IRC log for #asterisk on 20180927

00:11.18*** join/#asterisk war9407 (war@pool-173-72-178-211.clppva.fios.verizon.net)
00:21.09*** join/#asterisk infobot (ibot@208.53.50.136)
00:21.09*** topic/#asterisk is #asterisk The Open Source PBX and Telephony Platform (asterisk.org) -=- LTS: 13.23.1 (2018/09/20) 16.0.0-rc3 (2018/09/20), Standard: 15.6.1 (2018/09/20); DAHDI: DAHDI-linux 2.11.1 (2016/03/01), DAHDI-tools 2.11.1 (2016/03/01); libpri 1.6.0 (2017/01/27) -=- Wiki: wiki.asterisk.org -=- Code of Conduct: bit.ly/1hH6P22 -=- Logs: bit.ly/1s4AKKu
00:38.31*** join/#asterisk r0ckp3arl_ (~admin@203.149.66.146)
02:35.55*** join/#asterisk [TK]D-Fender (~joe@64.235.216.2)
02:46.48*** join/#asterisk duo_ubuntu (b61ecc7e@gateway/web/freenode/ip.182.30.204.126)
04:30.54*** join/#asterisk cemotyz09 (~cemotyz09@cpe-70-121-128-59.satx.res.rr.com)
04:55.23*** join/#asterisk luckman212_ (luckman212@unaffiliated/luckman212)
05:01.27*** join/#asterisk luckman212_ (luckman212@unaffiliated/luckman212)
05:08.06*** join/#asterisk duo_ubuntu (b61ecdf2@gateway/web/freenode/ip.182.30.205.242)
05:36.34*** join/#asterisk jkroon (~jkroon@165.16.204.162)
05:40.53*** join/#asterisk luckman212 (luckman212@unaffiliated/luckman212)
05:45.51*** join/#asterisk sibyakin (~sibyakin@188.162.228.107)
05:49.04*** join/#asterisk gerhard7 (~gerhard7@ip5657ee30.direct-adsl.nl)
05:54.51*** join/#asterisk samwierema (~samwierem@195.240.143.134)
06:32.34*** join/#asterisk Penguin (~xwQ5kwYl6@our.systems.are.full.of.penguins.at.penguinsystems.net)
06:50.39*** join/#asterisk war9407 (war@pool-70-106-229-104.clppva.fios.verizon.net)
07:25.15*** join/#asterisk jkroon (~jkroon@165.16.203.126)
07:32.00*** join/#asterisk samwierema (~samwierem@195.240.143.134)
07:50.11*** join/#asterisk mou (~donwillia@37.97.50.103)
08:05.47*** join/#asterisk hehol (~hehol@gatekeeper.loca.net)
08:20.10*** join/#asterisk DanB (~DanB@clt-195.192.206.178.ip-anschluss.net)
08:58.42*** join/#asterisk miralin (~Thunderbi@194.8.128.67)
09:55.52*** join/#asterisk Mr_Pleb_Mgoo (~jakeb@103.46.212.49)
10:21.55*** join/#asterisk pa (~pa@unaffiliated/pa)
10:48.28*** join/#asterisk lankanmon (~LKNnet@CPE64777d632383-CM64777d632380.cpe.net.cable.rogers.com)
12:10.02*** join/#asterisk sahmed (~sahmed@37.19.10.33)
12:11.14sahmedanyone experience "503 Service Unavailable" during browse http://doxygen.asterisk.org/
12:12.44fileit's been brought up on the dev list awhile ago, http://lists.digium.com/pipermail/asterisk-dev/2018-September/076983.html
12:13.42sahmedcould not understand if I search anything on google just point out those links
12:15.04sahmedok understood the issue, thanks @file
12:43.13*** join/#asterisk sekil (~sekil@nat-73.net011.net)
13:08.34*** join/#asterisk [TK]D-Fender (~joe@216.191.106.165)
13:16.12*** join/#asterisk brad_mssw (~brad@66.129.88.50)
13:22.17*** join/#asterisk scroll2 (sid313090@gateway/web/irccloud.com/x-eriuyigddvipdmjj)
13:49.29*** join/#asterisk AsteriskRoss (259d3426@gateway/web/freenode/ip.37.157.52.38)
14:00.26*** join/#asterisk yokel (~yokel@unaffiliated/contempt)
14:00.33*** join/#asterisk kharwell (kharwell@nat/digium/x-zwvnkbexigppsldr)
14:00.33*** mode/#asterisk [+o kharwell] by ChanServ
14:20.52*** join/#asterisk n8n8 (~n8@208.69.84.142)
14:26.03n8n8Hi all. I'm in a bit of a bind, my customers cannot make outbound calls this morning. I have a complicated setup. I believe the issue is related to my realtime config. The delta in my environment is that for about 2 hours last night asterisk would have been connected to my read-only replica database, which is where realtime would access the pjsip configs, as well as update them. After I restored the switch, I've restarted asterisk, and it is now talking to the
14:26.03n8n8primary RW db server for realtime. Inbound calls all seem to be working, but when an endpoint tries to dial out, it is always being told it is unauthorized. What I need the most help with is how to troubleshoot why asterisk/pjsip is determining the endpoint is unauthorized. The endpoints are able to register, and receive calls, just not send...
14:27.36*** join/#asterisk Da-Geek (~Da-Geek@81.145.134.175)
14:28.29*** join/#asterisk bford (uid283514@gateway/web/irccloud.com/x-gscgjgtkwjllzbhq)
14:28.30*** mode/#asterisk [+o bford] by ChanServ
14:45.42*** join/#asterisk rmudgett (rmudgett@nat/digium/x-ltnhavzqmfevviwa)
14:45.42*** mode/#asterisk [+o rmudgett] by ChanServ
15:05.53*** join/#asterisk cemotyz09 (~cemotyz09@cpe-70-121-128-59.satx.res.rr.com)
15:09.10igcewielingI can set the values of multiple dialplan variables at the same time using ARRAY when using AGI.    Can I GET values of multiple variables at the same time?
15:12.03*** join/#asterisk cresl1n (uid299068@asterisk/libpri-and-libss7-expert/Cresl1n)
15:12.03*** mode/#asterisk [+o cresl1n] by ChanServ
15:17.18SamotWith AGI?
15:17.42SamotI mean you could break out into an AGI script to write code that would set all them via a loop...
15:17.59SamotBut I don't think Asterisk has the concept of arrays in that regard.
15:18.17*** join/#asterisk Han (~han@unaffiliated/han)
15:26.04sibiriaigcewieling: if you're already in an AGI application you most certainly have handier options for pulling variables
15:26.42sibiriai'm trying to figure out what you actually want to do
15:26.51sibiriayou want to read multiple variables, so why not just read them?
15:30.27n8n8anyone have any thoughts as to why an endpoint would be able to register with pjsip, and be able to receive calls, but not be able to make them? When they try to call they are getting an unauthorized response...
15:30.56Samotn8n8: Show a failed called with the log and pjsip logger enabled.
15:31.51igcewielingsibiria: nothing all that important.   Having a zillion GETs and SETs can cause excessive AGI debug output.  I've reduced the issue by using ARRAY with SET when setting variables.   I was wondering if I could do the same for GETing variable.
15:32.21*** join/#asterisk stefanauss (~stefanaus@host222-246-static.59-217-b.business.telecomitalia.it)
15:32.29sibiriaigcewieling: ah, no, you can't in any other way than concatenating the channel variables yourself and then parsing them afterwards
15:33.16sibiriaas Samot pointed out, the asterisk dial plan "language" doesn't really have a concept of arrays in that sense
15:33.38igcewielingMy SETs look like SET VARIABLE ARRAY(MASTER_CHANNEL(sm_call[1][start_timestamp]),MASTER_CHANNEL(sm_call[1][call_direction]),MASTER_CHANNEL(sm_call[1][btn]),MASTER_CHANNEL(sm_call[1]....
15:33.38sibiriathe ARRAY function really only translates to MSet() effectively. same thing
15:33.54igcewielingI create my own arrays.
15:34.06igcewielingthat wasn't my question
15:34.15SamotNo.
15:34.24SamotThere is nothing to do the opposite of what ARRAY does.
15:34.39igcewieling*nod* I accept that.
15:34.41sibiriaonly thing you can do is to build your own CSV string and then deparse it
15:34.50sibiriathat will be one "function call"
15:36.33SamotWell.....
15:37.05SamotIf in the first part you SET six different VARS with ARRAY and in the second part you want to GET those VARs.....
15:37.30SamotWhy do you need a mass GET?
15:37.37SamotYou can just call on them. They're set.
15:37.49sibiriahe wants to reduce the log output by not issuing 20 gets in a row
15:38.21SamotBut they are already SET..
15:38.28n8n8Here is a snapshot of the full log with at least one test call being made (live system, lots of things trying). https://pastebin.com/U7SuYmux
15:38.32SamotSo wouldn't just doing ${VAR} be getting it?
15:38.48sibiriahe probably wants them in his AGI script, and an AGI instance doesn't automatically import channel variables
15:38.49n8n8I suspect there is some kind of data corruption in the realtime tables....
15:38.59sibiriayou need to make a Get call for each one
15:39.32sibiriaso one way would be to build that string in the dialplan in advance
15:39.39sibiriaf.e. by Set(BLAH=${abc};${def};${ghi};etc.)
15:39.50sibiriaand then Get it from the AGI script and split it accordingly
15:40.04sibiriaresult: one Get, not 20 Gets
15:40.17SamotLook at his set...
15:40.24SamotHe's setting MASTER_CHANNEL vars.
15:40.42SamotSo he would need two SETs.
15:41.09*** join/#asterisk stefanauss (~stefanaus@host222-246-static.59-217-b.business.telecomitalia.it)
15:42.07Samothttps://www.irccloud.com/pastebin/l3XUbjDD/
15:42.18Samotn8n8: The issue is the other side.
15:42.26SamotYour PBX is sending back a 401
15:42.48SamotInstead of re-sending the INVITE with auth headers the other side is just ACKing the 401.
15:43.31SamotUser-Agent: Grandstream GXP2160 1.0.7.97 <-- So look at the Grandstream.
15:44.08n8n8Samot: this worked yesterday... and the phones didn't change, d
15:44.20n8n8sorry about the extra d...
15:44.24SamotOK.
15:44.29SamotHave you rebooted the phone?
15:45.05n8n8not the NKB one, but I have had some ppl reboot them, and no change. The configs are centrally managed, and fetched on reboot
15:45.19n8n8the passwords match ps_auth and the config files
15:45.40n8n8the phones can register, just not call out...
15:45.45SamotWell you got other issues as well
15:46.06Samothttps://www.irccloud.com/pastebin/h9z7nxEi/
15:46.28n8n8yes, I saw the OWC thing with the internal server error, but that seems isolated to them, everyone is having the auth issue
15:46.28SamotI see a bunch of SUBSCRIBEs then 401s and when the SUBSCRIBE is resent with proper auth headers....
15:46.35SamotYou're returning 500 Server Errors.
15:47.04Samotn8n8: You sent the grandstream phone a 401 challenge. That's what should happen.
15:47.38SamotThe phone sent back an ACK
15:47.46SamotNot a new INVITE
15:48.07n8n8I believe the subscribe 401's have always been that way....
15:48.41SamotWhat that you respond to a proper subcribe request with a 500 error?
15:48.48SamotThats wrong.
15:49.55SamotSo this issue is for one client? Or all users?
15:50.01n8n8OK, I see what you mean about the ACK to the 401 on INVITE, but, most imporantly at this time is getting back to my working, if crappy, setup...
15:50.08n8n8all users AFAICT
15:50.14igcewielingsibiria: This does not work in AGI GET VARIABLE ${abc}
15:50.51igcewielingAGI GET VARIABLE abc is correct, but it does not support something like AGI GET VARIABLE abc,myvar2
15:51.01SamotSo all users are not replying to 401 challenges on INVITEs?
15:51.13n8n8That seems to be the case.
15:51.26igcewielingthis is a minor optimization I thought about, not some important feature I desperatly need.
15:51.29SamotThey arent sending new INVITES?
15:51.37n8n8Samot: Correct
15:52.18sibiriaigcewieling: yes i know, neither the dial plan onr AGI was ever spec'd to have a notion of arrays. it's why i told you the only way to avoid having to issue loads of GETs is to build a CSV string in the dial plan, and then split it up in the AGI script
15:53.39n8n8so far as I know, no config has been changed, only a 2 hr window of realtime state being "lost" while asterisk was pointed at the readonly replica DB.
15:54.26SamotSo phones have been rebooted?
15:54.51n8n8Many have, yes, but they'd pull the same config as before the DB pointing issue.
15:56.38SamotAre the phones pulling their profiles from you?
15:57.20n8n8yes. I generate them, and my PBX DB config at the same time (when I deploy them). These generally do not change.
15:57.29SamotOk.
15:57.35n8n8If they do, it is explicit and done by me.
15:57.45SamotSo have yoy used a older profile and have them pull that?
15:57.58SamotBecause the phones are the issue
15:58.42n8n8there is only one profile for each phone, and none were changed in the last few days. each phone has its own MAC addr named config file.
15:58.55SamotDo yoy have access to a phone?
15:58.59n8n8yes
15:59.04SamotCheck it
15:59.29n8n8that is the NKB_7162004262_104 phone.
15:59.42SamotThey could have made a pull request during your window of error and screwed up their settings
16:00.42n8n8the phone configs are static files, not generated on request.
16:01.47SamotOk look man
16:01.52SamotThe phones are the issue
16:02.09SamotEither look at one to see what is up or dont
16:02.25SamotThey are not handling the 401 properly
16:05.26n8n8I'm not sure what I'm going to look for in the phone config. Most settings are left to defaults, my configs are very small, and do not touch most settings. I've just factory defaulted one phone, and it should grab its config, but I do not expect a change in behavior. I know that my overall setup has some problems, I am just trying to get back to a working state. I get where you are coming from with pointing at the phones, but if this one comes back from a facto
16:05.26n8n8ry default + reprovision and still does not work, I think there is something deeper going on
16:08.10Samotn8n8: What have you actually done with a phone that is having this issue? Troubleshooting wise? What steps have been taken?
16:11.15n8n8Samot: so far, I've watched the pjsip logs, and run packet captures. I've tested inbound calls, and done/undone small DB config changes to observe different behaviors. The most interesting so far has been to add an entry to endpoint_identifier_ip to see a different query flow
16:11.31n8n8Also, turned on db query logging for this database.
16:11.57n8n8From a phone side, I'm really not sure what else to modify / test. I'm completely open to suggestions
16:12.50*** join/#asterisk twanny796 (~Twanny@195.158.111.16)
16:13.08n8n8I'm trying not to get too carried away with changing things without restoring them back to where they were, as yesterday everything was working....
16:13.37n8n8This phone is almost ready to re-provision.... the grandstreams take such a long time to reset & boot
16:15.37*** join/#asterisk twanny796 (~Twanny@195.158.111.16)
16:17.43SamotHave you tried your own phone?
16:18.01SamotDo you have issues calling out the system from a test phone?
16:19.37n8n8Samot: factory default and reprovision complete. the phone I rebuilt is: NKB_7162004262_102 here is the log data from a test call on that phone (on my desk next to me): https://pastebin.com/TcbARHRi
16:20.09n8n8yes, this phone cannot be called out of. I'll catch an inbound to it and paste that so you can see
16:23.24SamotDid you use your profile?
16:23.57SamotYeah, you did. You said.
16:23.58n8n8Here is the inbound call to this phone. https://pastebin.com/R23H6ABc
16:24.21SamotSo now factory default and manually set the creds up
16:24.48SamotInbound and outbound ate not the same
16:24.48n8n8ok
16:24.54SamotAre*
16:32.17n8n8Samot: hand provisioning does the same thing
16:32.59SamotAfter a factory reset?
16:33.22n8n8correct. I just put in the sip server, username and password
16:33.30SamotShow the call
16:34.03n8n8ok. I there a way to isolate the call in the logger?
16:34.59Samotpjsip set logger ?
16:35.14SamotTo get all the options but yes you can limit it to an IP/contact etc
16:35.43n8n8https://pastebin.com/pdszg3tT
16:36.03n8n8ah, so host will limit to just one client IP...
16:36.53n8n8sorry to put you through all the extra junk in there then
16:40.43n8n8Here is a dump of the logs with pjsip limited to just my IP, but this time I let the phone time out to the fast busy tones: https://pastebin.com/mvb6S2v7 in case that provides any more info
16:42.17SamotShow the AOR setup for this device.
16:42.54SamotI'm unsure why all these Grandstreams are not sending a new INVITE when challenged.
16:43.13SamotYou've always challenged them before right?
16:43.47n8n8here you go: https://pastebin.com/hb1VuRF3
16:44.33n8n8I'm honestly not sure I've always challenged them before. That's why I'm trying to dig into the realtime tables in the DB. What settings would affect that?
16:45.00SamotWell do the endpoints have an auth setting?
16:45.41n8n8in their config or in the ps_endpoints config?
16:45.50n8n8their being the XML file for the phone
16:46.07SamotIn their PJSIP config
16:46.15SamotHave they always had an auth setting?
16:48.04n8n8well, nothing should have been changed. I know that we didn't run anything to change the data in the realtime tables. I'm really not sure what asterisk gets up to with realtime though
16:48.26SamotYou're not answering my question.
16:48.45SamotI'm asking if you've always auth challenged your end users for outbound calls.
16:49.38n8n8I don't know. I'm checking the DB table now to see what is in there and checking my config gen code to see what it puts in there
16:50.08SamotYou don't know you're own security setup for outbound call requests?
16:51.01n8n8not off the top of my head. I built this system a few years ago....
16:51.19SamotUhm.
16:51.26SamotOK.
16:52.24n8n8we're a small WISP. VOIP is an additional feature customers demanded. I was too cheap to buy a system, so I built a multi-tenent one, that has worked pretty well. Most of our security is managed by firewalls as wo own the network
16:52.30n8n8we*
16:52.45SamotOK
16:53.14SamotWell I can't see anything wrong with the 401 challenge.
16:53.39SamotOutside of the fact the phones are ACKing it instead of re-sending a new INVITE with auth headers.
16:54.12SamotI don't use Real-Time because, why. It's not really a benefit overall.
16:55.15n8n8I had big plans to put a web-UI and self service on top of this, and web+db is familiar to me. Perhaps I could have done all of it with flat files....
16:55.52n8n8it looks like I've always been populating the auth field
16:56.42n8n8could some other setting precluded the auth from being looked at? when I messed with the endpoint_identify_ip table, I saw a change in behavior
16:59.06SamotDo you have a non grandstream?
17:00.01n8n8no... when I took the auth value out of ps_endpoints, outbound worked.
17:00.08SamotYes
17:00.19n8n8something must have been overriding / ignoring that...
17:00.27SamotBecause youre not challenging them anymore
17:00.33SamotDude
17:00.45n8n8I get that. the question is why wasn't I
17:00.51SamotThe phones were not responding to the auth challenge right
17:01.03n8n8that I also understand
17:01.36SamotBecause now youre just open for fraud
17:02.23n8n8I'm clearing the field to restore service. I agree, and I'll figure out the correct way to have the auth working
17:02.53n8n8what I'd like to understand is how I would have had that populated, but it was apparently being ignored....
17:03.27SamotWell who said it was?
17:03.43SamotWho is to say that yesterday they were responding to auth challenges right
17:04.37n8n8No, I think that yesterday there were no auth challenges for reasons TBD, and that for some reason pjsip decided to start sending them
17:04.46n8n8that's my current theory
17:05.06n8n8as that config field is clearly being populated
17:06.04n8n8I think there is something in the pjsip identifiers
17:06.37n8n8that was preventing authentication. (I'm not saying it was correctly set up, just that it might have been there)
17:08.33igcewielingIf I had this issue, I'd manually generate the pjsip conf file from the realtime DB so we can get all the database bs out of the way.
17:08.55n8n8Samot: In any case. I really appreciate you're taking the time to go through all this with me. I'd be glad to send you some beverages of your choice.
17:09.16n8n8My problem is not the DB, but which data to look for in it.
17:09.19SamotWhat igcewieling said
17:10.17SamotAnd I would also update from a know bug/security vulnerability version to something current.
17:10.23Samot13.17 had some serious bugs
17:10.28n8n8I've been a DB guy for far longer than I have asterisk (as I'm sure that's apparent by now....)
17:10.36n8n8oh, that's good to know....
17:10.43SamotNot to mention the RTP bleed issue that was fixed for all previous version before 13.17
17:10.47SamotIncluding 11
17:10.51n8n8I wonder how many of them I depend upon
17:11.13SamotWell the current version is 13.23.1
17:11.28SamotYou're on a version that is almost 2 years old.
17:12.19n8n8I'll build and get that into my test env... as I said, I put this together some time ago, and pretty much locked it behind our firewalls.... I guess it's time to update.
17:12.26SamotWell
17:12.52SamotThe "set it up, forget about it and let it run" plan works great until something happens.
17:13.44SamotAnd you find out you are now seriously behind in releases and that support for an out of date system becomes rare to find.
17:13.51n8n8Well, it has always worked better for asterisk than it has a great many other things I depend upon.... * is some solid software in my experience
17:14.01SamotI'm not say it's not solid.
17:14.05n8n8it being set and forget
17:14.07SamotI've used it for over a decade.
17:14.18SamotBut that doesn't mean it shouldn't be kept current.
17:14.30SamotOr at least in a state where it's not open to bugs and security holes.
17:14.53SamotOr in a state where you can't remember a thing about the setup because you've left it sitting for 2 years.
17:15.19SamotHard to troubleshoot something that is broken if you can't remember how it worked correctly in the first place.
17:16.34SamotOtherwise you end up with a client base that can't make outbound calls for half the day because you weren't sure how it was working correctly or supposed to be working correctly.
17:16.57*** join/#asterisk degenerate (~degenerat@S0106cc2de0099182.no.shawcable.net)
17:17.15n8n8agreed. I've just been burnt by so many other packages over the years when an upgrade breaks something, that I'm always reluctant. but this just goes back to having solid test infrastructure in place.... the problem with being a small business owner.... its really tricky to be as thorough as one should
17:18.00SamotI am a small business owner running a voice network for multiple small ITSPs/MSPs....
17:18.23SamotOn top of 90% of their networking.
17:19.19SamotUnfortunately, I've had to make a rule about dealing with WISPs overall.
17:19.33Samot"We want to add voice to our services"
17:19.47SamotAnd that right there is the most planning or thought they put into it.
17:21.37SamotI mean outside of the fact they generally have zero voice experience among them....
17:22.23SamotIt usually turns out the network is setup poorly and really can't support voip or they don't actually have enough bandwidth / coverage to support the needs for voip.
17:27.53*** join/#asterisk jameswf (uid27319@gateway/web/irccloud.com/x-ahaahjopujwjvrgn)
17:27.54n8n8Sorry, had a phonecall. I know I don't have enough voice experience, bit I can say that my network has been supporting voice for 8+ years pretty well. I had to learn a lot at the beginning, but I've got law firms, call centers, and a fair number or other businesses (we're 95% business customers) using VOIP without issue over my net. My weakness is in the intricacies of the asterisk config. My biz partner hacked some freepbx's together to meet initial demand, an
17:27.54n8n8d they've been the source of many of my headache's
17:28.52n8n8that's why I stood up my multi-tenant asterisk system. to get out of reverse engineering freepbx...
17:35.31*** join/#asterisk samwierema (~samwierem@195.240.143.134)
17:35.31SamotI use rather highly customized versions of freepbx
17:36.38SamotAnd in a "multi-tenant" setup
17:39.45igcewielingI use Bicomm PBXware, (multi-tennant version)  which is a non-FreePBX GUI for Asterisk for "hosted" customers.
17:41.58SamotI also use a proxy so the users don't directly register to Asterisk/FreePBX
17:42.25SamotWhich makes the peers on the PBX side just something there for states/tracking.
17:43.42SamotFor the SIP trunk or plain, simple "voip" accounts (resi/soho) the PBX is mainly just used to pass the call through so I can have media control.
17:44.12SamotAnd voicemail for resi/soho
17:44.26SamotEverything else for them is dealt with outside of the PBX.
17:45.28igcewieling*nod* that makes a lot of sense if you have lots of registrations.
17:45.56SamotI just turned up a hotel yesterday that added 107
17:46.43SamotBased on current new sales, I should have close to 1,000 new users before the end of the year.
17:47.24igcewielingOut of 286 peers, 6 of them register.   Most of the peers are Adtran SIP/Analog gateways (Netvanta or Total access) so there is often a dozen "lines" per peer.
17:47.48SamotWell it's not just registrations.
17:47.49igcewielingSometimes I wish I had your sales people, Samot
17:48.07SamotYou still have to deal with SIP pings, etc.
17:48.25igcewielingI'm curious how you handle call routing when you don't send calls to the PBX
17:48.36SamotI do send calls to the PBX
17:48.41igcewielingevery single one?
17:48.47SamotYes.
17:49.11SamotThey are either a SIP trunk/resi/soho which means it's going to the PSTN or voicemail
17:49.33igcewielingwhat about inbound calls?
17:49.33SamotOr they are "hosted" and they are calling each other and I need to track states for BLF and other monitoring.
17:49.52SamotPSTN -> Switches -> PBX -> Proxy -> Devices
17:50.10igcewielingSwitches = Ethernet Switches not Telephony Switches?
17:50.17SamotSoftswitches
17:50.36SamotMy PBXes sit between proxy/switches
17:51.52SamotI still make peers for those that needed
17:51.58SamotMainly for the hosted side.
17:52.09SamotSo the calls can be controlled and handled by that peer.
17:52.18SamotFor calls from the proxy to the PBX
17:52.28igcewielingOur hosted stuff is an entirely different box from our main call routing box.
17:52.29SamotAnd send calls to that peer to the proxy...
17:52.45SamotMy main routing box is a pair of softswitches
17:53.01SamotThe routing that is on the PBX is internal for that PBX/users.
17:53.11igcewielingYou are able to match incoming packets to the correct peer on Asterisk since all the packets are coming from the IP of your proxy?
17:53.20SamotYes.
17:53.27SamotIt's called "friend" for Chan_SIP
17:53.32Samotusername,ip
17:54.09igcewielingI have no idea what you mean
17:54.25SamotWhen use setup a chan_SIP peer with the type=friend
17:54.30SamotAnd host=ip
17:54.46igcewielingdon't you mean host=ip_of_your_proxy ?
17:54.49SamotWhen multiple calls from that IP come, Asterisk will match the peers based on username first
17:54.52SamotYes.
17:55.03SamotThen it will match against the IP
17:55.08igcewielinghow many friends do you have with the same IP?
17:55.13SamotAll of them.
17:55.38SamotIt will look at the peer name/user first
17:55.41SamotThen the IP
17:55.51SamotSo I can name the peers whatever I want.
17:56.21SamotBecause the proxy will go "Oh user 100@domain1 is peer 901001 on the PBX, let me fix my from user to that"
17:56.38SamotAnd when it hits the PBX it looks the from is now 901001@domain1
17:58.03SamotAnd when a call from the PBX hits the proxy for 901001@domain1 it goes "Oh that's really 100@domain1 let me send it to that AOR/location"
17:58.24igcewielingis AFK on a call
18:09.38*** join/#asterisk kharwell (kharwell@nat/digium/x-dfvjsofbpisexvag)
18:09.38*** mode/#asterisk [+o kharwell] by ChanServ
18:09.59*** join/#asterisk samwierema (~samwierem@195.240.143.134)
18:19.37igcewielingSamot: Have you tried running kamailio on the same box as Astersik?
18:19.53SamotRecently? No.
18:20.02igcewielingI'll need 4 kamailio hosts, one for each of our main Asterisk servers.
18:20.04SamotBecause I have multiple Kamailio boxes for failover.
18:20.09SamotWhy?
18:20.20igcewielingfor redundancy of course.
18:20.38SamotBut why do you need one for each Asterisk box?
18:20.42igcewielingseems kind of pointless to have 4 call routing boxes when they all go through 1 single proxy.
18:21.49SamotYou shouldn't need one per Asterisk box, how many you want for failover is your call.
18:22.12igcewielingright now we failover between 4 asterisk boxes.
18:22.13SamotA single Kamailio box can handle 1,000's of sessions  without breaking a swear.
18:22.15SamotA single Kamailio box can handle 1,000's of sessions  without breaking a sweat.
18:23.23igcewielingthe HARDWARE can fail.
18:23.31SamotI understand that.
18:23.35SamotBut four boxes?
18:23.52SamotThat's neither here no there.
18:24.12SamotHow many you want to run in the failover is a variable that can change.
18:24.28SamotHowever, you will need a central MySQL server.
18:24.42igcewieling*nod*  I currently run 4 main call routing servers, with 3 of them normally active.
18:24.43Samotso they can all use that for their backend.
18:25.01igcewielingI don't need an mysql server for this
18:25.19SamotYou need a central DB for them to access.
18:25.25SamotWith all the same information.
18:25.25igcewielingWHY?
18:25.32SamotBecause it requires a DB backend
18:25.33igcewielingThe peers are not registering
18:25.42SamotSure but other things are in there.
18:25.45SamotLike your domains
18:25.48SamotThe attributes
18:25.52SamotThe routes you want to use
18:26.00SamotThe subscribers (users/domains)
18:26.04SamotIf you're using that
18:26.11SamotIf this is for call routing you need to be able to route the calls
18:26.18SamotSo that is in a backend
18:26.22igcewielingthe routes are handles in the asterisk dialplan.
18:26.31SamotThen what is the point of the kamailio box?
18:26.37igcewielingexactly!!!!!
18:26.45SamotOK
18:26.46SamotDude.
18:26.53SamotAsterisk as a softswitch is not right
18:26.57SamotYou should be using Kamailio
18:27.05SamotI thought you were asking because you were looking to actually migrate.
18:27.18igcewielingeveryone tells me that, but none of them have compelling reasons except for registration
18:27.22SamotOr some sort of softswitch/sbc
18:27.28SamotCall routing
18:27.34SamotBecause Asterisk is NOT a switch.
18:27.47igcewielingI'm trying to find a reason to spend the time to setup Kamailio since everyone tells me how much it is needed.
18:27.57SamotBecause it makes things easier.
18:28.02SamotTakes load off of Asterisk
18:28.09SamotDoes the job more properly
18:28.14igcewielingI don't need to take load of the Asterisk servers.
18:28.17SamotOK
18:28.23SamotThen why did you ask?
18:28.33SamotBecause it seems no matter what valid reasons, you're going to reject them.
18:29.20igcewielingThose are valid reasons for you.   If I had dozens or hundreds of registrations I might feel differently.
18:29.33SamotI have peers as well
18:29.47SamotStill does a lot of the routing and call management
18:29.48*** join/#asterisk salviadud (~ralfalfa@187-167-69-132.static.axtel.net)
18:29.58SamotHandles the IP auth
18:30.10SamotVerifies the fromuser, etc.
18:30.35SamotDeals with the sip pings for those that are behind NAT
18:30.47SamotDeals with NAT repair
18:31.03SamotProxying the media so no matter what Asterisk box the call comes from, to the end user it's all the same IP.
18:31.04igcewielingEach time I look at Kamailio I realize I'm trying to re-implement all the stuff I've already implemented in Asterisk dialplan
18:31.09SamotRight
18:31.25SamotBut in a more manageable and really, better way.
18:31.31SamotBecause it's like what it was designed for.
18:31.34SamotBut cool.
18:31.46SamotAsterisk as a switch did wonders with A2Billing.
18:31.52SamotThey're still kicking, right?
18:32.30igcewielingno idea, I don't care anything about A2billing, we use OmniBill for billing and use CDRs from the carrier rather than our own CDRs
18:32.40sibiriai think kamailio and similar software are great tools, but certainly not needed for all scenarios. we do hundreds of thousands outbound calls a month originating from our system, and for our use case kamailio or some other SIP proxy would be more of an obstacle
18:32.46SamotWell I use my own CDRs.
18:32.53SamotBecause, I have a lot of carriers.
18:33.09SamotI'm not playing the "lets consolidate these usage reports to get our billing done"
18:33.52Samotsibiria: How is it an obstacle?
18:34.06SamotHow is a SIP Proxy/SBC an ostacle?
18:34.09SamotHow is a SIP Proxy/SBC an obstacle?
18:34.28sibiriaSamot: for us it would be one more piece of software to maintain, that wouldn't be needed for our platform to do its job
18:34.38SamotOK.
18:34.50SamotSo we got 4 Asterisk boxes running for call routing.
18:34.59SamotI have two kamailio boxes, one is failover.
18:35.02SamotDoes the same job.
18:35.08SamotHow is that more hardware?
18:35.21sibiriaif we were to use multiple carriers, or direct routing to this or that operator, i would be looking into kamailio as the first thing on the agenda
18:35.33SamotOh and the kamailio box can handle more call volume and routing vs Asterisk with the same resources.
18:35.41sibiriabut we only do simple app2phone stuff originating from our own platform
18:37.02sibiriaor rather, the stuff we do is sort of complicated, but on SIP/asterisk level it's just running various dial plans on recipient numbers
18:41.36igcewielingMostly my main call routing (non-freepbx) boxes accept a call, runs an AGI, runs dialplan based on the channel variable the AGI set.    Only 500 lines or so of AEL
18:42.20igcewieling2,500ish lines of PHP AGI
18:44.11sibiriafastagi i hope :P
18:45.25igcewielingno, but if the system load increases enough the first thing I'll try it turning it into FastAGI
18:45.46SamotExcept with Kamailio, call comes in, I tag it with all the needed user details, handle the call limits (if channels are limited), check blacklists for the user blocking calls, etc etc then I send the call to the PBX with some extra headers....
18:46.04Samot2500 lines of code.
18:46.04sibiriain this scenario PHP has the problem of being quite hungry on startup
18:46.27SamotBut that also includes all the outbound, failover, firewalling and extra functionality ..
18:46.37SamotNot just handling the routing of inbound calls.
18:47.37igcewieling*nod*  Where do you store your call forwarding info?
18:47.55SamotIn a database?
18:48.10igcewielingthe kamailio database?
18:48.17SamotSure.
18:48.28igcewielinghow does Asterisk access the forwarding info?
18:48.43SamotIt depends on the setup
18:48.48SamotSee you don't have "registered" users
18:48.59SamotSo in the SIP trunk world, why would Asterisk need it?
18:49.36SamotIn those scenarios Asterisk is acting more like a media system than anything else.
18:49.47SamotKamailio handles all that end user stuff.
18:50.29SamotIn a hosted scenario, I let them setup CF via feature codes or use AstDB to set it
18:50.34SamotAnd then do it at the PBX level.
18:50.50igcewieling*nod*  I think the issue is my reluctance to throw out 8 years of work to switch to Kamaililo
18:51.05SamotYou can do more
18:51.07SamotEasier
18:51.10SamotMore manageable
18:51.17Samotwith less.
18:51.26SamotAsterisk is a B2BUA
18:51.30SamotNot a switch/proxy
18:51.35SamotIt is not as scalable.
18:51.40igcewielingI already have a full GUI for managing call routing
18:51.54SamotWhy do you think there are so many Kamailio+Asterisk designs out there?
18:52.15robmaligcewieling: Can you show some pics?
18:52.24robmalBTW. hi :-)
18:52.33SamotAnd it's not about a GUI
18:52.40SamotI'm not talking about management in that way
18:53.01SamotI'm talking about managing the actual routing and the results of the request
18:53.06igcewielingrobmal: do you have a preferred imagebin?
18:53.15SamotInstead of getting back a 503 and then having to Dial() out again
18:53.25SamotThus tearing down and setting up a new channel in Asterisk..
18:53.41SamotKamailio just goes "Oh, do I have another route for this? Cool, let me send it there"
18:53.51SamotAsterisk thinks the call is still just in progress..
18:53.53Samot1 channel
18:54.09SamotThat doesn't have to be stopped and created again with each failure.
18:54.35robmaligcewieling: Any will do.
18:54.57SamotBeing able to do more with the actual transaction and dialog then just INVITEs when it hits the box.
19:14.41*** join/#asterisk LoKoMurdoK (~LoKoMurdo@fedora/LoKoMurdoK)
19:16.08*** join/#asterisk stefanauss (~stefanaus@160.97.54.39)
19:26.05*** join/#asterisk samwierema (~samwierem@195.240.143.134)
19:33.30*** join/#asterisk miralin (~Thunderbi@194.8.128.67)
19:51.33*** join/#asterisk adeel (~adeel@2602:ffc1:1:face:e495:1b1e:9ed2:bc33)
20:11.59*** join/#asterisk NightMonkey_ (~NightMonk@pdpc/supporter/professional/nightmonkey)
20:23.08*** join/#asterisk Penguin (~xwQ5kwYl6@our.systems.are.full.of.penguins.at.penguinsystems.net)
21:24.24*** join/#asterisk [TK]D-Fender (~joe@64.235.216.2)
21:30.50*** join/#asterisk shootbird (~quassel@beepbeep.serverpit.com)
22:08.51*** join/#asterisk tomaluca95 (~quassel@kde/developer/tomaluca)
22:16.06*** part/#asterisk kharwell (kharwell@nat/digium/x-dfvjsofbpisexvag)
22:17.30*** join/#asterisk kunwon1 (~kunwon1@unaffiliated/kunwon1)
22:50.39*** join/#asterisk gtjoseph (~gtjoseph@unaffiliated/gtj)
22:50.40*** mode/#asterisk [+o gtjoseph] by ChanServ
23:27.11*** join/#asterisk gtjoseph (~gtjoseph@unaffiliated/gtj)
23:27.11*** mode/#asterisk [+o gtjoseph] by ChanServ
23:31.38*** join/#asterisk Juggie (~Juggie@unaffiliated/juggie)
23:32.11*** join/#asterisk gtjoseph (~gtjoseph@unaffiliated/gtj)
23:32.11*** mode/#asterisk [+o gtjoseph] by ChanServ
23:57.23*** join/#asterisk cemotyz09 (~cemotyz09@cpe-70-121-128-59.satx.res.rr.com)

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.