My pleasure to confirm bug :) As additional information, this server has a quite a few SMPP (~20), so I guess that would match the same pattern here.
Concerning the queued message, it seems like the "queued" message actually came back so it may be only that they weren't finished loading ? The specific connection that had the problem here was using EMI2 however, if that is of any importance... I'm also using store-file in this case. I didn't suppose that the loading ok 23k sms should take 10 mins, unless there's a good reason to load them slowly ? ------- Mathieu Bruneau -----Original Message----- From: Stipe Tolj [mailto:[EMAIL PROTECTED] Sent: Wednesday, October 08, 2008 6:06 PM To: Mathieu Bruneau Cc: [email protected] Subject: Re: Losing queued message ? Mathieu Bruneau schrieb: > Hi, > I notice that we are losing queued message on a few occasion over the > last few days and I was wondering if it could be a bug with Kannel? > > It seems like that when kannel is restarted we are losing "queued" > message sometimes (as reported by status) , they seems however to > still be in the kannel.store file (as reported by store-status). > > I'm using a debian box so the init script are the one from their > package and on shutdown kannel throw this: > > 2008-10-08 16:02:20 [8551] [10] PANIC: gwlib/gw-rwlock.c:137: > gw_rwlock_rdlock: Assertion `lock != NULL' failed. > 2008-10-08 16:02:20 [8551] [10] PANIC: > /usr/sbin/bearerbox(gw_panic+0xcc) [0x80cc73c] > 2008-10-08 16:02:20 [8551] [10] PANIC: > /usr/sbin/bearerbox(gw_rwlock_rdlock+0x73) [0x80c1973] > 2008-10-08 16:02:20 [8551] [10] PANIC: > /usr/sbin/bearerbox(route_incoming_to_boxc+0x76) [0x8057346] > 2008-10-08 16:02:20 [8551] [10] PANIC: > /usr/sbin/bearerbox(bb_smscconn_receive+0x118) [0x8059028] > 2008-10-08 16:02:20 [8551] [10] PANIC: /usr/sbin/bearerbox [0x8086a81] > 2008-10-08 16:02:20 [8551] [10] PANIC: /usr/sbin/bearerbox [0x80878e5] > 2008-10-08 16:02:20 [8551] [10] PANIC: /usr/sbin/bearerbox [0x80c2e70] > 2008-10-08 16:02:20 [8551] [10] PANIC: /lib/tls/libpthread.so.0 > [0x400abb63] > 2008-10-08 16:02:20 [8551] [10] PANIC: > /lib/tls/libc.so.6(__clone+0x5a) [0x404de18a] When you start it back > up Kannel report : store size 23560 > But the individual connection report : ( online 579s, rcvd 1, sent 82, > failed 0, queued 0 msgs ) > > Since obvisouly the message are still there, when/how does kannel is > suppose to requeue them? Kannel will requeue them as soon as the bearerbox is restarted and the storage subsystem (store-file or spool) is reloaded. The shutdown problem of bearerbox has been confirmed. We see this on various deployments with a larger set of SMPP based SMSC connections. The smsbox_list get's empty and while still receiving MOs in the SMSC module instances we run into the rwlock assertion panic. I need to clarify why the shutdown is causing problems. Thanks for confirming that this has not been an isolated issue on our side. Alex confirmed also the shutdown problematics. Stipe -- ------------------------------------------------------------------- Kölner Landstrasse 419 40589 Düsseldorf, NRW, Germany tolj.org system architecture Kannel Software Foundation (KSF) http://www.tolj.org/ http://www.kannel.org/ mailto:st_{at}_tolj.org mailto:stolj_{at}_kannel.org -------------------------------------------------------------------
