Hi, I’m new to this project, and i’m from China.
I joined a server to the CN zone last week. The server score never stay above 10 for a long time. As soon as it reach 10 point, a timeout happens, and the score took a dive. It happens about each tow hours. I readed former post about scores, and I don’t know if the problem is on the monitoring station. Could anybody help me out of this ?
My server status is here: https://www.ntppool.org/scores/106.39.20.237
I got some packages, about 260p/s, but no more.
I changed the server registration from DNS name to IP address. The scores are still rolling, but I get more packages, about 3kp/s. I’d wait to see if it can reach 10k.
And I’m strill curious about the rolling score, if somebody knows, please tell me.
According to your MRTG graph, it is possible that your server just overloaded and started dropping query packets in the spiky periods. Set your server bandwidth to 384k and see if it can keep up with incoming queries.
I tried to change my server bandwidth to 384k as you sugested. Unfortunately the scroe keep rolling.
Also, I jioned the beta monitor system, and it works pretty good. Screen shots as blow.
The CN region just having too much traffic…
I have few servers in CN as well, and soon after they got score over 10, I had to apply traffic limit on NTP requests to avoid all my bandwidth been used out, and then I’m having requests / score graph very similar with yours, that’s a normal appearance when the server / network can’t handle all those requests.
Thank you two for help. Before today my server’s CPU utilization never above 30%, and the network traffic always below 1MB/s，so I didn’t think it’s a performance problme. But today, the CPU utilizaion came to near 100% for about an hour, and scores on both monitor dorp hardly. I’ll try the ECMP seen by another post. It may take a while to figure out how to accomplish in an all VM envrionment. And I’ll post whatever come out.