Vicidial and virtualization 2020

[tails84]

Greetings,

What is the feasability of running vicidial whole or in part on virtual machines? We currently run about 70 agents across 1 primary server (web,sql,dialer) and 4 clusters (just dialer) works reasonably well, but I am seeing some random Lagged paused codes appear. Today I hope to find what is causing the lagged status, but if the issue is the platform, then i might need to make other plans. I notice a bunch of posts from 2010 about virtualization not being ideal, but that was a decade ago.

VERSION: 2.14-721a
BUILD: 191015-1620

Cheers

[williamconley]

"Not being ideal".

Um ... not being viable.

If you have issues with server load (too much work, too little server) switching to a virtualized kernal is the opposite of a good idea. At 70 agents, you are approaching the barrier where everything must be pretty darn close to perfect. Networking between servers should be gigabit and no firewall. Agent web/phone connections must be clean. Servers should be routinely cleaned of excess audio files and logs. The DB server (especially) must be rid of excess table data and optimized regularly. Note that having millions of leads can beging to bog down your system, and logs can do the same thing.

Also consider that enterprise level settings exist for both dialer and web to allow more users than usual. Check for apache errors related to running out of network ports, etc. Trace down one problem at a time before you begin a rebuild. Consider making sure ALL services have redundancy (including a replication server to run reports off the main db server, which can be problemetic when running report while dialing).

[tails84]

Thanks for the guidance mate. I will update once we have a poke around.