------------------------------------------------------------------
wapme.net - wherever you are
--- Begin Message ---
The following NEW bug has been ADDED.
=======================================================================
http://bugs.kannel.org/view_bug_page.php?f_id=80
=======================================================================
Reporter:                   tolj
Handler:                    
=======================================================================
Project:                    Kannel
Bug ID:                     0000080
Category:                   General
Reproducibility:            always
Severity:                   crash
Priority:                   immediate
Status:                     new
=======================================================================
Date Submitted:             08-20-03 14:14 BST
Last Modified:              08-20-03 14:14 BST
=======================================================================
Summary:                    bearerbox runs out of thread when smsbox connections are 
sequentially established and dropped
Description: 
Ok, what I do is to use an exernal "smsbox" (actually it is an Exim
transport) that connects to bearerbox, identfies itself (for
smsbox-routing) and then issues the message to bearerbox and reads the acks
for it. Then disconnects and so on...

This is used for an SMTP-to-SMS chaining setup via Exim's transports.

Here is the bearerbox.log:

2003-08-20 14:48:43 [0] INFO: ----------------------------------------
2003-08-20 14:48:43 [0] INFO: Kannel bearerbox II version cvs-20030811
starting
2003-08-20 14:48:43 [0] INFO: MAIN: Start-up done, entering mainloop
2003-08-20 14:48:43 [0] DEBUG: smsc_fake: start called
2003-08-20 14:48:43 [6] DEBUG: sms_router: time to sleep
2003-08-20 14:48:43 [6] DEBUG: sms_router: list_len = 0
2003-08-20 14:48:58 [5] INFO: Fakesmsc client connected from 127.0.0.1
2003-08-20 14:49:06 [2] DEBUG: HTTP: Creating HTTPClient for `127.0.0.1'.
2003-08-20 14:49:06 [3] DEBUG: HTTP: Destroying HTTPClient area
0x816fc10.
2003-08-20 14:49:06 [3] DEBUG: HTTP: Destroying HTTPClient for
`127.0.0.1'.
2003-08-20 14:49:42 [4] DEBUG: Started thread 7 (gw/bb_boxc.c:function)
2003-08-20 14:49:42 [7] INFO: Client connected from <127.0.0.1>
2003-08-20 14:49:42 [7] DEBUG: Started thread 8
(gw/bb_boxc.c:boxc_sender)
2003-08-20 14:49:42 [7] DEBUG: boxc_receiver: got boxc_id <exim> from
<127.0.0.1>
2003-08-20 14:49:42 [7] DEBUG: boxc_receiver: unlocking sender
2003-08-20 14:49:42 [8] DEBUG: boxc_sender: sender unlocked
2003-08-20 14:49:43 [4] DEBUG: Started thread 9 (gw/bb_boxc.c:function)
2003-08-20 14:49:43 [9] INFO: Client connected from <127.0.0.1>
2003-08-20 14:49:43 [9] DEBUG: Started thread 10
(gw/bb_boxc.c:boxc_sender)
2003-08-20 14:49:43 [9] DEBUG: boxc_receiver: got boxc_id <exim> from
<127.0.0.1>
2003-08-20 14:49:43 [9] DEBUG: boxc_receiver: unlocking sender
2003-08-20 14:49:43 [9] DEBUG: boxc_receiver: sms received
2003-08-20 14:49:43 [5] DEBUG: smsc_fake: sending message to client
2003-08-20 14:49:43 [10] DEBUG: boxc_sender: sender unlocked
2003-08-20 14:49:43 [10] DEBUG: send_msg: sending msg to boxc: <exim>
2003-08-20 14:49:43 [9] DEBUG: boxc_receiver: sms received
2003-08-20 14:49:43 [5] DEBUG: smsc_fake: sending message to client
2003-08-20 14:49:43 [10] DEBUG: boxc_sender: sent message to <127.0.0.1>
2003-08-20 14:49:43 [10] DEBUG: send_msg: sending msg to boxc: <exim>
2003-08-20 14:49:43 [9] DEBUG: boxc_receiver: sms received
2003-08-20 14:49:43 [5] DEBUG: smsc_fake: sending message to client
...
2003-08-20 14:50:34 [5] DEBUG: smsc_fake: sending message to client
2003-08-20 14:50:34 [110] DEBUG: boxc_sender: sent message to <127.0.0.1>
2003-08-20 14:50:34 [110] DEBUG: send_msg: sending msg to boxc: <exim>
2003-08-20 14:50:34 [109] DEBUG: boxc_receiver: sms received
2003-08-20 14:50:34 [5] DEBUG: smsc_fake: sending message to client
2003-08-20 14:50:34 [110] DEBUG: boxc_sender: sent message to <127.0.0.1>
2003-08-20 14:50:34 [110] DEBUG: send_msg: sending msg to boxc: <exim>
2003-08-20 14:50:34 [109] DEBUG: boxc_receiver: sms received
2003-08-20 14:50:34 [5] DEBUG: smsc_fake: sending message to client
2003-08-20 14:50:34 [110] DEBUG: boxc_sender: sent message to <127.0.0.1>
2003-08-20 14:50:34 [110] DEBUG: send_msg: sending msg to boxc: <exim>
2003-08-20 14:50:34 [109] INFO: Connection closed by the box <127.0.0.1>
2003-08-20 14:50:34 [110] DEBUG: boxc_sender: sent message to <127.0.0.1>
2003-08-20 14:50:34 [110] DEBUG: Thread 110 (gw/bb_boxc.c:boxc_sender)
terminates.
2003-08-20 14:50:34 [109] DEBUG: Thread 109 (gw/bb_boxc.c:function)
terminates.
2003-08-20 14:50:35 [4] DEBUG: Started thread 111 (gw/bb_boxc.c:function)
2003-08-20 14:50:35 [111] INFO: Client connected from <127.0.0.1>
2003-08-20 14:50:35 [111] PANIC: cannot allocate wakeup pipe for new
thread
2003-08-20 14:50:35 [111] PANIC: System error 23: Too many open files in
system

see the thread numbers in []!!

Which means threads seem to stay alive for boxc connections that have
already disconnected. At some time bearerbox will run out of resources and
will PANIC that no more threads can be established.

Stipe

=======================================================================


--- End Message ---

Reply via email to