Kong: fail to handle USR2 signal

[qiangcz]

Summary

when i send USR2 signal to kong, i get error below:

[notice] 103716#0: signal 12 (SIGUSR2) received from 25984, changing binary
[notice] 103716#0: changing binary
[notice] 103716#0: start new binary process 48397
[alert] 48397#0: execve() failed while executing new binary process "nginx" (2: No such file or directory)
[notice] 103716#0: signal 17 (SIGCHLD) received from 48397
[notice] 103716#0: new binary process 48397 exited with code 1

because kong start nginx with relative path, how to resolve it.

Steps To Reproduce

1. kill -USR2 {nginx master pid}
   2.
   3.
   4.

Additional Details & Logs

• Kong version (1.1.0)
• Kong debug-level startup logs ($ kong start --vv)
• Kong error logs (<KONG_PREFIX>/logs/error.log)
• Kong configuration (the output of a GET request to Kong's Admin port - see
  https://docs.konghq.com/latest/admin-api/#retrieve-node-information)
• Operating system

[hishamhm]

Could you use kong restart instead?

[qiangcz]

Could you use kong restart instead?

no, I have to make sure the business is continuous without LB or others.

[qiangcz]

is anyone here?

[p0pr0ck5]

Try setting the process PATH to include /lusr/local/openresty/nginx/sbin, which is where the Kong nginx binary lives.

[qiangcz]

Try setting the process PATH to include /lusr/local/openresty/nginx/sbin, which is where the Kong nginx binary lives.

thanks for your reply, but unfortunately it does not work

[p0pr0ck5]

@thibaultcha I recall recently seeing your patch to nginx-devel about USR2 handling; any ideas here, since you've been in that code recently?

[thibaultcha]

@p0pr0ck5 The patch is for a different issue, but https://github.com/Kong/kong/pull/5657 should do the trick for this one.

[qiangcz]

thanks all