httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From William A Rowe Jr <wr...@rowe-clan.net>
Subject Re: Worker states for balancer members
Date Wed, 03 Feb 2016 16:15:51 GMT
On Feb 3, 2016 6:25 AM, "Jim Jagielski" <jim@jagunet.com> wrote:
>
>
> > On Feb 3, 2016, at 3:10 AM, Plüm, Rüdiger, Vodafone Group <
ruediger.pluem@vodafone.com> wrote:
> >
> >
> >
> >> -----Original Message-----
> >> From: Jim Jagielski [mailto:jim@jaguNET.com]
> >> Sent: Dienstag, 2. Februar 2016 22:44
> >> To: httpd
> >> Subject: Worker states for balancer members
> >>
> >> Getting back to this discussion...
> >>
> >> Right now we have:
> >>
> >>   o #define PROXY_WORKER_DISABLED       0x0020
> >>   o #define PROXY_WORKER_STOPPED        0x0040
> >>   o #define PROXY_WORKER_IN_ERROR       0x0080
> >>   o #define PROXY_WORKER_HC_FAIL        0x0400
> >>
> >> Are these bit flags part of our API?? Could I, for example,
> >
> > I think they are as they are part of mod_proxy.h and other 3rd party
> > proxy modules may have consumed them.
> >
> >
> >> make PROXY_WORKER_DISABLED == PROXY_WORKER_STOPPED == 0x0060
> >> for example?
> >
> > Hm. What if 3rd party modules only set 0x0040 or 0x0020 as they are
using the old
> > values? Wouldn't that create some possible weird side effects?
> > I mean no issue for trunk, but don't think this would be correct for
backporting.
> >
>
> Well, that's why I proposed setting them as such, so that
> a bit check would always be true whether testing for 0x20 or
> 0x40 (the old values).
>
> I WAS thinking about basically "making" HC_FAIL STOPPED because
> that mode can only be set/cleared during configuration (the
> balancer-manager doesn't provide for changing that bit) and so
> for the life of me I can't figure out how or why anyone would be
> using that status since it never changes...
>
> Maybe we can just say that STOPPED is there for potential
> 3rd party uses and be done w/ it :)

If we believe one is of very little use, it might not be a bad idea to
@deprecate it and yank it in trunk eventually.  I think that is more robust
than mapping two statuses as equality.

Mime
View raw message