Here’s a weirdness – I got a call from someone who couldn’t get to my info line earlier, I tried it and it was busy tone.

Being on a layby beside a road on a mobile on a long journey, my only real option was a remote server reboot so I couldn’t diagnose further.

That fixed it, but here’s the weirdness:

[Aug 23 14:16:59] NOTICE[5923] res_pjsip/pjsip_distributor.c: Request
‘INVITE’ from ‘”thecallerID” ‘ failed for
‘46.31.X.X:5060’ (callid: VF0046af0fc3f12e7dddb70d62974exx@voipfone) – No matching endpoint found
[Aug 23 14:17:25] NOTICE[5923] res_pjsip/pjsip_distributor.c: Request
‘INVITE’ from ‘”thecallerID” ‘ failed for
‘46.31.X.X:5060’ (callid: VFdfc3ae4343390139cb285533bc97xx@voipfone) – No matching endpoint found
[Aug 23 14:18:01] NOTICE[5923] res_pjsip/pjsip_distributor.c: Request
‘INVITE’ from ‘”thecallerID” ‘ failed for
‘46.31.X.X:5060’ (callid: VF9f956b236cef2b7609037335db7xx@voipfone) – No matching endpoint found
[Aug 23 14:28:54] NOTICE[4014] res_pjsip/pjsip_distributor.c: Request
‘INVITE’ from ‘”thecallerID” ‘ failed for
‘46.31.X.X:5060’ (callid: VFb797ae4a6e90a05a480de806afxxe@voipfone) – No matching endpoint found

All I can tell you is that this server has been up and running for 3 months without problem (occasional reboot).

The host’s monitors showed no strange spikes, it didn’t run out of memory, and I can find nothing in syslog related to Asterisk between a known “good”
time and the time of the failed calls.

And of course, the Asterisk process didn’t crash, otherwise it wouldn’t be writing to logs.

The only thing that might be relevant is that yesterday, the ITSP
increased its IP address ranges, but I already had those ranges listed in iptables and pjsip.conf.

And besides, if the traffic hadn’t been blocked, then, again, it wouldn’t have made it through to Asterisk to get logged. Am I correct?

So, it’s like it somehow got amnesia. The only other strange thing I am doing is running this script at just after midnight each night, although it’s been running just fine for a couple of weeks now.

(It’s a workaround for a moh fiddle I’m running – see a previous post but I
can’t see it as being relevant).

pkill mpg123 >> /root/tidy.log
/usr/sbin/asterisk -x “moh reload” >> /root/tidy.log
/usr/sbin/asterisk -x “database query \”DELETE FROM \`astdb\` WHERE \`key\`
LIKE \’%Last%\’;\”” >> /root/tidy.log

So… it just seems like Asterisk got pjsip endpoint amnesia – is that possible? Anyone come across this before?

Thanks.