/ Zope / Apsis / Pound Mailing List / Archive / 2007 / 2007-09 / possible minor pug in pound/poundctl

[ << ] [ >> ]

[ X-Forwarded-for header behaviour / Alex Hunsley ... ] [ Multiple HeadRequire in one service / "Nigel ... ]

possible minor pug in pound/poundctl
Alex Hunsley <alex.hunsley(at)blueyonder.co.uk>
2007-09-27 14:10:23 [ FULL ]
Been looking at the source code and playing and I think I've worked out 
what seemed like odd behaviour to me....

I have the following service running in pound, as a test:

  0. http Listener 192.168.0.39:8123 a
 -1. Global services
    0. Service active (5)
      0. Backend PF_INET 192.168.0.39:65 active (5 0.000 sec) alive


If I now disable the backed with:

  ./poundctl  -b 0 0 0

(i.e. listener 0, service 0, backend 0), then run poundctl again, it says:

  0. http Listener 192.168.0.39:8123 a
 -1. Global services
    0. Service active (5)
      0. Backend PF_INET 192.168.0.39:65 active (5 0.000 sec) alive

i.e. it still shows backend 0 as being active.

Now, if I disable the backend with this command instead:

  ./poundctl  -b -1 0 0

(i.e. listener -1, server 0, backend 0), then run poundctl again, it says:

  0. http Listener 192.168.0.39:8123 a
 -1. Global services
    0. Service active (0)
      0. Backend PF_INET 192.168.0.39:65 DISABLED (5 0.000 sec) alive

.. so now it is correctly reporting the backend as disabled.

So is this expected behaviour?
I would have expected "-b 0 0 0" to disable the backend, according to 
the help text that poundctl outputs.... (i.e. "-b n m r - disable 
back-end r in service m in listener n").


thanks,
Alex

Re: [Pound Mailing List] possible minor pug in pound/poundctl
Robert Segall <roseg(at)apsis.ch>
2007-09-27 19:12:06 [ FULL ]
On Thu, 2007-09-27 at 13:10 +0100, Alex Hunsley wrote:[...]

Yes.
[...]

The "general" service (catch-all) is -1, exactly as shown.[...]

MailBoxer