On Sat, Feb 09, 2013 at 10:47:23AM +0100, Marc-Antoine Perennou wrote:
> I'll do even simpler, when submitting V2 of this patchset with the EINTR
> issue,
> I'll basically forard every arg given to the wrapper to all the children
> themselves.
> The tool is not really meant to be used by the user himself though, but
> rather to
> be launched by the systemd service.
I understand but it's quite common to expect to reuse the same command
line everywhere.
> Is SIGUSR2 ok here ? I first did it with SIGUSR1 but then children couldn't
> bind
> to this signal on reload, since it was already a USR1 action, so I took the
> first one not colliding.
6 signals are used by haproxy :
- SIGTTOU is used to make the old process temporarily release the ports
it is bound to so that a new process may attempt a startup
- SIGTTIN is used to tell the old process to rebind the ports that were
unbound by SIGTTOU and continue its job as if nothing happened
- SIGUSR1 is used to make the old process unbind and wait for the last
connection to complete before exiting
- SIGTERM is used to make the old process go away without waiting for
connections to complete
- SIGHUP is used to dump server states in the logs
- SIGQUIT is used to dump memory pool usages to the logs
Additionally, I'd rather keep SIGUSR2 available for later use, because
the master-worker rework that Simon Horman did makes use of this signal
to restart the process and reopen config files.
However I think that your usage of SIGUSR2 is exactly the same as for
the master-worker mode so it seems to make a lot of sense to use it in
the wrapper.
Willy
