Subject: Some ideas for #11579
Dears,
First of all, thanks for Wenjun Ruan's advice, allowing me to have the
opportunity to say some of my ideas about the issue:#11579
<https://github.com/apache/dolphinscheduler/issues/11579>
Here are some of my thoughts, please review them:
1) I think that the current service start, stop and service check scripts
are not named accurately enough to convey the meaning. Could we renamed to
start-cluster.sh、stop-cluster.sh and status-cluster.sh, or just cluster.sh
start|stop|status.
2) I think it is better to store only the start、stop and status scripts of
the cluster service in the /bin directory. Move the tool script such
as scp-hosts.sh and remove-zk-node.sh to /tools dir. This can help users
focus on the operation scripts of the cluster service, while shielding
users from the tool scripts in the process.
3) Could we consider adding entry scripts for starting and stopping a
single application in the /bin directory, not just
{application}/bin/start.sh
4) The printing format of some log in the current script can also be more
elegant and tidy. And i think write to a log file at the same time should
be better for product env.
Sure, personal suggestion, for reference only.
Thanks.
B. R.
Yann
