Elasticsearch run as service

[Yungyoung_Ok (ooyy)]

I installed elasticsearch with a tar.gz package.
I want to run as a service so that es can run automatically when the system reboots.
How should I set it up?

elasticsearch: 7.6
os: centos7

Instead of using rpm, we need a method that runs as a service in the current installation plan.

[spinscale (Alexander Reelsen)]

I don't fully understand, why can't you use RPM?

If you cannot use it, you can properly take a look at the RPM and reuse the systemd startup script from the package.

[Yungyoung_Ok (ooyy)]

Yes I changed the installation method to rpm.

I have two questions.
First, a bootstrap check fail occurs.
Error: memory locking requested for elasticsearch process but memoty is not locked.
Here is my config.
Which settings should I add?

(1) /etc/security/limits.conf
elasticsearch - nofile 65336
elasticsearch soft memlock unlimited
elasticsearch hard memlock unlimited
elasticsearch - nproc 4096

(2) /usr/lib/systemd/system/elasticsearch.service
LimitMEMLOCK=infinity

(3) /etc/sysctl.conf
vm.max_map_count=262144

Second, I changed the path.logs(/data) setting, but gc.log is still logged in /var/ log.
How should I set it so that gc.log can also be written to the changed path(/data)?

[spinscale (Alexander Reelsen)]

See https://www.elastic.co/guide/en/elasticsearch/reference/7.6/setup-configuration-memory.html

Did you check that setup and not only the configuration files?

Regarding the gc.log, check the jvm.options file.