I made some changes to Rspamd and now the server is running normal.
My VPS had 6GB memory and it was eaten to the last bit, making it impossible for the server to run normal.
Solution, goto /etc/rspamd and copy the from modules.d the file neural.conf into override.d
Edit the file so this section looks like this:
Code: Select all
neural {
#servers = 127.0.0.1:6379; # Redis server to store learning data and ANN
enabled = false;
train {
max_trains = 1k; # Number of trains per epoch
max_usages = 20; # Number of learn iterations while ANN data is valid
learning_rate = 0.01; # Rate of learning (Torch only)
max_iterations = 25; # Maximum iterations of learning (Torch only)
}
timeout = 20; # Increase redis timeout
And last, logrotate is also a big problem, the files grow too big and then it starts zipping them on rotate.
I suggest you rotate them more often, so they do not grow that much.
That will help the CPU to stay fast when it's rotating.
Hope it helps others.
Bas.