CentOS 7 Cassandra systemd service file

cassandra.service

# /usr/lib/systemd/system/cassandra.service

[Unit]
Description=Cassandra
After=network.target

[Service]
PIDFile=/var/run/cassandra/cassandra.pid
User=cassandra
Group=cassandra
ExecStart=/usr/sbin/cassandra -f -p /var/run/cassandra/cassandra.pid
StandardOutput=journal
StandardError=journal
LimitNOFILE=100000
LimitMEMLOCK=infinity
LimitNPROC=32768
LimitAS=infinity
Restart=always


[Install]
WantedBy=multi-user.target

Thank you!! This was helpful.

worth adding:
EnvironmentFile=-/etc/default/cassandra

Thank you, it's very helpfull ! :)

Thank you, very helpfull,
I was facing with a strange problem with Cassandra, nomore managed correctly as service on Centos7 because such file was missing after a yum update.
Putting that file under /usr/lib/systemd/system/ and doing a "systemctl daemon-reexec" solved my problem :-)

Not sure why a PID file is being set if the unit is left to default to simple?

After some tries I've created upgraded version of this unit file.

Changes:

• use background run + Type=forking because -p works only without -f
• use CASSANDRA_* env variables to load cassandra configs from /etc/cassandra/conf, those are created by cassandra RPM package
• upd 28.08.2020: add stop timeout
• upd 28.08.2020: set successful exit code

Tested with cassandra 3.11.5 & centos 7.7

# /usr/lib/systemd/system/cassandra.service

[Unit]
Description=Cassandra
After=network.target

[Service]
RuntimeDirectory=cassandra
PIDFile=/var/run/cassandra/cassandra.pid
Environment=CASSANDRA_HOME=/usr/share/cassandra
Environment=CASSANDRA_CONF=/etc/cassandra/conf
Environment=CASSANDRA_INCLUDE=/usr/share/cassandra/cassandra.in.sh
Environment=CASSANDRA_OWNR=cassandra
User=cassandra
Group=cassandra
Type=forking
ExecStart=/usr/sbin/cassandra -p /var/run/cassandra/cassandra.pid
StandardOutput=journal
StandardError=journal
LimitNOFILE=100000
LimitMEMLOCK=infinity
LimitNPROC=32768
LimitAS=infinity
Restart=always
TimeoutStopSec=180
SuccessExitStatus=143


[Install]
WantedBy=multi-user.target

Tested on CentOS 8, works!

On my sysV file I call the following command when stopping cassandra:
nodetool -u cass** -pw pass** -h 127.0.0.1 decommission

Do I need to do this in systemd file as a ExecStopPost command?

Should be useful in forking mode:

TimeoutStartSec=300

But foreground mode seems more stable (imho).