Hi, glad to be around again, it's been a while : )
It's a fresh, stock ViciBox 6.0.1 cluster install.
1 Web Server
1 DB Server
2 Telephone servers - 192.168.200.21 (T1) and 192.168.200.22 (T2).
SVN 2163, vicidial.php 2.10-412c 140811-1219, admin.php 2.10-448a 140706-0927.
Crated 2 new SIP phones- 11101 and 21101.
Created a new phone alias, named 102, with a list of phones 11101,21101.
Noone else is loged in.
When for both telephone servers Active Agent Server = Y, and for agent login the alias 102 is used, T2 accepts the call. This seems normal looking at vicidial.php as results from the SQL query are traversed in reverse order.
However, if T2 is set to Active Agent Server = N, while T1 is set to Y, a login error appears, rather failing over to T1.
If T2 is set to Active Agent Server = Y, while T1 is set to N, T2 accepts the call and no error appears.
My idea is to use phone alias, and if anything goes wrong with any of the telephone servers, to set it to Active Agent Server = N and Active Asterisk Server = N and pull it off the cluster while being rebuilt or whatever, at the same time expecting phone alias logins to ignore/skip the removed/disabled server.
My question is is the above described an expected behaviour, or I am missing something.
If it's expected behaviour, what would be the best practise to take temporarily a telephone server off the cluster, without breaking or editing phone alias logins?
Or maybe it's a candidate for a future request?