logs archiveBotHelp.net / Freenode / #2600hz / 2015 / August / 14 / 1
pramsky
hey everyone, I have a few numbers assigned to a PBX in the PBX connector. I moved the number from one PBX to another and now calling the number returns busy. Freeswitch says : UNALLOCATED_NUMBER . Any ideas ?
stormqloud_
make sure caches are flushed
hmmhesays
well after 4 hours of poking around monster I might be able to use this
I'm hoping app development isn't too daunting
pramsky
stormqloud_ , I flush wh_cache , whapps_config , ecallmgr_config , whistle_couch_maintenance , couch_mgr , the works
stepswitch_maintenance flush as well
stormqloud_
how did you transfer the number?
HTTP API? Kazoo-ui?
you should look in the mailing list, there is something vaguely famailiar about your question
pramsky
kazoo-ui
under the pbxs connector clicking Move to another PBX and selecting the PBX from the dropdown
if I restart kz-ecallmgr and kz-whistle_apps
stormqloud_
ah, I havn't tried that exactly
pramsky
it works
stormqloud_
ah, so it is a cache thing then
pramsky
it is
stormqloud_
reconcile
?
try one of the sup number reconcile ones
pramsky
ok
do you have a url to the docs on that ?
stormqloud_
sec
sup whistle_number_manager_maintenance reconcile
somethign liek that
pramsky
k
haha searching google for that returns this great link: https://github.com/stormqloud/kazoo-tgbbok/blob/master/sup.md
thanks!
stormqloud_
sup trunkstore_maintenance flush
that might be what you need
pramsky
did that too, still shows as unallocated number
in freeswitch logs
stormqloud_
did you put kazoo in debug for syslog
then look in /var/log/2600hz/kazoo.log
depending on how many servers you have..
the problem is that the transaction end up spread among different servers logs
FYI on that
pramsky
yea, i have 2 media servers, had the freeswitch console open on both and kazoo log being tailed.
says: "Disposition":"UNALLOCATED_NUMBER"
stormqloud_
why don;t you try just to sto and start one of them
kz-ecallmgr and kz-whistle_apps
then at least you have it down to 50%
pramsky
only one app server , i should have setup a second app server. restarting those services drops registrations. I'll do that after 5pm then
stormqloud_
should have asked what version you are on
pramsky
3.19
stormqloud_
ah
pramsky
one app, two freeswitch, two couch
stormqloud_
registrations are kept in kamailio I think
I don;t want to mess up your callers tho
pramsky
yea, I'll restart the app at 5:30 and keep my fingers crossed.
stormqloud_
do what works for you1
pramsky
will do. its definitely a cache issue
stormqloud_
I havn;t really used the trunking module stuff in kazoo.
you are west?
hmmhesays
registrations are kept in ecallmanager aren't they?
stormqloud_
I think they are kept in kamailio since that's what handles it. I think ecallmgr has access to the data tho as well. It's a bit beyond my pay grade right now!
when phones register to kamailio kamilio keeps track of it and also pushes it into amqp
then the other kamailio and ecallgr learn about it
so it's not really just kept in 1 place
kamailio also has the abilit to store it's registration info on disk
so when it starts up again it knows the info faster
the features really depend on what version of kazoo
and whether you have th eknobs turned on or off
http://www.vintagecomputer.net/MITS/8800a/MITS_Altair8800a_toggle-switch-nuts.jpg
voxter
any of you dudes kamailio experts?
stormqloud_
only MR L I think
maybe I can help, I've been using it in non kazoo for many years
but if it's realted to your MWI woes, I'm not the guy for sure
lazedo
voxter: wazzup ?
voxter
lazedo: I was just a little impatient and needed to implement somethign that would handle a pickup 302 without redirecting back to the same To/Contact
But, I didnt realize I was doing it in the wrong place - it seems to work now
the best fix might be to actually do a kazoo query and just direct-dispatch to the correct freeswitch box
but temporarily i just implemented a little thing that modifies 302 redirects for park pickups to go to sip:pickup+*31@etc and then strip that back off and dispatch to X-Redirect-Server from there
lazedo
voxter: because some UA dont like to 302 to same server right ?ยด
voxter
lazedo: ya, you got it - apparently when To: and Contact: are the same as the original invite, some UA's think its a loop and wont re invite
lazedo: they just end the call
luckily - our largest deployed brand does this.... LOL :(
lazedo
voxter: your fix should work
voxter
hey also
I dont know how much is coming for kazoocon
but we found a bunch of tiny little tweaks when we were doing registration/call load testing the other day
lazedo
voxter: were working on 302 redirect within kamailio without going back to UA
voxter
where certain conditions would allow for credential caching to break, or would pass on requests straight to whapps and overwhelm things
we went from 150reg/sec before latency increased
up to 2300reg/sec
which i am pretty sure was maxing out the one sipp instance we had running
lazedo
voxter: credential cache to break, that im curiuos
voxter
lazedo: it was just the way it was querying the location db, if a device had not gone through nathelper, and had no received= entry in the db, the cache lookup would fail
lazedo: I'll do some pull requests for our changes, and if they can also be adopted into any new work, thats great
lazedo: im so excited to kazoocon this year - you're going to be there again right??
lazedo
voxter: got my flight booked
voxter
niceeee
i might venture out and test out using google cloud print for faxboxes tonight once i get this 302 patch tested
lazedo
voxter: not sure i follow the nathelper thing and its relation to credential cache
voxter
one sec ill paste the line
if ($sht(auth_cache=>$Au) != $null && registered("location", "$rz:$Au", 6) == 1) {
line 80, registrar-role
that last argument to registered() is 6 which means the contact and the received headers have to match
lazedo
voxter: youre right
voxter
received will be NULL if you are not passed through nathelper
lazedo
voxter: we are actually changing it to 2 for vNext
« prev 1 2 next »