errors in systemctl services and documentation (OneProvider 20 and 19)

[AJRubio-Montero]

Good evening.

I have noticed (installing from CentOS 7 repos) that OneProvider' services return errors when the systemctl status command is used.

There are some errors that appear only in cluster_manager and op_panel logs, and we don't know their source:

$ systemctl status cluster_manager
...
 su[4363]: (to cluster_manager) root on none
...
$
$ systemctl status op_panel
...
 su[6473]: (to cluster_manager) root on none
...

And there are other that appear in the three services:

...
 cluster_manager[4352]: Starting cluster_manager: touch: cannot touch ‘/opt/onedata/onedata2002/root/var/lock/su…directory
...
 op_worker[4905]: Starting op_worker: touch: cannot touch ‘/opt/onedata/onedata2002/root/var/lock/subsys/op_wor… directory
...
 op_panel[831]: Starting op_panel: touch: cannot touch ‘/opt/onedata/onedata2002/root/var/lock/subsys/op_panel’… directory

But they can be easily corrected if the corresponding directories are made:

$ mkdir -p /opt/onedata/onedata2002/root/var/lock/subsys/op_panel
$ mkdir -p /opt/onedata/onedata2002/root/var/lock/subsys/op_worker
$ mkdir -p /opt/onedata/onedata2002/root/var/lock/subsys/cluster_manager

Moreover, I have seen in 20.02.04 release that the op_panel service is now able to start the other required services. If you try to start (or to stop) services in the ordered fashion suggested at the current documentation, it results in many errors and orphan processes (not guided by systemctl). I suggest to include this note in that section:

$ systemctl disable couchbase-server
$ systemctl disable cluster_manager
$ systemctl disable op_worker
$ systemctl enable op_panel
$ # reboot 

Cheers and happy new year.