Hi Ken, Looking at the timestamps on the Sprout logs, it looks like your two Sprout nodes have significantly different times. Can you check whether this is true, and synchronise them if so?
A large time difference (greater than 1sec) will affect the Chronos service. As background, when a subscriber registers, a Chronos timer is set to pop when the registration expires. When the timer pops, the registration is deleted. What (might be) happening is that your subscriber registers on the Sprout that has an earlier time. However, the timer gets sent to Chronos on the other Sprout that has a newer time, and gets immediately popped as the Chronos thinks that time has passed. This then causes the registration to be immediately expired, so the later INVITE gets rejected with a 480. Ellie From: Antonio Ken Iannillo [mailto:[email protected]] Sent: 03 December 2014 09:12 To: [email protected] Cc: Eleanor Merry Subject: Re: [Clearwater] 480 Temporarily Unavailable Hi Ellie, I agree this is a registration problem. I already run the nodetool facility on hs and, as you said, I get both nodes with UN state, identical load and 100% of "owns". Same for sprout cluster settings: they're identical on both nodes. Here there are all the logs for a single call: https://app.box.com/s/gxinfcfd2ojm0hmv0f1a Hope this can help, KEN 2014-12-02 19:13 GMT+01:00 Eleanor Merry <[email protected]<mailto:[email protected]>>: Hi Ken, This sounds like the terminating user isn't registered, so the call is rejected with a 480. Can you check that the Homestead nodes have been clustered correctly? If you run 'nodetool status' on each homestead node I'd expect you to see both Homestead nodes owning 100%, roughly equal load and in the UN state. Can you also check the Sprout clustering? Each Sprout node should have the file /etc/clearwater/cluster_settings, and it should have a single line of the form "servers=<Sprout-1 IP address:11211>,<Sprout-2 IP address:11211>,...". The order of the Sprouts should be identical between the Sprout nodes. If these are OK, then can you please send the logs (for a single call attempt) as well? Thanks, Ellie -----Original Message----- From: [email protected]<mailto:[email protected]> [mailto:[email protected]<mailto:[email protected]>] On Behalf Of Antonio Ken Iannillo Sent: 02 December 2014 11:30 To: [email protected]<mailto:[email protected]> Subject: [Clearwater] 480 Temporarily Unavailable Dear All, Dear Eleonor, I resolve the previous problem, thought I think it just needed time for the dns name propagation. However I came up with another issue that is, after the first INVITE and the 100 response, the reception of an unexpected message. 480 Temporarily Unavailable This is totally random and it is not associated to a particular user. In the deployment there are two copy of each node (but Ellis) on two different servers. In the log of the sprout nodes I found that - sprout node gets the state registered in hs for user 2010000001 - memcached fetches result for key reg\\sip:2010000001@..., CAS=9828 - data store returns the record - then scscf (from sproutletproxy) send the 408 response Thanks for your help, Ken _______________________________________________ Clearwater mailing list [email protected]<mailto:[email protected]> http://lists.projectclearwater.org/listinfo/clearwater _______________________________________________ Clearwater mailing list [email protected] http://lists.projectclearwater.org/listinfo/clearwater
