1. Fault description:
Label a host with the EFK label, and let this host deploy data, master, and client containers.
As a result, the pod on this host could not get up. The pod keeps reporting CrashLoopBackOff.
2. Solutions:
Don’t panic if you encounter a problem. The first thing is to look at the log first, check the pod’s log, and report this error
max file descriptors [50000] for elasticsearch process is too low, increase to at least [65536], start looking for this error Solution.
3. Solution:
(1) Check that the error report means that the maximum file descriptor value of the elasticsearch process is too low and increased to at least [65535].
(2) Check the information and consult a colleague. Finally, I modified the following two parameters in the docker.service file to 65535
LimitNOFILE=65535
LimitNPROC=65535
(3) Restart docker. All three pods of elasticsearch are up. The problem is solved.