On Thu, Sep 25, 2008 at 11:36:36PM +0200, Florian Forster wrote:
> I see your point, I'm in the same situation myself. However, I feel
> somehow uncomfortable about making this ``fast exit'' feature the
> default behavior. Can't tell you precisely why, it just doesn't feel
> ``right'' ;)
> 
> In any case, the user should have a choice. Using a different signal for
> this is a good idea, though I'd use SIGUSR1 or SIGUSR2 for this. SIGTERM
> and SIGINT aren't really different enough for this to make much sense
> (imho).

I understand what you guys ar saying.. I'm fine with preserving the same
default behavior.  I like the idea of using a signal for this, since it's
easy to invoke when you only know the PID (i.e. at system shutdown)..  So
I'll use SIGUSR1 for that.

> Another, possibly additional, idea would be to add `SHUTDOWN' and
> `BUGGEROFF'[*] commands.

I think a signal is easier for this.

On the same note, I was thinking of a command "FLUSHALL" which triggers an
immediate flush to disk similar to what happens during shutdown.  What do
you think?  (per earlier discussions on read/read+write control channels,
I think this would be appropriate for the read+write channels only).

-- 
 kevin brintnall =~ /[EMAIL PROTECTED]/

_______________________________________________
rrd-developers mailing list
[email protected]
https://lists.oetiker.ch/cgi-bin/listinfo/rrd-developers

Reply via email to