Re: [WIP] pg_ping utility

On Sat, Nov 17, 2012 at 2:48 AM, Phil Sorber <phil(at)omniti(dot)com> wrote:

> On Thu, Nov 15, 2012 at 10:55 PM, Michael Paquier
> <michael(dot)paquier(at)gmail(dot)com> wrote:
> > On Fri, Nov 16, 2012 at 12:34 PM, Phil Sorber <phil(at)omniti(dot)com> wrote:
> >> On Thu, Nov 15, 2012 at 9:23 PM, Michael Paquier
> >> <michael(dot)paquier(at)gmail(dot)com> wrote:
> >> > 3) Having an output close to what ping actually does would also be
> nice,
> >> > the
> >> > current output like Accepting/Rejecting Connections are not that
> >>
> >> Could you be more specific? Are you saying you don't want to see
> >> accepting/rejecting info output?
> >
> > OK sorry.
> >
> > I meant something like that for an accessible server:
> > $ pg_ping -c 3 -h server.com
> > PING server.com (192.168.1.3)
> > accept from 192.168.1.3: icmp_seq=0 time=0.241 ms
> > accept from 192.168.1.3: icmp_seq=0 time=0.240 ms
> > accept from 192.168.1.3: icmp_seq=0 time=0.242 ms
> >
> > Like that for a rejected connection:
> > reject from 192.168.1.3: icmp_seq=0 time=0.241 ms
> >
> > Like that for a timeout:
> > timeout from 192.168.1.3: icmp_seq=0
> > Then print 1 line for each ping taken to stdout.
>
> How does icmp_seq fit into this? Or was that an oversight?
>
Yes, sorry it doesn't fit in this model. Please forget about it.

> Also, in standard ping if you don't pass -c it will continue to loop
> until interrupted. Would you suggest that pg_ping mimic that, or that
> we add an additional flag for that behavior?
>
By targeting pg_ping as a clone of ping, yes it would mean that we target
it to loop indefinitely if no c flags is given.

> FWIW, I would use 'watch' with the existing output for cases that I
> would need something like that.
>
watch allows you to launch a program given a certain time period. I am not
sure this is related with pinging a server.
When pinging a server, what you are looking for is not only the
connectivity to it but also the latency you have with it, no?
--
Michael Paquier
http://michael.otacoo.com