[squid-users] Trouble with TIMEOUT_DIRECTs

[Andre Kajita]

Greets,

I have 4 squids on the same subnet with the same distance to my uplink, 2 of them as transparent proxies (NAT firewalls) to our internal networks and the other 2 are stand-alone proxies used by our external networks - they're 3 Redhat 7.3 and a single 8.0 machines and they're all configured to use each other as peers (everyone is a sibling to everyone else). An example:

In squid.univap.br's /etc/squid/squid.conf:
cache_peer orthanc.univap.br sibling 3128 3130
cache_peer academico.univap.br sibling 3128 3130
cache_peer tiziu.univap.br sibling 3128 3130

I used to have the 'no-query' option for each sibling on every machine but I noticed that, despite having heavy usage to the same basic sites they weren't sharing cached information so I configured all the ICP ports and options and removed the 'no-query' from everyone and restarted.

Things are working fine and dandy with no noticable differences, just as they were before but now I'm picking up hundreds of "TIMEOUT_DIRECT" messages on each of the proxies. I added the 'no-query' option again and the messages stopped but so did the ICP traffic between the machines (that I logged for a while and were, as I could tell, working).

Here's a sniplet of a Calamaris report from two of the machines for the past few hours:
>>
DIRECT Fetch from Source 105555 99.99 888426K
DIRECT 80736 76.48 687990K
TIMEOUT_DIRECT 24819 23.51 200436K
>><<
DIRECT Fetch from Source 20556 99.93 141170K
DIRECT 15394 74.84 94164325
TIMEOUT_DIRECT 5162 25.09 50393972
<<

First of all, what the heck does "TIMEOUT_DIRECT" mean?

Second, what does removing 'no-query' from the peer settings have to do with TIMEOUT_DIRECTs?

I've scoured Google trying to find this out and I've picked up quite a few Calamaris reports, some with siblings and parents but very few of them come close to having 20%+ TIMEOUT_DIRECTs and those have no aparent siblings.

I don't know if I f*cked some configuration option (which would explain the fact that there are no Google records on this type of problem) or I have such a unique topology (unlikely, but hey, who knows). I'm really stuck here, so far it hasn't affected anything but I got a feeling that so many TIMEOUT_DIRECTs on a stable link aren't normal.

Thanks in advance for any tips on figuring this one out.

Andre.
--
Andre Kajita
[EMAIL PROTECTED]
--
In Googlis non est, ergo non est.