[00:28] CunningPike (n=CunningP@204.239.8.149) left irc: "And so, to bed" [00:30] CunningPike (n=CunningP@S010600095b33697f.vc.shawcable.net) joined #openser. [00:53] bkw_ (i=brian@ppp-70-128-123-137.dsl.tulsok.swbell.net) left irc: "Leaving" [01:40] _VoiceMeUp_COM (n=_VoiceMe@145-27.mc.cite.net) joined #openser. [01:48] maxedout (n=maxstout@user-38lm3ok.cable.mindspring.com) joined #openser. [03:31] JT (n=jon@unaffiliated/jt) left irc: No route to host [03:36] DoberMann (n=james@AToulouse-156-1-118-28.w90-30.abo.wanadoo.fr) joined #openser. [03:39] JerJer[mobile] (n=jj@dhcp-11-150.hstf.interop.net) joined #openser. [03:43] JT (n=jon@CPE-203-45-105-79.nsw.bigpond.net.au) joined #openser. [03:48] DoberMann[ZZZzzz (n=james@AToulouse-156-1-2-168.w81-250.abo.wanadoo.fr) left irc: Read error: 110 (Connection timed out) [03:59] JerJer[mobile]_ (n=jj@45.128.1.225) left irc: Read error: 110 (Connection timed out) [04:03] maxedout (n=maxstout@user-38lm3ok.cable.mindspring.com) left irc: Client Quit [04:23] Zaaaz (n=1046D5BC@201-95-180-153.dsl.telesp.net.br) joined #openser. [04:28] Defraz (n=t0tal@24-116-152-177.cpe.cableone.net) joined #openser. [04:49] CunningPike (n=CunningP@S010600095b33697f.vc.shawcable.net) left irc: "And so, to bed" [04:51] CunningPike (n=CunningP@204.239.8.149) joined #openser. [04:57] hi guys [05:11] unspin (n=unspin@207-232-98-114.ip.van.radiant.net) joined #openser. [05:12] greetings =D [05:12] does anyone know anything about the PRE-RFC3555 SDP semantics for negotiating a G729 audio session? [05:13] specifically related to G729 and G729B [05:13] i'm trying to establish what the expected behaviour, PRE-RFC3555 is for a SDP payload that indicates support for G729 but does not contain the "annexb=no" codec parameter [05:13] primarily because i think RFC3555 came along and messed everything up (by indicating that the lack of "annexb=no" means that annexB should be used by default) and i want to establish, outside of my own observations, that 1) before RFC3555 lack of annexb=no didn't mean use annexb and 2) that i'm not crazy ;/ [05:14] i've seen plenty of devices that dont take the lack of the "annexb=no" parameter to mean that annexB G729 needs to be used (namely all Pre 12.3 cisco AS5300 and AS5400 gateways) but i'm trying to find some standard or specification to support this behaviour so i can better quantify it (instead of just the anecdotes of my observations) [05:15] i've got a bunch of Cisco AS5400s that simply insist on strict RFC3555 compliance, however most other devices don't employ the semantics described in rfc3555 (namely that they don't indicate "annexb=no" when they don't want to use g729b) [05:15] the issue is that i end up needing to use g729b, and this is creating some sound quality issues for me [05:16] err, i do use OpenSER quite extensively so there is some association in all that =D [05:18] JerJer[mobile]_ (n=jj@192.16.170.13) joined #openser. [05:19] maxdoubt (n=mackstou@169.198.254.6) left irc: "Leaving" [05:28] erider (n=erider@unaffiliated/erider) left irc: "I don't sleep because sleep is the cousin of death!" [05:37] erider (n=erider@unaffiliated/erider) joined #openser. [05:38] JerJer[mobile] (n=jj@dhcp-11-150.hstf.interop.net) left irc: Read error: 110 (Connection timed out) [05:38] JerJer[mobile] (n=jj@192.16.170.13) joined #openser. [05:39] JerJer[mobile]_ (n=jj@192.16.170.13) left irc: Read error: 110 (Connection timed out) [05:58] JerJer[mobile] (n=jj@192.16.170.13) left irc: Read error: 110 (Connection timed out) [06:37] Defraz (n=t0tal@24-116-152-177.cpe.cableone.net) left irc: Read error: 54 (Connection reset by peer) [06:45] L|NUX (n=linux@unaffiliated/lnux/x-10290) joined #openser. [06:52] Zaaaz (n=1046D5BC@201-95-180-153.dsl.telesp.net.br) left irc: Read error: 110 (Connection timed out) [07:25] Defraz (n=t0tal@24-116-152-177.cpe.cableone.net) joined #openser. [07:41] dhakatel (n=ashrar@58.65.224.5) joined #openser. [07:59] Ded (n=ded@36.248-106-89.FTTH.rus-com.net) joined #openser. [08:02] Hi all. Help please. Is there anyone willing to dump files tables for Postgres 8.x for OpenSER 1.1.x? [08:09] stimpie (n=michiel@ip565faf27.direct-adsl.nl) joined #openser. [08:11] emp^ (n=emp@35.248-106-89.FTTH.rus-com.net) joined #openser. [08:26] CunningPike (n=CunningP@204.239.8.149) left irc: "And so, to bed" [08:33] Help please. Is there anyone willing to dump files tables for Postgres 8.x for OpenSER 1.1.x? [08:36] dude.. can you not find the mysql examples and simply convert them? [08:49] Nick change: DoberMann -> DoberMann[PullA] [08:53] Surely this would be the right thing to do. So I did. But I have a problem with the links between tables. You may already have done this work. It is necessary ... Please. [08:53] p.s. Sorry for my English :) [08:56] nicox (n=nicox@83-64-42-210.prater.xdsl-line.inode.at) joined #openser. [09:05] r_open (n=root@58.65.224.5) joined #openser. [09:08] r_open (n=root@58.65.224.5) left irc: SendQ exceeded [09:08] r_open (n=root@58.65.224.5) joined #openser. [09:14] r_open (n=root@58.65.224.5) left irc: "ircII EPIC4-2.6 -- Are we there yet?" [09:15] ffs not again [09:25] Loady (n=lasse@noc.mtulink.no) joined #openser. [09:25] toex (n=toex@pd95b297b.dip0.t-ipconnect.de) left irc: "Leaving" [09:26] Hello [09:26] can anybody help me with High CPU load on openser? [09:27] and also a problem with memory, Openser tells me that there is not enough memory [09:33] nicox: how many concurrent calls and what hw specs? [09:34] I am having some problemes with the dr_rules table, when i add more than 12 rows in the database it wont start. [09:34] 0(0) ERROR:drouting:load_routing_info: nul column [09:35] Loady: are you using mysql? and that's a strange error [09:35] codestr0m: im useing postgres sql. [09:36] yes, i have mysql [09:36] can you pastie your config at p.caboo.se [09:36] hm, concurrent calls is to 60 calls [09:36] about 800 user registred [09:37] nicox: I faced an io problem the other day when I had massive debugging on for a production system. 60 calls isn't much nor 800 registered users [09:37] and the user send every 15 seconds a notify, that they are alive [09:37] nicox: what hardware.. . I mean if it's on a soekris 4801 it would have high cpu ;) [09:37] Sipura/Linksys adapters [09:38] Pentium D 3 GHz [09:38] 2 GB RAM [09:38] nicox: can you install sysstat and run iostat -x 5 [09:38] yes, ofcourse, but i have to test it in the night, because that a testing system which i have to make to a producteion system [09:38] but thats a big problem [09:39] nicox: well.. I'm here now.. installing that won't cause any extra load [09:39] it just pulls some info from proc and displays it every 5 seconds [09:40] codestr0m: what spesfic part of the config du you need. i need to do a lot of censorship before i can paste it [09:40] no, now is not possible, to get the load, i have to make it to the production system, change the ip and so on, so its not possible now, i have to try it in the night [09:41] nicox: it sounds similar to what I faced.. [09:42] try to minimize your disk writes and track slow queries.. cause once the problem starts it's hard to recover from [09:42] Loady: well. make a testcase creating this problem at the smallest level and I'll take a look at it.. I plan to ferret out any postgres problems [09:45] i don't think there is a problem with the system, thats a clean new installed system. [09:45] can you have a look on the configuration? [09:46] codestr0m: hmmm.. i seem to have found another error as well, when browsing the logs. this might be more helpful, but i am not sure. [09:46] nicox: do you have a lot of debugging on.. I mean if you look at the logs. .are they just flying off the screen? and if so.. turn debugging off and see if that helps. that's what I'm saying [09:46] 0(0) ERROR:drouting:load_routing_info: 32 records found in dr_gateways [09:46] 0(0) ERROR:drouting:load_routing_info: 0 row DST_ID: 300 [09:46] Defraz (n=t0tal@24-116-152-177.cpe.cableone.net) left irc: Read error: 113 (No route to host) [09:46] hm, its a little bit better if the debugging is off [09:46] but not okay [09:47] it means openser makes a CPU-load of 50 Percent [09:50] nicox: what else do you have on this box? [09:50] nothing [09:50] Loady: I'd need a testcase [09:50] nicox: where is the db? [09:50] on an other server [09:50] sorry, we have fire alarm [09:50] i have to go [09:51] cpu caught fire [09:51] JT: ;) that's not funny [09:51] you're laughing [09:52] JT: and what if the asterisk guys get wind that openser is so hot? [09:52] then we'll have all those guys over here [09:52] codestr0m: ok thanks. im gonna see if i can get any info on the last one first. I'll post the config if i get stuck again, but thanks a bunch for your offer. [09:52] Loady: well. please report back what the problem was.. as it may be helpful to others [09:52] anyone knows a commandline tool for linux to do a call check, something like sipsak but with rtp? [09:52] codestr0m: indeed [09:53] martin--: well.. it may or not not be useful, but the quick and dirty way is tcpdump filtering by ip [09:54] codestr0m: but i need a tool to establish a connection first [09:54] i need this for a small check script [09:54] martin--: sipp [09:54] sipp.sf.net [09:55] which tries to establish a call every minute [09:55] martin--: sipp can do exactly that if you also tie it into a cron.. or maybe other ways [09:55] ah thanks [09:55] codestr0m: luckily asterisk isn't a direct competitor [09:55] , but don't ask me to help with the config.. it's xml mess [09:56] JT: asterisk is cool for what it does, but I use it on several systems.. [09:56] and wish I could replace it [09:56] SIPp can also send media (RTP) traffic through RTP echo and RTP / pcap replay. Media can be audio or audio+video. [09:56] that's what i'm looking for [09:56] codestr0m: hehe [09:56] martin--: are you going to calculate mos? [09:57] would be my next question [09:57] how can i do this [09:57] martin--: :) [09:58] we have ~35.000 customers on a voip platform, there are often calls of them complaining of bad quality [09:59] martin--: yeah. I know the feeling.. for this there are two choices.. and unfortunately nothing open source yet [09:59] that I'm aware of.. I could be wrong [09:59] replace the Internet? [10:00] we've bought a test system now [10:00] JT: no.. one is a free tool which will do what he wants, but no source is included. the other is a secret :) [10:00] to check the sip/mgcp signalling, rtp streams and ss7 signalling [10:01] martin--: sounds fancy.. I'm always curious about such things to see how others are handling it [10:01] martin--: are you guys up to 1B cdrs per month? [10:01] codestr0m: stop complaints about VoIPoI quality, sounds like an amazing product [10:02] JT: who said stop? ha ha ha.. no way, but identifying poor quality grey routes can help alot [10:02] a lot* [10:02] i dont know how much records we have, i have nothing to do with the billing [10:02] :) [10:02] codestr0m: i see [10:02] qdk (n=qdk@0x50c627be.bynxx11.adsl-dhcp.tele.dk) left irc: Read error: 113 (No route to host) [10:02] we have 2000 channels to the pstn [10:03] qdk (n=qdk@213.150.62.32) joined #openser. [10:03] martin--: interesting.. is that 3-4 ds3 or fibre.. and if you're in the US. don't tell anyone.. cause verizon did win against vonage.. pstn > sip is not safe these days [10:04] it's not us [10:05] and we use cisco as voip platform, so technical patents would be the problem auf cisco [10:05] as 5400? [10:05] yes [10:05] good kit.. expensive though [10:06] how much? [10:06] we have 6 of them [10:06] full of dsp's and E1 interface cards [10:06] JT: contact your local cisco drug dealer fo rmore info ;) [10:06] bah :P [10:06] i wanted a rought figure [10:06] JT: expensive because you have to pay extra for dsp and interfaces [10:07] yeah cisco like to sell empty metal boxes [10:07] martin--: well for those that don't want to spend the time to develop their own solution it makes it rather nice.. ds3 on one side and sip on the other [10:08] indeed, we have a similar setup. about 4500 ptsn channels about 30-50 as5300, with dsp cards. [10:09] Loady: you need so many as35300 for 4500 pstn channels? you use other codecs than g711? [10:09] so per E1, are we talking $5k, $10k, $20k? [10:09] you guys do realize there is open source pcb layouts for the cards you need and boxes from IBM and or other vendors which can easily handle this at a fraction of the cost [10:09] we only use g711, so we can put more channels on one as5400 [10:09] JT: quit asking that here ;) [10:09] codestr0m: why? [10:10] well... we use g711, but we try to plan a bit ahead. [10:10] JT: i guess a box witha sterisk with interfaces is cheaper :) [10:10] martin--: i don't doubt it [10:10] Loady: ah k [10:10] Loady: we install new ones when needed [10:10] i'm interested in how it stacks up to some of the other stuff like audiocodes and mediatrix [10:10] martin--: well. it depends. if you go the sangoma route with onboard echo cards. those can be pricing.. [10:10] i guess we need some new soon [10:11] we will get 5000 new customers this week [10:11] sangomas are cheaper than digium though [10:11] digium is crap or at least was last I had to deal with it [10:12] their drivers cause problems and really screwed this one client that was trying to do SS7 [10:12] i'm not debating that sangoma make good cards [10:12] and there are other problems with asterisk/ss7: i think this isnt certified to use with carriers [10:14] martin--: well. from first hand experience. over a year ago helping a client. replaced the digium cards with sangoma and the server was able to handle the load.. there were a couple reasons for this which basically boiled down to drivers and hw [10:14] and you're correct.. I wouldn't recommend SS7 to anyone yet [10:14] or is anyone using asterisk for carrier <-> carrier interconnections? [10:15] of course.. there are many carriers doing ss7 with asterisk for national to internation grey routes in europe [10:15] , but a lot of them have private stacks [10:16] it costs much time to even build a interconnection via cisco to the deutsche telekom [10:17] you have to do many test scenarios with them [10:17] martin--: never tried.. and are you referring to the interopt testing? and isn't there legislation on voip in germany.. my last dealing was several years ago [10:17] any good major carrier is going ot require you to do interopt tests.. track calls and such. it's a great practice.. and raises the bar to play the game.. [10:18] yes i'm referring to the interop testings [10:18] cisco things were tested before, so we dont need all tests [10:18] but it's till enough [10:18] martin--: hw.. configs.. and many other things can factor in [10:18] __Henning__ (n=henning@et-1-5.gw-nat.bs.ka.oneandone.net) joined #openser. [10:19] i i'm wondering if anyone did this before with deutsche telekom in germany [10:19] martin--: cisco gear does break.. you can get bad cards.. it helps bring these issues up [10:19] for asterisk [10:19] deutsche telekom sucks as ex monopolist [10:19] they try to slow down all as much as possible :) [10:20] <__Henning__> telekom.. :-/ [10:20] yeah.. all telecoms do that... which company do you work with if you can say... [10:20] query [10:21] __Henning__: i guess everyone who has to work with them, hates them :) [10:22] <__Henning__> martin: even as private customer they suck :-) [10:23] yep [10:28] <__Henning__> martin--: do you work for an call center company? [10:28] no [10:28] i work for a carrier [10:29] <__Henning__> ah ok. you mentioned friday that you have many call center agents as customers, so i thougt.. [10:30] we have some call centers as customer [10:30] <__Henning__> ok [10:31] codestr0m, Please could you send me to dump tables for postgres to email delidov.george [at) gmail.com [10:32] bdrubel (n=root@58.65.224.5) joined #openser. [10:32] Ded: this will be the least of your concerns.. if you can't figure this out.. I'm sure you will have more trouble after that.. I'm sorry [10:32] anyone is using openser + redundancy? [10:33] i want use linux-ha heartbeat for redundancy [10:34] sorry, for the long time, my configuration is at http://p.caboo.se/55839 [10:35] matrin: what type of redundency u need [10:36] i use heart beat previonsly [10:36] now i have my own way [10:36] failover from one box to another [10:36] no data replecation? [10:37] this also [10:37] i thought about mysql replicat [10:37] e [10:37] only register information? or [10:37] oonly subscriber + location table [10:37] rtpproxy and mediaproxy informaion [10:38] we dont use rtp/mediaproxy [10:38] ok [10:38] i used heartbeat previonly [10:38] Nick change: fulgas -> FuL|OUT [10:38] then i dig heart bit [10:39] and found arping is doing all thing [10:39] yes [10:39] u can configure heartbeat for openser [10:40] or use a normal script [10:40] that ping two machine [10:40] yea ok [10:40] my first complaint about heartbeat is that it won't compile under a uclibc system.. so I can't put it on my routers.. keepalived works well for me so far [10:40] and is case of failover use arping to take over [10:41] i already have tried heartbeat with virtual machines [10:43] there is a function in openser.cfg to replecate registed call [10:45] oh, really? [10:46] yap [10:47] it only replecate the register information what u see in (opserctl ul show) [10:52] toady (n=toady@251.9.39-62.rev.gaoland.net) joined #openser. [10:52] hi [10:58] <__Henning__> toady: hi sebastien [10:58] hey hey :) [10:58] r_open (n=root@58.65.224.5) joined #openser. [11:00] <__Henning__> toady: you are right with your last point, "libapr is another discussion".. lets see whats the opionion of the other devels about the autoconf/ cmake issue is, and then we could proceed [11:00] yep [11:01] whatever build system is no problem for me, as long as we have something that can handle the job easier [11:01] <__Henning__> i don't now thats the current state is such a problem for you [11:01] well [11:01] it is not _such_ a problem :) [11:02] I just want to make sure things are done well [11:02] and the first step towards libification is a build system [11:02] cmake is out of the question for me.. having your build system depend on c++ among other limitation doesn't make sense.. even as I don't like autocrap. it's the preferred choice if we have to move [11:02] otherwise that will make it tricky [11:02] Ded (n=ded@36.248-106-89.FTTH.rus-com.net) left irc: Read error: 104 (Connection reset by peer) [11:02] toady: you aren't the guy who dreams of a windoz port too are you? [11:02] codestr0m, well, developers can get anything required for building. That somethings users are not going to see [11:03] codestr0m, I don't mind a windows port since I am not going to use it :) [11:03] toady: some dependancies are rediculous.. and cmake while easier isn't standard imho.. it also doesn't build cleanly on a hardened system because of various reasons.. mostly not to blame on cmake [11:04] a building system is _not_ a dependency [11:04] <__Henning__> build dependency :-) [11:04] there are build time dependancies and run time [11:04] however if it doesn't build on some plateform we are targeting then this is a problem :) [11:04] run time ? come on :) [11:05] toady: ? [11:05] well, cmake is just for generating makefiles [11:05] <__Henning__> yes [11:05] so there is nothing impacting runtime [11:06] <__Henning__> or visual c project files.. [11:06] toady: did you even read what I wrote? [11:06] I said build dep [11:06] __Henning__, which can be very interesting if openser wants Windows [11:06] codestr0m, ok, you mind build dep then :) [11:06] who here is the one wanting a windoz port? [11:06] <__Henning__> i'm not :-) [11:06] good [11:06] codestr0m, I just think this is not a problem because this concern [11:07] codestr0m, well, this is a possibility [11:07] codestr0m, I am not a windows user though [11:07] well. I forget who it was saying this. I have to refer to past conversations, but I'm going to set ignore to whoever it was :) [11:07] in fact, I don't care of having openser on windows. Windows is a really dangerous platform. [11:08] codestr0m, ahah [11:08] it's a waste of time and resources [11:08] codestr0m, ok I think this way: [11:08] no one here is interested by windows [11:08] so no one is going to work on it [11:08] there was one person on the list that mentioned it.. [11:09] however, if one wants, why not letting him do the job ? [11:09] and maybe cmake is something to consider for this guy [11:09] <__Henning__> codestr0m: im build the hosting list (from friday), can you provide me your real name and your e-mail address? [11:09] toady: because once you open that box.. windoz users will always want it maintained. they will want binaries. they won't be able to give proper debugging. it's a ball of wax I have seen before [11:10] codestr0m, but they will do the job [11:10] codestr0m, let's put windows folks on work! [11:11] codestr0m, hum, maybe not otherwise they would have joined opensource ;) [11:11] toady: that may be true short term, but long term no guarantees.. they could abandon it. leaving the rest of us with a bunch of whiney users.. I won't even mention it again.. [11:11] <__Henning__> the mindset is different in the windows world, you are right [11:12] codestr0m, and that will be their problem :) [11:12] __Henning__: my offer for hosting was much less than anyone else.. I won't give a dedicated box. I was offering more a fully managed system I plug into my existing framework [11:12] and if they want it maintained, they will have to switch for a better OS [11:12] toady: are you a contributor on many projects? [11:13] codestr0m, quite a lot yes :) [11:13] codestr0m, I am maintaining Linux PAM along with 3 other folks [11:13] and how many have windoz versions? [11:13] codestr0m, and I work on prelude IDS [11:13] codestr0m, well Linux PAM.. hum :) [11:13] codestr0m, and Prelude is unix only too [11:13] <__Henning__> codestr0m: ok, i understand. so you don't want to included in this list? [11:14] so have you faced this problem before? [11:14] codestr0m, despite we'd like to have windows users [11:14] codestr0m, not at all [11:14] __Henning__: you can include me, but I'd rather not look like the cheap bastard and just want to make sure I'm clear.. most of this is security related [11:15] what do you think? [11:15] codestr0m, you faced problems with windows users already right ? :) [11:15] __Henning__: also was there a list of official requirements made up.. [11:16] <__Henning__> codestr0m: i understand [11:16] what ended up being the list of requirements.. [11:17] DanB (n=dan@87.139.12.167) joined #openser. [11:17] and I'm away for a few hours. I'll read when I get back.. let me know your email so I can send you more info if needed [11:17] Hello All! [11:20] Defraz (n=t0tal@24-116-152-177.cpe.cableone.net) joined #openser. [11:21] <__Henning__> codestr0m: henning.westerholt@1und1.de, but i think i'll just include that what you said to the list, security is important, and if you don't want this access, no problem. [11:23] <__Henning__> DanB: hey Dan [11:26] Hey _Henning_, congrats for the way the meeting on Friday went... [11:28] <__Henning__> DanB: thank you! It was a little bit chaotic sometimes, but i'm satisfied too [11:28] bdrubel (n=root@58.65.224.5) left irc: "ircII EPIC4-2.6 -- Are we there yet?" [11:31] bdrubel (n=root@69.88.13.17) joined #openser. [11:31] r_open_ (n=root@69.88.13.17) joined #openser. [11:31] dhakatel (n=ashrar@58.65.224.5) left irc: "I quit" [11:32] dhakatel (n=ashrar@58.65.224.5) joined #openser. [11:32] bdrubel (n=root@69.88.13.17) left irc: Client Quit [11:33] bdrubel (n=rubel@69.88.13.17) joined #openser. [11:36] <__Henning__> DanB: what do you think, for the next time: stay with this mode, or use set the channel to moderated? [11:39] Hello, anyone there who can help me with high CPU load and very high RAM consumption of openser? [11:42] __Henning__: I felt it was pretty well behavied and there was no need of moderation. There were some general questions, which had to be non-moderated... like what is your experience with OpenSER .. as far as I remember. Of course u should know better if you were able to achieve your targets... [11:43] miconda (n=daniel@81.180.83.75) joined #openser. [11:44] __Henning__: yes.. security for all my hosted apps is #1.. we have a number of strict policies I/we follow.. and just handing out dedicated unmanaged boxen for no real reason doesn't make sense. I'll email my details a little later today.. [11:45] <__Henning__> codestr0m: ack, good. I want to talk with you about the build services later you offered. I'll wait for you mail. [11:46] also nicox.. I'm going to guess based on what you've told me that openser isn't directly the problem.. it's a causality from something else.. such as db lookups blocking and that causing this cascade.. I have seen similar.. I could be totally wrong though.. [11:47] <__Henning__> DanB: targets.. :-) The agenda in the wiki was open for everyone to object or edited, so i hope you don't think of it that the purpose of this meeting was to get my targets through. I see the purpose of this meeting more on the community side. [11:48] r_open (n=root@58.65.224.5) left irc: Read error: 110 (Connection timed out) [11:50] __Henning__: missunderstand, I was thinking you from the point of view of moderator ... not as your own purposes, but community ones... [11:53] <__Henning__> DanB: ok, i also don't want to be misunderstood :-). Everyone in the opensource business faces these problems: company vs. community.. [12:03] __Henning__: gotcha , I faced many similar issues myself... [12:04] r_open_ (n=root@69.88.13.17) left irc: "ircII EPIC4-2.6 -- Are we there yet?" [12:21] miconda (n=daniel@81.180.83.75) left irc: Remote closed the connection [12:48] Nick change: __Henning__ -> _Henning_ [12:48] Nick change: _Henning_ -> henning__ [12:50] Nick change: henning__ -> __Henning__ [12:53] can anybody help me about the memory use of openser? [12:54] why the memory use increasing by sequential call [12:55] is there any memory bug or my configuration error? [13:05] <__Henning__> bdrubel: what version do you use? [13:05] i have the same problem, openser uses 1.5 Gig of memory [13:06] i'm using openser 1.2.0 [13:06] <__Henning__> hmm, anything special in your setup? Are you using presence? [13:08] i'm using openser 1.1.0 [13:09] no I am not using presence [13:09] but using dialog module [13:12] <__Henning__> nicox, there was a memory leak/ fragmentation issue with 1.2.0 related to AVPs, see the user list on 15.03 [13:14] but I use 1.1.0 is there any memory leak? [13:14] <__Henning__> sorry, i don't remember anything, have you tried to reproduce this with the current 1.1 branch? [13:15] <__Henning__> or at least 1.1.1? [13:15] no [13:16] I also found Memory debug but I can't get any problem. [13:16] then why the memory is rising? [13:17] <__Henning__> bdrubel: probably a bug :-) [13:17] <__Henning__> bdrubel: how much memory the server uses? [13:17] yap I think that [13:18] 1GIG [13:18] but for each 500 concurrent call increasing memory [13:21] <__Henning__> bdrubel, there was two fixes regarding memory leaks in 1.1.1, perhaps you can try this version? [13:21] thanks [13:21] <__Henning__> bdrubel: sorry, i need now to get some work done.. and if the problem persists, please report your problem to the user list [13:21] ok thanks [13:36] NormB (n=NormB@smoothwall.goes.com) joined #openser. [14:12] miconda (n=daniel@81.180.83.75) joined #openser. [14:19] i only use modparam("nathelper", "received_avp", "$avp(i:42)") [14:28] bdrubel (n=rubel@69.88.13.17) left irc: "ircII EPIC4-2.6 -- Are we there yet?" [14:51] __Henning__ (n=henning@et-1-5.gw-nat.bs.ka.oneandone.net) left irc: Remote closed the connection [14:51] DanB (n=dan@87.139.12.167) left irc: Read error: 104 (Connection reset by peer) [14:58] Chris-NB (n=chris@ng1.kurtkrenn.com) joined #openser. [14:58] hi [14:58] how can I integrate asterisk as voicemail system into openser? [14:58] in asterisk I can say, dial that nr for 30 seconds, after that, dial voicemail [14:59] is this possible with openser? [14:59] if user do not pick up the phone in defined time, call voicemail at asterisk? [15:01] you can implement that using call forwarding (by playing with tm timers and failure routes) [15:02] I've just came around that. (tm timers) [15:25] dhakatel (n=ashrar@58.65.224.5) left irc: "Hey! Where'd my controlling terminal go?" [15:27] osochebol (n=osochebo@58.186.22.133) joined #openser. [15:29] L|NUX (n=linux@unaffiliated/lnux/x-10290) left irc: "Leaving" [15:46] osochebol (n=osochebo@58.186.22.133) left irc: "Leaving" [15:48] L-info (n=l-info@87-194-69-156.bethere.co.uk) joined #openser. [15:49] erider (n=erider@unaffiliated/erider) left irc: Read error: 110 (Connection timed out) [15:50] erider (n=erider@unaffiliated/erider) joined #openser. [16:31] Defraz (n=t0tal@24-116-152-177.cpe.cableone.net) left irc: "Trillian (http://www.ceruleanstudios.com" [16:35] Chris-NB (n=chris@ng1.kurtkrenn.com) left irc: Read error: 113 (No route to host) [17:25] Defraz (n=t0tal@24-116-152-177.cpe.cableone.net) joined #openser. [17:52] Defraz (n=t0tal@24-116-152-177.cpe.cableone.net) left irc: Read error: 104 (Connection reset by peer) [17:53] bdrubel (n=xman@202.53.162.177) joined #openser. [17:54] bdrubel (n=xman@202.53.162.177) left #openser. [18:05] _0penser_ (n=xman@202.53.162.177) joined #openser. [18:08] escribzz (n=escribzz@VDSL-130-13-37-65.PHNX.QWEST.NET) left irc: "Trillian (http://www.ceruleanstudios.com" [18:29] __Henning__: ping [18:32] osochebol (n=osochebo@58.186.22.133) joined #openser. [18:33] in regards to the build service.. I was talking with one of the maintainers for a rather large *nix dist and he pointed me to a much better solution which looks like it will be easy to build bins for Debian Etch, FC 5/6, Mandriva, SLES 9/10 and Opensuse 10.. RHEL is missing, but that's for political reasons atm.. I haven't asked further [18:34] DoberMann_ (n=james@AToulouse-156-1-163-244.w90-38.abo.wanadoo.fr) joined #openser. [18:35] toady (n=toady@251.9.39-62.rev.gaoland.net) left irc: "Leaving" [18:38] <_0penser_> can anybody help me about openser memory lick [18:38] _0penser_: are you using 1.2? [18:38] <_0penser_> yes [18:40] <_0penser_> I think 1.1 contain memory lick [18:41] <_0penser_> is it true? [18:41] earlier someone said it was discussed on the ml.. on 15.03 or something.. I haven't looked.. [18:41] using valgrind you can find out for sure [18:43] <_0penser_> in openser have builtin malloc debugger [18:44] I'm reading about it on the wiki.. I haven't had to face this issue, but looks like others are having it pop up [18:44] <_0penser_> yes u r right [18:45] <_0penser_> it is in wiki [18:46] DoberMann[PullA] (n=james@AToulouse-156-1-118-28.w90-30.abo.wanadoo.fr) left irc: Read error: 110 (Connection timed out) [18:46] <_0penser_> I also use dialog module is there any problem like memory lick [18:47] _0penser_: well.. if you're reading the wiki and have access to valgrind.. maybe you can test and report back.. ? [18:48] DanB (n=danbogos@87.139.12.167) joined #openser. [18:49] <_0penser_> ok thanks [18:56] qdk (n=qdk@213.150.62.32) left irc: Read error: 113 (No route to host) [19:08] osochebol (n=osochebo@58.186.22.133) left irc: "Leaving" [19:14] qdk (n=qdk@193.164.155.27) joined #openser. [19:23] CunningPike (n=CunningP@dhcp-10-234.district.north-van.bc.ca) joined #openser. [19:39] miconda (n=daniel@81.180.83.75) left #openser. [19:52] Defraz (n=t0tal@fw.fuzecore.com) joined #openser. [19:55] _0penser_ (n=xman@202.53.162.177) left irc: "ircII EPIC4-2.6 -- Are we there yet?" [19:56] good evening boys & girls [20:02] Nick change: DoberMann_ -> DoberMann [20:38] OK, I want to have a machine that I funnel all my outbound calls thru. I want that machine to load balance (and retry as necessary) across my 4 SIP providers. Is OpenSER the right tool? If not, recommendations? [20:39] My peak call loads are about 2000 simultaneous calls and a call initiation rate of 10 calls per second... [20:41] wwalker: the recommendations will be always for openser here... [20:41] hehe [20:41] DanB: :) so, all I need is 20 to 50 lines of config to route across 3 or 4 providers? [20:42] depinds of what u want to do, u can make it with only one line: t_relay() if you will feel better... [20:43] there's no openser standard config [20:43] you will need to "code" your config [20:44] choose the appropriate routing policy and then implement it on the openser.cfg [20:44] that's the beauty of openser ;-) [20:44] I have in my config 938 lines after "optimisation" [20:44] any fun debugging that? :p [20:45] no need of debuging anymore, now I dream it ;-) [20:46] mine: 941 for openser.m4 => 784 for openser.cfg [20:47] wwalker: you will be amased to find out that openser doesn't implement sip... you need to know sip to make openser implement this ... openser will give u just the tools to do it [20:47] osas: so we will need to define a good config arround 800 then ;-) [20:47] heh [20:49] I will need to add around 200 more lines for the m4 config file (including comments) [20:50] so where is the list of providers that t_relay uses defined? [20:50] wwalker: are you familiar with openser? [20:51] escribzz (n=escribzz@VDSL-130-13-37-65.PHNX.QWEST.NET) joined #openser. [20:52] osas: trying to get there. [20:52] in the topic.. should we include the link to that automagical config generator.. does anyone think it would be useful to others? [20:53] wwalker: you can start from here: http://old.iptel.org/ser/doc/seruser/seruser.html [20:53] codestr0m: yes. ;-) [20:53] once you get the basic, read the documentation for modules on the openser website [20:53] codestr0m: it should be already on the openser website [20:54] and by the way, read several times the SIP rfc ... [20:54] http://www.openser.org/mos/view/News/NewsItem/OpenSER-Configuration-Generator-for-v1.2.0/ [20:54] after that, you are ready to create/debug your scripts ;-) [20:56] I think the best tutorial on openser/ser is from the old iptel.org website: http://old.iptel.org/ser/doc/seruser/seruser.html [20:57] that will give anyone the basic understanding of openser/ser config engine [20:57] after that, it is just module documentation reading [21:10] Nick change: DoberMann -> DoberMann[Flim] [21:14] qdk (n=qdk@193.164.155.27) left irc: Read error: 110 (Connection timed out) [21:15] OK, so - will OpenSER, by itself, with a good config file, do what I need (act as a gateway between my non-routable call generation servers and my internet available voip providers)? [21:16] Before I burn 2 days reading docs... [21:17] wwalker: openser together with a rtpproxy module (rtpproxy or mediaproxy - by your wish) [21:23] OK, anyone want to help me get this up and running? A config file for openser and for rtpproxy? price? I tried the sipwise tool - "Application error (Rails)" on the very last page [21:28] JerJer[mobile] (n=jj@199.45.11.90) joined #openser. [21:28] here's another start: http://www.voip-info.org/wiki/view/OpenSER [21:30] wwalker: if you want commercial support check out http://www.voice-system.ro [21:30] qdk (n=qdk@0x50c627be.bynxx11.adsl-dhcp.tele.dk) joined #openser. [21:32] wwalker: for questions we can happily help... [21:34] DanB: have you read the sip rfcs? It's like 200 pages and a total mess.. unless you're subscribed to the sip implementors list. I'd never recommend such punishment ;) [21:35] so has anyone ever noticed that if one starts openser directly (no fork, log to stderr) that TCP and TLS support does not get bound to the IP address? [21:35] v1.2 [21:36] JerJer[mobile]: I did after 2 days of network down :-( [21:36] ugh [21:36] Oooops ... [21:37] JerJer[mobile]: same story for 1.1 [21:37] is there like a bug depository somewhere ? [21:37] or should i just drop a message to the dev list ? [21:38] or is this a feature ? [21:38] :) [21:38] luckily I had not so much tcp uas in the network [21:41] rt [21:41] codestr0m: not completly read them, only parts... [21:43] the sip rfcs suck, they say nothing :) [21:43] oh no, they say all [21:43] but nothing [21:43] JerJer: the fork=no option works that way by design [21:43] JerJer[mobile]: I was crying some days ago that 1.2 version has a bug in TCP, but it came to be this forking issue... I didn't get the time to look through docs if this should be the way or not ... since it is inherited from before, not new... so not sure if it is to be reported as a bug, or not [21:44] anyway, I just know that the openser reports as bounded to TCP, but netstat shows not... [21:44] http://www.openser.org/dokuwiki/doku.php/core-cookbook:1.2.x#fork [21:44] unspin: thxs, it looks like a bug then... [21:46] dhakatel (n=dhakatel@202.168.246.206) joined #openser. [21:47] that url doesn't make any reference to not binding to TCP ports when fork == no [21:49] JerJer[mobile]: yes, this is the reason I came to the conclusion it should based on documentation... [21:51] JerJer[mobile]: you were asking for the link for submiting new bug: http://sourceforge.net/tracker/?func=add&group_id=139143&atid=743020 [21:53] danke [21:54] err, what i mean is that, since OpenSER spawns seperate processes to handle incoming udp and tcp connections [21:54] the behaviour of fork=no would have the side effect of not listening on tcp (if the listen statement doesn't specify a protocol, udp would start up first) [21:57] but the print statement states that its listening on the various methods [21:58] giving one the impression that it ~should~ be listening on TCP/TLS [22:01] has anyone here used shype of palmOS [22:02] unspin: I remember that I have specifically bind openser to tcp, no udp, but still was not running on tcp [22:03] bound, sorry [22:03] if the first listen statement is explicitly set to listen on tcp does it the problem occur? [22:03] JerJer: i agree, its pretty vague and the documentation doesn't fully describe the behaviour [22:04] unspin: this is what I remember... [22:04] let me check [22:04] Good morning all [22:04] hello [22:05] hmm [22:05] with only a tcp listen statement defined it wont start (1.1) [22:05] no fork mode and no udp listen address found [22:05] seems that its designed to only listen on udp when fork=no is defined [22:06] good night all :) [22:07] we should probably update the wiki with this information [22:08] unspin: shall I repeat that I remember that one also now :-) [22:08] unspin: leme test that [22:08] ok [22:08] are you running 1.2? [22:08] JerJer[mobile]: yes [22:08] unspin: yes [22:09] svnrevision: 2:1916M [22:09] version: openser 1.2.0-tls (i386/linux) [22:10] ok on 1.2 i get udp only [22:10] with a listen = tls:foo first [22:10] alright, i'll update the wiki so at least the current behaviour is documented [22:11] i'm gonna blog about it too [22:13] ok guys, looks like my "kopf" doesn't help me anymore, so I will go offline now... wish you all a good night/day! [22:14] DanB (n=danbogos@87.139.12.167) left #openser. [22:15] the wiki has been updated, can you have a look to verify that it describes the behaviour you are seeing? [22:31] Chris-NB (n=chris@85-126-34-233.static.xdsl-line.inode.at) joined #openser. [22:38] escribzz (n=escribzz@VDSL-130-13-37-65.PHNX.QWEST.NET) left irc: Read error: 110 (Connection timed out) [22:42] Chris-NB (n=chris@85-126-34-233.static.xdsl-line.inode.at) left irc: Read error: 54 (Connection reset by peer) [22:50] NormB (n=NormB@smoothwall.goes.com) left irc: "Leaving" [22:59] L-info (n=l-info@87-194-69-156.bethere.co.uk) left irc: Read error: 110 (Connection timed out) [22:59] Chris-NB (n=chris@85-126-34-233.static.xdsl-line.inode.at) joined #openser. [23:00] Nick change: DoberMann[Flim] -> DoberMann [23:22] Nick change: DoberMann -> DoberMann[ZZZzzz [23:54] stimpie (n=michiel@ip565faf27.direct-adsl.nl) left irc: Remote closed the connection [00:00] --- Tue Apr 24 2007