Re: Re: Bug#727708: tech-ctte: Decide which init system to default to in Debian.
>> You don’t want anything like these in your local init service. For such
>> tests you have Nagios, Icinga or similiar daemons. And they can do much
>> deeper checks, e.g. can you login into your webservice because your
>> database backend on a different server is available.
>
> Once your monitoring system – your costly monitoring system with someone
> behind it just to check whether your buggy scripts have failed to start
> your service — has detected the service is not running, what will you do
> to understand what has gone wrong?
Obviously, if you ask such a question - you never had any good
expirience with mentioned above "costly monitoring solution". Then
why you claim that other people are "incompetent"? You are.
FYI. The monitoring system will restart your service on failure, just
like systemd. Or it will run gdb on core and send you a message, or
do whatever you configure it to do. And the point is what tests
(and actions) could be much more configurable than systemd's. The best you
can do with systemd's "monitoring" functional is to switch it off.
>> I don’t mind if you want systemd, but don’t force it on others. There are
>> no features in systemd that I would dump the well known sysvinit.
>
> I don’t mind if you don’t want these features, but don’t force
> others to run their Debian systems with software from the previous
> century that lacks them.
If they want systemd - fine. I'll happy to support this as a variant
for init system. It's already in the Debian, please use this.
But why you force others to this overbloated system?
> Nobody is forcing you to use these features.
How I can be sure? There is the "Restart" configuration directive. Yet
another "competent" (yes, just like you) Debian maintainer could add
a default service file with Restart=on-failure, for example.
Why not? It's "cool" and from this century, right?
Reply to: