Hi All,
I have the following:
FusionPBX 5.0.3.
Ubuntu 20.04 (Not my choice./ Stuck with it)
Freeswitch 1.10.7
Cache is file mode under /var/cache/fusionpbx.
I have moved the coredb and the Internal registrations in to Posgresql.
There is not a lot of call volume and dial plan is very basic but we have a large amount of registrations. ~5000 over 4 domains.
The extensions were setup with SIP Force Contact = 'Rewrite Contact IP and Port'. As this was causing a 30 second expiry back to the clients, I edited 'Sip Force Expires' per extension to 65. I tested on a few extensions manually and saved. This would cause the directory.extension file in the cache to be removed and then replaced when the client re-registered. The file had the new expiry timeout. All good.
I then ran a mass update on the v_extensions table in Postgres to mass update all the extensions in the domains I wanted changed. That went ahead fine. I did realise this would not update the cache files.
After business hours, I cleared the cache from the GUI, and all the files in the directory were cleared.
However, then, 'fs_cli -x "show registrations" | grep total' showed registrations down to 6 ( from ~5000), and did not recover. There were no new files being created in the cache folder. I tried reloadacl and reloadxml. No joy. I tried connecting with Bria on a working account and could not connect.
I had to restart Freeswitch to get connections working again.
Logging is not up high enough unfortunately so I have nothing to look at in the logs.
Question: Did I do steps in the wrong order? When updating the Postgresql DB for mass changes, is there a preferred way to "enable" the change in cache? Or did I do things correctly and I need to troubleshoot this further with better logging?
Cheers.
I have the following:
FusionPBX 5.0.3.
Ubuntu 20.04 (Not my choice./ Stuck with it)
Freeswitch 1.10.7
Cache is file mode under /var/cache/fusionpbx.
I have moved the coredb and the Internal registrations in to Posgresql.
There is not a lot of call volume and dial plan is very basic but we have a large amount of registrations. ~5000 over 4 domains.
The extensions were setup with SIP Force Contact = 'Rewrite Contact IP and Port'. As this was causing a 30 second expiry back to the clients, I edited 'Sip Force Expires' per extension to 65. I tested on a few extensions manually and saved. This would cause the directory.extension file in the cache to be removed and then replaced when the client re-registered. The file had the new expiry timeout. All good.
I then ran a mass update on the v_extensions table in Postgres to mass update all the extensions in the domains I wanted changed. That went ahead fine. I did realise this would not update the cache files.
After business hours, I cleared the cache from the GUI, and all the files in the directory were cleared.
However, then, 'fs_cli -x "show registrations" | grep total' showed registrations down to 6 ( from ~5000), and did not recover. There were no new files being created in the cache folder. I tried reloadacl and reloadxml. No joy. I tried connecting with Bria on a working account and could not connect.
I had to restart Freeswitch to get connections working again.
Logging is not up high enough unfortunately so I have nothing to look at in the logs.
Question: Did I do steps in the wrong order? When updating the Postgresql DB for mass changes, is there a preferred way to "enable" the change in cache? Or did I do things correctly and I need to troubleshoot this further with better logging?
Cheers.
