I am trying to set up a server for rehearsals of our choral. The choral consists of around 20 people all living in the same area in the near of Cologne, Germany.
I have set up a server as a virtual private server hosted by Contabo. I asked them to set up my machine in Düsseldorf, Germany, as in my tests the Düsseldorf location delivered the best ping times of around 15 ms .
I installed a Ubunutu 18.04 LTS, and for the Jamulus package have been following the installation guidelines for a headless server as provided in the wiki: https://jamulus.io/wiki/Installation-for-Linux
Then I installed the low latency kernel as recommended: "linux-lowlatency-hwe-18.04"
For starting the jamulus process I am using the service script provided here: https://jamulus.io/wiki/Tips-Tricks-More
From my win10 client I got the above mentioned ping time of around 15ms, overall delay ist around 55ms.
The first attempts of a rehearsal with 4-5 persons seemed promising, so we tried the next step with 10 to 12 people.
With more clients joining , we get continuous cracks and noises which make it impossible to use, ping time changes dramatically , up to several hundred ms, the same with overall delay time.
I look at the CPU load on the server, it shows between 40% and 80% for the Jamulus process when the first client connects, and then I see CPU load increasing far beyond the 100% load!
These are the basic data for the VPS:
6 vCPU cores
16 GB RAM
400 GB SSD
400 Mbit/s Port
Is ist possible that I fooled something up during installation or is the VPS too small?
Kind regards
Michael
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Did you add the -T or --multithreading at the server run line in the service file te enable multiple vCPU's. In the install documentation on the wiki it is not listed, but because you have 6 cores you have to tell jamulus to use them.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
My provider says their CPU is based on AMD epyc 7282, and the jamulus task ought to have no problems. But to my mind it could also be that the host my VPS was running on had a problem, as last night they migrated my VPS to a different host, and now I see - 4 clients connected - a CPU load of around 20%. This looks much better than before, and the ultimate test on Monday with 10+ active clients will hopefully bring about more reliable results.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello forum,
I am trying to set up a server for rehearsals of our choral. The choral consists of around 20 people all living in the same area in the near of Cologne, Germany.
I have set up a server as a virtual private server hosted by Contabo. I asked them to set up my machine in Düsseldorf, Germany, as in my tests the Düsseldorf location delivered the best ping times of around 15 ms .
I installed a Ubunutu 18.04 LTS, and for the Jamulus package have been following the installation guidelines for a headless server as provided in the wiki:
https://jamulus.io/wiki/Installation-for-Linux
Then I installed the low latency kernel as recommended: "linux-lowlatency-hwe-18.04"
For starting the jamulus process I am using the service script provided here:
https://jamulus.io/wiki/Tips-Tricks-More
From my win10 client I got the above mentioned ping time of around 15ms, overall delay ist around 55ms.
The first attempts of a rehearsal with 4-5 persons seemed promising, so we tried the next step with 10 to 12 people.
With more clients joining , we get continuous cracks and noises which make it impossible to use, ping time changes dramatically , up to several hundred ms, the same with overall delay time.
I look at the CPU load on the server, it shows between 40% and 80% for the Jamulus process when the first client connects, and then I see CPU load increasing far beyond the 100% load!
These are the basic data for the VPS:
6 vCPU cores
16 GB RAM
400 GB SSD
400 Mbit/s Port
Is ist possible that I fooled something up during installation or is the VPS too small?
Kind regards
Michael
Do you know the server specifications? Processor and clock...
Did you add the -T or --multithreading at the server run line in the service file te enable multiple vCPU's. In the install documentation on the wiki it is not listed, but because you have 6 cores you have to tell jamulus to use them.
Yes, I tried with -T and without. Made no difference, though. And I emailed my provider as to the specifications, and now I'm waiting for their reply.
My provider says their CPU is based on AMD epyc 7282, and the jamulus task ought to have no problems. But to my mind it could also be that the host my VPS was running on had a problem, as last night they migrated my VPS to a different host, and now I see - 4 clients connected - a CPU load of around 20%. This looks much better than before, and the ultimate test on Monday with 10+ active clients will hopefully bring about more reliable results.