hi all,
we are using
Kannel 1.1.5 (no patches) and SMPP protocol and Red Hat Linux
7.1.
in last 8 days, smsbox process disappear
several times (not bearerbox).
we check smsbox.log and foundsome panic
message:
2001-10-01 01:08:35 [3] PANIC: mutex_unlock:
Mutex
Hi,
The version of SMPP in kannel is 3.4. The SEMA SMPP implementation checks
the version and if it is 3.4 the SEMA implementation stores the message
contents in the message payload field. This is why the sm_length is set to 0
as required by the SMPP specification. The message payload paramter
Bi Bruno,
do you get the panic even if you don't specify the dlr_* in http push ?
yes, even for normal cgi-bin/sendsms this happens.
Stipe
[EMAIL PROTECTED]
---
Wapme Systems AG
Münsterstr. 248
40470 Düsseldorf
Tel:
Did you check if the problem still occurs when you comment fakesmsc group
entries ? I noticed that bearerbox crashes with cgi-bin/status on Redhat 6.2
with the last CVS version abd only when you add fakesmsc group.
PS : I am using emi2 SMSC only
-Message d'origine-
De : [EMAIL
Hi. I'm using a falcom A2d1 modem with recent CVS kannel (sep28),
and have noticed the following weird behaviour:
LF (0xa) and CR (0xd) becomes switched around !!?
When I send an SMS with CR out through kannel, my normal phone gets an LF.
And vice versa.
This does NOT happen, when I use
Hi all,
Did you check if the problem still occurs when you comment fakesmsc group
entries ? I noticed that bearerbox crashes with cgi-bin/status on Redhat 6.2
with the last CVS version abd only when you add fakesmsc group.
There is no fakesmsc group in the config file. There is definitly
It's now on CVS a patch to support ACK without timestamp.
Please note that Delivery Reports wouldn't work without that timestamp,
so http fields are irrelevant.
--
Bruno David Rodrigues
BEGIN:VCARD
VERSION:2.1
N:Rodrigues;Bruno David Simões
FN:Bruno David Simões Rodrigues
ORG:Vodafone
hi Ian,
thanks for your reply.
i do log in debug mode.
2001-10-08 17:57:50 [8] DEBUG: HTTP: Sending request:
2001-10-08 17:57:50 [8] DEBUG: Octet string at 0x808eb18:
2001-10-08 17:57:50 [8] DEBUG: len: 76
2001-10-08 17:57:50 [8] DEBUG: size: 77
2001-10-08 17:57:50 [8] DEBUG:
It's now on CVS a patch to support ACK without timestamp.
I'll try that one and report.
Stipe
[EMAIL PROTECTED]
---
Wapme Systems AG
Münsterstr. 248
40470 Düsseldorf
Tel: +49-211-74845-0
Fax: +49-211-74845-299
E-Mail: [EMAIL
I'm now looking at your code.
You are forcing receiving sms's through ME memory, right ?
What about if some mobile doesn't work correctly with it ?
quote(SMSs go through memory - this is reliable, but only implemented and
+ tested on an Ericsson gm25 modem and falcom */)
can should create a
Usually (at least with UCP), it's the smsc that limits you by delaying the
ACK for a certain amount of time, so you don't have to worry about that.
But if UCP SMSC is configured to use windowed-mode, I (or someone)
would have to make a patch to smsc_emi2 to define the window to use.
If it's
I'm now looking at your code.
You are forcing receiving sms's through ME memory, right ?
What about if some mobile doesn't work correctly with it ?
quote(SMSs go through memory - this is reliable, but only
implemented and
+ tested on an Ericsson gm25 modem and falcom */)
can should
I thought of not auto-detecting so I can force the old behaviour or force
a ME or SIM usage (someone told about a limit of read/writes in SIM memory).
Of course the code could auto-detect that if I try to force ME and mobile
doesn't
support ME memory, kannel should use other mode.
--
Bruno
Usually (at least with UCP), it's the smsc that limits you by delaying the
ACK for a certain amount of time, so you don't have to worry about that.
But if UCP SMSC is configured to use windowed-mode, I (or someone)
would have to make a patch to smsc_emi2 to define the window to use.
If it's
Are you sure that smsc aren't in stop-and-wait mode and you are in windowed
mode ?
If that's the case, you could be loosing messages, like
send msg 01
send msg 02
send msg 03
receive ack 01
send msg 04
send msg 05
send msg 06
receive ack 04
...
message 02,03, 05 and 06 were missed because smsc
Are you sure that smsc aren't in stop-and-wait mode and you are in windowed
mode ?
Yes. I'm sure.
If that's the case, you could be loosing messages, like
send msg 01
send msg 02
send msg 03
receive ack 01
send msg 04
send msg 05
send msg 06
receive ack 04
...
message 02,03, 05 and 06 were
2001-10-08 17:57:50 [8] DEBUG: HTTP: Sending request:
2001-10-08 17:57:50 [8] DEBUG: Octet string at 0x808eb18:
2001-10-08 17:57:50 [8] DEBUG: len: 76
2001-10-08 17:57:50 [8] DEBUG: size: 77
2001-10-08 17:57:50 [8] DEBUG: immutable: 0
2001-10-08 17:57:50 [8] DEBUG: data: 47 45 54
When I download and compile the latest CVS, it
runs only the run-bearer and run-smsbox binaries,
and not the bearerbox or smsboxes.
No logs are generated or touched.
The run-bearer and run-smsbox stay up forever,
apparently doing nothing. And I don't know how
to see why, short of running a
That's not bad, that's the way EMI protocol works.
And yes, kannel retries the message but once again, the only messages that
will pass are the ones sent after the last ACK/NACK.
--
Bruno David Rodrigues
- Original Message -
From: Andreas Fink [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Soon.
I've commited now a patch with preliminary code to do that.
Now we have to check in where (or which services) would we
want to bind to a ip. I know I want to bind bearebox-smsbox
to localhost only and I have a smsc connection that I want to bind
to a IP that is an alias (and linux insists
can you run bearerbox -v 0 path/kannel.conf and smsbox
by hand ?
--
Bruno David Rodrigues
- Original Message -
From: Steve Rapaport [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Sent: Monday, October 08, 2001 7:07 PM
Subject: Strange non-running state
When I download and compile the
Ok, now the code replaces it back to colon and replaces password
with * chars.
--
Bruno David Rodrigues
- Original Message -
From: Bruno David Rodrigues [EMAIL PROTECTED]
To: Gildas PERROT [EMAIL PROTECTED]; Devel@Kannel
[EMAIL PROTECTED]
Sent: Monday, September 24, 2001 4:37 PM
Once we've talked about moving sms_split code to smsc code because
it's smsc code that would encode the message (for example to gsm alphabet)
and only then we know the size of the message
Reference: a message with special characters - example lots of { would get
droped by emi2 if splitted.
But
Once we've talked about moving sms_split code to smsc code because
it's smsc code that would encode the message (for example to gsm alphabet)
and only then we know the size of the message
Reference: a message with special characters - example lots of { would get
droped by emi2 if splitted.
But
but what about if the charset used by the smsc is different ? we cannot
assume that it will be the default gsm alphabet as was told here.
--
Bruno David Rodrigues
- Original Message -
From: Andreas Fink [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Sent: Monday, October 08, 2001 10:28 PM
I remember one smsc that eats IA5 chars and internally converts to gsm
alphabet
(I remember it because I couldn't use my fields_to_dcs in it)
--
Bruno David Rodrigues
- Original Message -
From: Andreas Fink [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Sent: Monday, October 08, 2001 10:39
2001-10-08 Bruno Rodrigues [EMAIL PROTECTED]
* gwlib/http.c: When url was like http://user:pass@site, the first
colon was changed to @ to differentiate from the colon for port
number. Now the code replaces it back to colon and replaces password
with *.
Hi Bruno,
On 8 October 2001, Bruno David Rodrigues wrote:
I'm now looking at your code.
You are forcing receiving sms's through ME memory, right ?
No the code uses SM memory and defaults to that. All mobiles have
SM memory. There is the option of using ME memory for phones which
have it.
Hi Paul,
On 8 October 2001, Paul Keogh wrote:
I'm now looking at your code.
You are forcing receiving sms's through ME memory, right ?
What about if some mobile doesn't work correctly with it ?
quote(SMSs go through memory - this is reliable, but only
implemented and
+ tested on
comments included:
--
Bruno David Rodrigues
- Original Message -
From: Matt Flax [EMAIL PROTECTED]
To: Bruno David Rodrigues [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Sent: Tuesday, October 09, 2001 12:09 AM
Subject: Re: CVS patches
Hi Bruno,
On 8 October 2001, Bruno David
hi, Ian,
does smsbox quit when she get
error?
we have got 100,000 msg in last 8 days and met
the error only several times.so we think the url is correct.we think
the key was : 2001-10-08 17:57:50 [8] PANIC: mutex_unlock:
Mutex failure! 2001-10-08 17:57:50 [8] PANIC: System
2001-10-08 Bruno Rodrigues [EMAIL PROTECTED]
* gw/smsc_emi2.c, gwlib/cfg.def: Added throughput (messages per
second) variable to configuration. Delays 1/throughtput seconds
between sending messages.
* gw/smsc_emi2.c, gwlib/cfg.def: Added window configuration that
32 matches
Mail list logo