Closed victor-sudakov closed 3 years ago
Interesting case, I never thought of that. This make a case to simply drop the regex and consider the first dataset like any other datasets.
This would also simplify the installation of the template, I think I'll just do that.
Interesting case, I never thought of that. This make a case to simply drop the regex and consider the first dataset like any other datasets.
This would also simplify the installation of the template, I think I'll just do that.
It will probably break the detection of zpools? Doesn't the detection of zpools also rely on this regex?
Nope, the zpool LLD is based on zpool list
so that's completely unrelated, this will also solve #22 since the expression will no longer be necessary.
Nope, the zpool LLD is based on
zpool list
so that's completely unrelated, this will also solve #22 since the expression will no longer be necessary.
So, is it safe to remove those 2 filters from my production Zabbix system? And my top-level datasets like /users and /d01 will start being monitored?
@victor-sudakov : import the latest version of the template first, then you can remove the global expression. You should see at the next discovery cycle your "base" dataset.
@victor-sudakov : import the latest version of the template first, then you can remove the global expression. You should see at the next discovery cycle your "base" dataset.
It works! Thanks.
Looking at git diff 12eecbf..ffc0105
though, I could have simply deleted the filters without updating the template.
This template differentiates a zpool from a zfs dataset by the presence of "/" in the name, is this correct?
I happen to have some hosts where the zpool itself is the only dataset, like here:
Thus I'm missing useful statistics (e.g. compression ratio) about the dataset "users" because the template does not recognize "users" as a rightful dataset.
Is there a way to circumvent this other than create children datasets in the pool and move all data thereto (I would hate this)?