I start NIS with: -Xms512M -Xmx950M
I really don’t see it swapping much
@rigel: can you check if other nis logs than the one supplied contain the ip 52.16.44.214?
I see in the servant log that if the servant has to do 3 bandwidth tests in a row, then the calculation that is involved takes longer for the first one. That is due to the java class being paged out to hdd/ssd i think.
i can see no requests from 52.16.44.214
@rigel: i will let the master run on my local node too and see if it shows the same problems with your node.
@rigel7, can you supply the current nis log again? My local node rewards master is getting
2016-02-16 16:02:18.393 WARNUNG getChainAfter completed exceptionally, reason: org.nem.core.connect.BusyPeerException: java.net.SocketTimeoutException
when requesting a chain part from your node.
Edit: All in all your node does do better when the master runs on my local computer.
Looks like a network issue. I’m going to talk about this to the hoster.
Here is the log file: http://www.megafileupload.com/akfX/nis-0.log
I am setting up a more powerful server for the node rewards master 
@rigel: with the upcoming new node rewards server rigel7 is doing a bit better but still looks too weak. i personally have -xmx=768M set for nis on my 1GB servers, but they also have a hard time to pass all tests. Maybe you can play around with swapiness a bit.
The new node rewards server is online. The results should now be more constant.
Everything seems to be running well. Where are we now regarding Start date. tks
Next step is to evaluate the tests to flag which nodes are eligible for payout and which nodes alter their status (activated/deactivated). No ETA for start yet.
Hello, my node is orbedragones:
At first all the test were fine, but suddenly the “responsiveness” is too high, and I don´t find the problem… Which factors affects the “responsiveness”? How can I optimize it?
Thanks
Maybe a restart of nis helps. How much ram does your node have?
Same problem happened to me.
I doubled RAM to 2GB but responsiveness is still very high.
I’m afraid it is a networking issue but I had no time to investigate.
Perhaps my node is doing reverse DNS query of IP address of the node doing the test?
What request is used by responsiveness test?
I have a dedicated PC with 8Gb of RAM. I tried with Windows 10 and now I have installed Ubuntu server, but with the same result in responsiveness.
2 GB is definitely enough, there are a lot of vpses that are doing well with that amount of ram.
The responsiveness test is done by issuing 10 concurrent requests of the current block chain height.
How much RAM do you give to the java process? The responsiveness test shows that only a few of the requests are answered by your nis.
The request string is /chain/height?
This is not audited by default. I’ll start logging it
Any idea when the rewards program will start paying. I feel we may lose some super nodes if it doesnt start paying now the price is up there.
ryan_Todd, sorry we are not ready yet. I think i fixed all bugs now but we will need some time to code the automated payout.