Lemii
commented
6 months ago After creating this issue, I removed all pm2 processes using the pm2 delete command. Afterwards, I restarted them one by one. It's now a few hours later, and no explosion has occurred thus far.
Perhaps the problem was that a process with an old version was saved in the pm2 startup config, and that by deleting the processes (rather than restarting them) it forced pm2 to use the new version?
Either way, I see v0.7.2 has been tagged. I'll try that branch, re-sync and re-index the node from scratch, and will report back if I run into more problems.
shuse2
It is currently not possible to run a stable mainnet Lisk Core + Lisk Service server.
Expected behavior
Lisk Core and Lisk Service should run stable without problems.
Actual behavior
When running Core and LS together, at some point the whole server becomes slow and unresponsive (even though the CPU and memory usage do not appear high..... yet). From that point on, the memory usage of Core keeps increasing with hundreds of megabytes every few seconds. Even if you stop all processes except Core, the memory still keeps rising. It continues to do this until the server runs out of memory and all PM2 processes are forcefully stopped. There are no additional error logs.
After the explosion, PM2 restarts with all processes saved in the
pm2 saveconfiguration.This issue does not seem to occur when running Core without LS.
The issue also occurs while indexing.
Steps to reproduce
Which version(s) does this affect? (Environment, OS, etc...)