Daemon Memory leak - Preemptive maintenance

Our staff noticed some daemons requesting almost 5G of RAM of resident memory and almost 30GB of virtual memory. This node behaviour is very dangerours since it may lead to node shutdown if the system starve in resources.

Due to our security buffer not a single node was shutdown, but we had to impose a regular daemon restarting routine. Many users may notice 6-12 minutes of disconnection in their nodes.

We must inform that Horizen has done a great job improving the use of memory and daemon portability, but it seems that we can wait for more on that too!


Leave a comment