18.04 Sipxsupervisor and BLA Issues

[pmkr...@gmail.com]

We have a production 18.04 server that was upgraded from releases 14 and 16 and - the upgrade method was 18.04 new install followed by a restore of the system configuration file. The system was running fine for a several weeks. However this week the users reported that the BLA lines show red busy even though there are no calls. I registered two test phones remotely to the system with the two shared lines on each phone with multiple line key appearances. One of the line keys on each shared line shows solid red even though there are no calls on the system. A restart of the sipxsaa and sipxrls processes did not clear up the issue, nor did a reboot of the phones. I dropped the subscription database in Mongo; this did not correct the issue. This morning, I rebooted the voice server - the issue is still there. When BLA is turned off on the two shared lines and phone profiles sent to the test phones, the busy lights disappear. However when BLA is enabled again, the red lamps show up again within 60 seconds or less.

We did an audit of the system log files and noticed the following issues:

• In the /var/log/messages file, this message appears frequently -  'cf3[3534]:  Proposed executable file "/etc/init.d/mysqld" doesn't exist'. The configuration file from 15.08 had Sipcapture enabled which has since gone away. In the Servers->Services menu, the Sipcapture database is listed as a service but stopped. I assume this message is benign as mysqld is no longer used on the system.
  
• Sipxsupervisor does not start, even when trying to start manually (service sipxsupervisor restart | start). The sipxagent.log file shows no errors. Issuing a 'cf-promises -f /usr/share/sipxecs/cfinputs/promises.cf' returns no errors.
  
• We see this alarm from sipxregistrar "2018-12-07T14:28:22.915751Z":8:SIP:CRIT:uc.voice.xyz.com:SipRegistrar:7f08b0468700:sipxregistry:"[160-ENUM] SipRedirectorENUM::readConfig BASE_DOMAIN parameter missing or empty". The system is running with default registrar settings.
  
• We see this alarm from from sipxproxy "2018-12-07T12:06:53.935788Z":937:SIP:CRIT:uc.voice.xyz.com::7fbc76d1d700:sipxproxy:"ALARM_DNS_LOOKUP_FAILED DNS lookup failed for 'mwi.uc.voice.xyz.com'. No valid 'SRV' records found". The system is running the default DNS configuration generated  by Sipxcom and mwi is working properly.

Incoming and outgoing calls are being processed correctly, voicemail and AA are working.

Other than rebuilding the system from scratch (or restore from backup), is there a system configuration file or DB setting that will clear out the busy lamp issue and get sipxsupervisor running?

Many thanks in advance.

Peter

[pmkr...@gmail.com]

Brief update - I added a new test shared user extension to the problem system, defined this new extension on two test phones, and tested BLA - no issue. I then deleted one of the existing shared lines, pushed server profiles, restarted rls/saa, and then re-created the same shared user extension - the spurious red lights reappeared. For the moment, I have created new shared extensions and applied to the phones to work around this issue.

Is there a realtime file on Sipxcom that can be reset/edited/deleted to clear out this issue?

All the best

Peter