bind-chroot, runs, works, dies

Reindl Harald h.reindl at thelounge.net
Wed Aug 30 14:01:07 UTC 2017



Am 30.08.2017 um 15:21 schrieb Tony Finch:
> Petr Mensik <pemensik at redhat.com> wrote:
>>
>> But presence of pid files also work as notification of completed
>> initialization (which is done BEFORE forking and finishing ExecStart
>> command).
> 
> named writes its pidfile relatively early during startup. The parent
> doesn't exit until the child daemon has finished starting, so that's a
> better notification (though not what systemd wants...).
> 
>> Service type=simple is not good replacement of forking, because it does
>> not let you know when service failed to start at all.
> 
> Good grief

it's not true, it just don't let you know when start is finsished but 
other than sysvcrap systemd *always* let you know when a service failed 
and it's easy to write a simple script to catch that events from syslog 
and send periodically mails if that happens

what's better than forking/simple and so on is just native systemd 
support of the service to use type=notify like it is now poissble with 
MariaDB - that's it

maybe i should add that we run Fedora in production for a decade and 
with Fedora 15 i siwtched *every* service to native systemd-units since 
the Fedora mix of old compat services and new units was a unacceptable mess

so when i talk about systemd and servers i know what i talk about since 
we run httpd/mail/dns/epp/smb/afp and what not else likely as one of the 
first companies at all with exclusive native units and nothing else for 
production servers hosting hundrets of domains



More information about the bind-users mailing list