Re: [squid-users] parents and siblings in a complex enviroment

From: Marc Elsen <marc.elsen@dont-contact.us>
Date: Thu, 19 Jun 2003 15:23:41 +0200

Eric Wieling wrote:
>
> I have 4 caches. 2 caches at the HQ and 2 caches at a branch office.
> Here's the peer info for each of the caches.
>
> hq-1:
> cache_peer hq-2 sibling 3128 3130
>
> hq-2:
> cache_peer hq-1 sibling 3128 3130
>
> branch-1:
> cache_peer hq-1 parent 3128 3130
> cache_peer hq-2 parent 3128 3130
> cache_peer branch-2 sibling 3128 3130
>
> branch-2:
> cache_peer hq-1 parent 3128 3130
> cache_peer hq-2 parent 3128 3130
> cache_peer branch-1 sibling 3128 3130
>
> The two HQ caches are siblings, the two caches at the BRANCH are
> siblings and use the two caches at HQ as parents.
>
> When a request comes into a BRANCH cache, the cache queries it's sibling
> and both parents at HQ. Do the parents then query each other (since
> they are siblings)? If so, does anyone have ideas on the best way to
> prevent these duplicate queries?

 Going down the stairs in your house is an orthogonal sequence of events
 since usally what happens on the next steps is not related to what
happened
 3 steps ago,well...

>prevent< : don't make the HQ caches siblings.

 M.

>
> Thanks in advance,
>
> --Eric
>
> --
> BTEL Consulting
> 850-484-4535 x2111 (Office)
> 504-595-3916 x2111 (Experimental)
> 877-552-0838 (Backup Phone)

-- 
 'Love is truth without any future.
 (M.E. 1997)
Received on Thu Jun 19 2003 - 07:23:53 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:17:26 MST