On 24/08/17 01:42, Antony Stone wrote:
On Wednesday 23 August 2017 at 15:26:53, Verónica Ovando wrote:
This is the last request until it holdups:
1503416750.346 386 192.168.0.2 TCP_REFRESH_UNMODIFIED/304 235 GET
http://www.mysite.com user HIER_DIRECT/www.mysite.com - This is the next
request:
1503416781.864 141376 192.168.0.2 TCP_MISS/200 5599 CONNECT
play.google.com:443 user HIER_DIRECT/play.google.com -
Okay, so I see there's a 36 second gap between those two requests; were
clients sending requests during that time?
Ah, the second log entry is for a CONNECT tunnel that was started 111.5s
before the first request and took 140 seconds to deliver all its 5KB of
tunneled traffic.
The time it was open and not doing anything useful might be related to
the delays, but not an indicator of anything specifically related. It
could just be a client holding an HTTPS connection with keep-alive until
the TCP layer or Squids timeouts kill it.
(The available FD slots running out temporarily, or bursts of socket
FDs going through TCP *_WAIT states could be the reason for pausing).
Amos
_______________________________________________
squid-users mailing list
squid-users@xxxxxxxxxxxxxxxxxxxxx
http://lists.squid-cache.org/listinfo/squid-users
