[systemd-devel] How to restart my socket activated service safely ?

[Francis Moreau]

On Tue, Jul 27, 2021 at 11:12 AM Mantas Mikulėnas <grawity at gmail.com> wrote:
>
> On Tue, Jul 27, 2021 at 10:10 AM Francis Moreau <francis.moro at gmail.com> wrote:
> >
> > Hello,
> >
> > During my application update, I want to restart my service which is
> > activated by a socket but want to be sure that no request sent to my
> > service will be missed. I also want to restart the socket too so
> > systemd uses the latest version of the socket unit file.
> >
> > If I restart the socket when the service is still running then I get
> > an error message: "rotor.socket: Socket service rotor.service already
> > active, refusing."
> >
> > If I stop the service first and restart the socket then there's a
> > short time frame where requests can be lost.
>
> The old socket has to be unbound before a new one can be put in its
> place. Trying to keep the service alive (holding the old listener fd)
> would just result in systemd not being able to bind a new socket with
> the same address... (And even if that was possible, the old service
> wouldn't be able to handle requests arriving on the new socket
> anyway.)
>
> So whenever you restart a socket, there will *always* be a short time
> frame where the old socket is closed but the new one is not yet
> bound/listening. But as soon as the new one is listening, it'll start
> queuing the requests even if the service isn't yet running (since it's
> a socket-activated service after all) and the number of lost requests
> should be minimal.

Thank you for response.

If the service only is restarted, will be there also requests that can be lost ?

If the service is restarted then it's first stopped but in the
meantime the socket unit should be listening for new requests...

-- 
Francis