Server Spec for cluster

[pratikvicidial]

Hello Everyone

we are looking to setup the vici cluster for 300 reps and up to 3500 concurrent calls.
What would vici cluster architecture like how many servers are required etc.

Any help would be highly thank full.

TIA.

[carpenox]

Here would be our MINIMUM recommendations for a system with those requirements:
- 14 x dialers(single quad-core CPUs with 8GB RAM)
- 2 x webservers(single quad-core CPUs with 8GB RAM)
- 1 x DB server(quad-quad-core CPUs with 32 GB RAM and LSILogic MegaRAID card with 4 x fast SSD drives set up in two RAID-1 partitions, one for DB data and one for everything else)
- also recommended are another DB server to operate as a slave/reporting/backup DB, as well as an archive server to store recordings

We have set up many VICIdial systems of this size, and not using a configuration like the above WILL lead to problems with your system.

Contact us for help if you need

Chris

[williamconley]

Hello Everyone

we are looking to setup the vici cluster for 300 reps and up to 3500 concurrent calls.
What would vici cluster architecture like how many servers are required etc.

Any help would be highly thank full.

TIA.

Your database server (at least) should have nvme drives if you can get them. But I digress: If you are setting up a system of this size and you do not have the experience of managing a vicidial cluster of more than 150 agents with 10:1 autodialing or above before you begin, you will likely crash and burn if you don't have assistance. Seriously. Either contact The Vicidial Group or another professional Vicidial support organization or hire someone who has Vicidial server experience professionally (with hands-on the servers themselves). You will have concerns regarding bandwidth, connectivity, enterprise configuration requirements and ramp up for special uses (such as reports and manager accesses).

I would honestly recommend you split this into two or even three systems if your agents don't have to always share leads on the same campaign. Having less stress on three separate clusters can allow your IT staff and managers to sleep more nights without pulling out their hair. Consider that a single point of failure (one database server) could render all 300 agents useless for several hours in case of a system failure. Even if it's just a network card, it takes time to find, fix and get everyone back on the phones.

[carpenox]

hey welcome back bud, you been missing in action forever, glad to see ya back

[williamconley]

hey welcome back bud, you been missing in action forever, glad to see ya back

I'm not sure I'm supposed to talk to you: Alma Linux? What on earth made you decide on that distro? lol

But I digress: I got permission now that a bunch of "move-related" and major projects are almost completed.

[carpenox]

alma linux is the new redhat centos distro and performs over 300% better than opensuse

[williamconley]

Can you get 600 live calls on a single dialer?

For clarity: When I built my first installer, the goal was to give me as much control over the install as possible (since there was not, at that point a single viable ISO installer). We built it in Gentoo and once completed scraped out everything that was not completely necessary. Result? 600 calls on 4 core machines. Back in the robodial days that was very important. LOL