All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
What response time can be expected for OpenVZ budget VPS?
I have two servers with SparkVPS, one in Dallas and one in New York. I have a computer in my own home network that sends ping requests to both servers every 80 seconds. The response time from the one in Dallas is very good and averages at 45 milliseconds with very little deviation. The response time from the other server varies a lot and is never better than 85 milliseconds. New York is further away from my location so I expect the response to take longer.
The problem I have is that the two servers work in a cluster and synchronize data. A response not being received within 8 seconds makes the software assume the other server is dead.
So my question is, how long can one expect an OpenVZ budget VPS to be busy before my own process gets execution time and responds?
I have collected data for five days and I have seen responses that take more than 16 seconds. I find it a bit weird that the Linux kernel is not able to schedule processes within this time, even if I have a neighbor that consumes a lot of CPU.
Comments
have both vps in same location but different nodes.
Thanks Ehab, but I should probably have explained that I have the two servers for redundancy purposes and the idea is that one server should still work even if the complete network in the other location goes down. The software can recover from this, so my question is more theoretical, what can be expected from OpenVZ. I still want perfect uptime :-)
You havn't thought to move from a Openvz to a KVM ?
I'm sure that you are fully aware but I will just mention it anyway. ICMP (ping) is not a priority protocol and should not really be used solely for monitoring.
That being said it would depend on if both VPS servers have a dedicated amount of resources or are on a shared use base. You may find that the one in New York is just a fuller loaded server and or network or that the interconnection between the two providers are highly loaded.
There could be a number of different reasons but these are just some that I can think of right now.
No, 16 seconds is more indicative of packet loss or bad server. That's not normal or expected.
You can use uptime robot to monitor the servers and look for network issues, but you should also monitor resources on the VPS to see if CPU cycles are being starved, ran out of memory, or otherwise crashed.