nebula-graphd generates a large number of processes, and excessive use of the number of processes causes further system crashes

[IanLiYi1996]

Dear developers,
In actual use, we found that the database will automatically start a large number of processes, causing the machine to crash, but we don't know what the cause is. I hope you can help us explore and look forward to solving this problem together.
Good luck and look forward to your early reply

pstree.txt

[CPWstatic]

Would you please describe how can we reproduce the scenario? We'd appreciate if you could provide us information about your environment, deployment, queries.

[IanLiYi1996]

The machine information used by the deployment map database is as follows
environment：Ubuntu 16.04.7 LTS (GNU/Linux 4.4.0-131-generic x86_64)
deployment: docker-compose as the document recommended
queries: None

When we encountered the above problem, we installed nebula studio to manage the database, and the machine crashed within a day. The cause of the crash was consistent with what I described. It was caused by the abnormal start of too many child processes by the nebula service itself.

By the way, before the machine crashed, we did not perform any data operations

[dutor]

What do you mean by machine crash, some of the service processes crashed or the OS panic? If it is the panic case, could you provide related logging messages? Normally, you could find the log at /var/log/messages.

[dutor]

Otherwise, if the number of processes exceeds ulimit -u, you could change it to some larger number

[wey-gu]

Will close it as it's been inactive for days. Feel free to reopen it.