Are you checking the logs?? Is there any throttling errors there?

That's the only way to tell if  2 Msg/sec is enough limit or not. I
think there shouldn't be any unless your fellow SMSC's limits are that
low.

Anyway, if you receive messages at a higher rate that what you are
being able to send, sooner or later you'll get a whole pile of unsent
messages... The only way to solve that is either limit the incoming
throughput (translating the queueing problem to the SMSC in the
process) or to ask for a higher throughput to your SMSC operator.

Hope it helps,

On 8/10/05, Ady Wicaksono <[EMAIL PROTECTED]> wrote:
> Hi Alejandro,
> 
> I've been set
> 
> throughput=2
> max-pending-submits=1
> 
> with assumption SMSC limit transmitter to 2sms/s
> 
> but it's not help so much :), any idea ?
> 
> 
> > First of all, if you don't already know it, ask your operator what's you
> > maximum allowed rate of messages per second.
> >  The big problem with being throtled is the retries: You try to send more
> > messages than allowed, so the SMSC rejects some of them and they are
> > retried
> > again later.
> >  This happens again and again and you end up sending a lot less messages
> > that you'd do if you send them at the right rate.
> >  Setting the throttle to the "right" value will give you the best possible
> > throughput with the carrier, so maybe that's enough to solve your problem.
> >  If that's not enough, I'm afraid that you should ask them to increase
> > your
> > rate, since your service seems to need more throughput that what they are
> > giving you. Anyway, IMHO it's a good practice to set the throttle rate to
> > a
> > "realistic" value even if you are not being throttled. It will spare you
> > some headaches with carriers if your connection goes down and comes back
> > again with thousands of messages awaiting to be sent (the carriers does
> > not
> > usually like to be hit hard with huhdreds of messages per second).
> >  Hope it helps.
> >
> >
> >  On 8/9/05, Ady Wicaksono <[EMAIL PROTECTED]> wrote:
> >>
> >> Currently, point 1 is the big possibility
> >>
> >> //
> >>
> >> 1- Throttling could cause problems if you're receiving 5 messages per
> >> second
> >> but only allowed to send 1 per second...
> >>
> >> However.., this is rule from telco operator, so how could we minimize
> >> the impact of the rule ?
> >>
> >>
> >>
> >> Alejandro Guerrieri wrote:
> >>
> >> 1- Throttling could cause problems if you're receiving 5 messages per
> >> second
> >> but only allowed to send 1 per second...
> >>  2- If you have set too much or too many simultaneous connections, load
> >> could climb or requests could be serialized. Anyway, except you are
> >> receiving _very_ high loads of messages, this shouldn't be an issue.
> >>  I also use asynchronous and works very well.
> >>  Any hints from the log files?
> >>  Regards,
> >>
> >>  On 8/8/05, Ady Wicaksono <[EMAIL PROTECTED]>
> >> <[EMAIL PROTECTED]> wrote:
> >>
> >>
> >>
> >> 1. Yes, i'm being throttled, receiver is not throttled but transmitter
> >> is
> >> throttled. However i run this application as transciever.
> >> 2. Apache misconfigured -> what kind of misconfigured application ?
> >>
> >> I made application asynchronous, it means that kannel will forward sms
> >> to
> >> http application (which some of them made by PHP, perl, or JSP)
> >> however this application will print nothing, i've been set
> >> "omit-empty=true".
> >>
> >> Another process incoming SMS and send the reply to the sender.
> >>
> >> Thx
> >>
> >>
> >>
> >> Alejandro Guerrieri wrote:
> >>
> >> What it's obvious is that you're receiving messages faster than
> >> delivering.
> >> That means that sooner or later you get a whole bunch of queued
> >> messages.
> >>  A few things I'd check:
> >>  * Could be that you are being throttled? I don't think so (0.87
> >> msgs/second
> >> doesn't look like a suitable throughput for throttling). Check on the
> >> smsbox.log
> >>  anyway, and if that's the case adjust throttling to avoid this
> >> problem.
> >>  * How are your apps doing? I'd take a look at the applications, maybe
> >> they
> >> are responding too slowly and that's what it's causing the delays.
> >>  * A misconfigured Apache could be the cause also.
> >>  * Check calling the apps from a web browser, how long does it take to
> >> respond?
> >>  * Is there anything noteworthy on the logs? What does they say about
> >> the
> >> queued messages?
> >>  Hope it helps,
> >>  Alejandro.
> >>
> >>  On 8/8/05, Ady Wicaksono <[EMAIL PROTECTED]>
> >> <[EMAIL PROTECTED]> <[EMAIL PROTECTED]>
> >> <[EMAIL PROTECTED]> wrote:
> >>
> >>
> >> Hi Chriss, sorry for late response, i'm restarting and waiting this
> >> kannel
> >> eating my memory :)
> >>
> >> I did, look at this complete log
> >>
> >> Kannel bearerbox version `1.4.0'.
> >> Build `Jul 20 2005 19:06:58', compiler `3.2.2 20030222 (Red Hat Linux
> >> 3.2..2-5)'.
> >> System Linux, release 2.4.20-8smp
> >> , version #1 SMP Thu Mar 13 17:45:54 EST
> >> 2003, machine i686.
> >> Hostname ******, IP 127.0.0.1 <http://127.0.0.1/> <http://127.0.0.1/>
> >> <http://127.0.0.1/> <http://127.0.0.1/> <http://127.0.0.1/>
> >> <http://127.0.0.1/> <http://127.0.0.1/>.
> >> Libxml version 2.5.4.
> >> Compiled with MySQL 4.1.9, using MySQL 4.1.9.
> >> Using native malloc.
> >>
> >>
> >> Status: running, uptime 3d 19h 39m 43s
> >>
> >> WDP: received 0 (0 queued), sent 0 (0 queued)
> >>
> >> SMS: received 515767 (0 queued), sent 258900 (0 queued), store size 455
> >> SMS: inbound 1.56 msg/sec, outbound 0.78 msg/sec
> >>
> >> DLR: 2388 queued, using mysql storage
> >>
> >> Box connections:
> >> smsbox:(none), IP 127.0.0.1 <http://127.0.0.1/> <http://127.0.0.1/>
> >> <http://127.0.0.1/> <http://127.0.0.1/> <http://127.0.0.1/>
> >> <http://127.0.0.1/> <http://127.0.0.1/> (0 queued), (on-line 3d
> >> 19h 39m 41s)
> >>
> >>
> >> SMSC connections:
> >> ********* SMPP:******************* (online 329983s, rcvd 515767, sent
> >> 258900, failed 0, queued 454 msgs)
> >>
> >>
> >> Still have no IDEA :( huh
> >>
> >>
> >>
> >>
> >>
> >> Chris Dobbs wrote:
> >>
> >> Have you configured Kannel to use a database for DLR storage? if not it
> >> stored them in memory hashes until they are completed. I found Kannle
> >> used MUCH less memory once I had DLR's into the DB.
> >> -Chris
> >>   ----- Original Message -----
> >>   From: Ady Wicaksono
> >>   To: users@kannel.org
> >>   Sent: Thursday, August 04, 2005 12:22 PM
> >>   Subject: Kannel eat memory too big
> >>
> >>
> >>   Log at the "top" result below, bbox-smsc1 eat about 197Mbyte and smsc2
> >> eat 120M and growing
> >>   i use kannel 1.4.0. Is it normal or, kannel has a memory leak ?
> >>   Any information & suggestion is normal
> >>
> >>   Thx
> >>
> >>
> >>
> >>   ----------- cut here ---------
> >>    18:13:43  up 147 days, 15 min,  1 user,  load average: 0.02, 0.03,
> >> 0.00
> >>   539 processes: 538 sleeping, 1 running, 0 zombie, 0 stopped
> >>   CPU0 states:   2.3% user   2.0% system    0.0% nice   0.0% iowait
> >> 95.0% idle
> >>   CPU1 states:   3.3% user   1.0% system    0.0% nice   0.0% iowait
> >> 95.0% idle
> >>   Mem:  2064252k av, 2038984k used,   25268k free,       0k shrd,
> >> 294952k buff
> >>                      1217516k actv,       0k in_d,   45028k in_c
> >>   Swap: 4192924k av,    2508k used, 4190416k free
> >> 926492k cached
> >>
> >>     PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME CPU
> >> COMMAND
> >>   11205 smsc2      15   0  197M 197M  1436 S     0.7  9.7  28:58   0
> >> bbox-smsc1
> >>   16239 smsc1      15   0  120M 119M  1424 S     0.0  5.9  18:32   0
> >> bbox-smsc2
> >>   ----------- cut here ---------
> >>
> >>
> >>
> >>
> >>
> >> --
> >> Regards,
> >> Ady Wicaksono
> >> HP: +628562208680
> >>
> >>
> >>
> >>
> >>
> >>
> >> --
> >> Regards,
> >> Ady Wicaksono
> >> HP: +628562208680
> >>
> >>
> >>
> >>
> >>
> >>
> >> --
> >> Regards,
> >> Ady Wicaksono
> >> HP: +628562208680
> >>
> >>
> >
> >
> > --
> > Alejandro Guerrieri
> > Magicom
> > http://www.magicom-bcn.net/
> >
> 
> 


-- 
Alejandro Guerrieri
Magicom
http://www.magicom-bcn.net/

Reply via email to