There is currently a problem with stopping Bitbucket server instance in data center:
While starting a bitbucket server instance with "start-bitbucket.sh" in DataCenter supports "--no-search" option to not start a local elastic-search instance, the "stop-bitbucket.sh" does not support such an option.

As a result the script does successfully stops the bitbucket instance, but then tries to stop the not-existent elasticsearch instance resulting in an errorcode.

As a consequence, using the script in initd/systemd the "systemctl stop bitbucket" will "fail" and the service is marked as "failed", even if was successfully stopped. This is a very annoying situation.

Example of /etc/systemd/system/bitbucket.service:

[Unit]
Description=Bitbucket
After=syslog.target network.target[Service]
Type=forking
User=bitbucket
Environment=JAVA_HOME=/usr/java/default
PIDFile=/var/opt/atlassian/bitbucket/log/bitbucket.pid
ExecStart=/opt/atlassian/bitbucket/bitbucket_current/bin/start-bitbucket.sh --no-search
ExecStop=/opt/atlassian/bitbucket/bitbucket_current/bin/stop-bitbucket.sh
[Install]
WantedBy=multi-user.target
 

As in "start-bitbucket.sh" the "--no-search" should be implemented in stop-Bitbucket and - if set - skip the part to stop the (non-exsitent) local elasticsearch instance.