Not my department. I am but a humble hacker. I expect Hendrik or Duane will
actually correct all the mistakes I made and slap something in for the next
release. They're good like that.
D
Edwin Culp wrote:
> Extraordinary, as always, Thank you. Don't you think that should be
> included in the squid.conf?
>
> Thanks Again
>
> ed
>
> Dancer wrote:
> >
> > Alright:
> >
> > # Allowing or denying direct connection to sites based on access lists.
> > # This comes in two flavours. always_direct and never_direct
> > #
> > # allow_direct allow|deny [!]aclname
> > #
> > # If aclname matches and the rule is 'allow', then squid is permitted to
> > make direct
> > # connection to the target. If no allow_direct lines are present, then
> > the default is
> > # to use the neighbour-selection algorithm, as defined elsewhere in this
> > configuration.
> > # Rules are matched top-to-bottom (ie: Earlier rules take precedence over
> > later ones)
> > # If there is at least one allow_direct line, an implicit rule is created
> > at the end that is
> > # the opposite of the last explicitly stated line (deny if you said
> > allow, allow if you said
> > # deny [someone check me on this - Dancer]). If allow_direct rules exist
> > and there is
> > # no allow line matching the current request, the request must go through
> > a neighbour
> > # or a parent _regardless_ of other selection criteria.
> > # The default is to use the neighbour selection criteria defined
> > elsewhere in this config
> > #
> > #always_direct allow all
> > #
> > # never_direct allow|deny [!]aclname
> > #
> > # Never direct works identically to allow direct but the sense of allow
> > and deny are reversed.
> > # ('never_direct allow all' means 'always use parents or neighbours for
> > every request,
> > # and if it is not possible to do so, generate an error).
> > #
> > # The default is to use the neighbour selection criteria defined
> > elsewhere in this config.
> > #never_direct deny all
> >
> > I've had a few too many drinks this-eve with the boss. This explanation
> > is the best I can do before a meal and coffee. Hope it helps.
> >
> > D
> >
> > Edwin Culp wrote:
> >
> > > Daniel Flinkmann wrote:
> > > >
> > > > This would also not work.
> > > >
> > > > I found the wrong config.
> > > >
> > > > I have to change to the following:
> > > >
> > > > never_direct allow all
> > >
> > > I think a lot of us get stuck here. Couldn't someone add an
> > > example to the squid.conf to replace the XXX that it has?
> > > I would be glad to do it if someone will direct me as to
> > > the change procedures.
> > >
> > > Another place where I and from what I have seen on the list
> > > is with round-robin as suggested below.
> > >
> > > >
> > > >
> > > > then everything work via the proxy :-)
> > > >
> > > > Daniel
> > > >
> > > >
> > > MUNCH
> > > >
> > > > if you have more than one parent, then 'single_parent_bypass' is
> > > > meaningless.
> > > >
> > > > You probably want to use round-robin.
> > > >
> > > > cache_host foo parent 3128 3130 round-robin
> > > > cache_host bar parent 3128 3130 round-robin
> > > >
> > > > Duane W.
> > > >
> > > > ---------------------------------------------------------------
> > > Has anyone been able to get round-robin to distribute load. I tried
> > > most of the day yesterday and it would use the first parent only.
> > > It doesn't seem to be doing what it says in distribute the load
> > > to the parent that has been consulted less.
> > >
> > > I was trying 1.2beta14 yesterday, but have tried almost all and have
> > > never been successful.
> > >
> > > thanks
> > >
> > > ed
> >
> > --
> > Did you read the documentation AND the FAQ?
> > If not, I'll probably still answer your question, but my patience will
> > be limited, and you take the risk of sarcasm and ridicule.
-- Did you read the documentation AND the FAQ? If not, I'll probably still answer your question, but my patience will be limited, and you take the risk of sarcasm and ridicule.Received on Mon Feb 09 1998 - 06:52:31 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:38:49 MST