This error is extremely hard to reproduce, and I found it can be cleared by restarting squid, which seems to make it go away. It seems to take several hours of non-stop proxy use then once the error occurs the we browser reports "too many redirects" and certificate errors. Doing a restart on Centos 7 clears it: # systemctl restart squid The log shows some sort of "refresh unmodified state before it happens: 1509690588.252 167 127.0.0.1 TAG_NONE/200 0 CONNECT events.bouncex.net:443 - HIER_DIRECT/35.190.62.200 - 1509690588.272 210 127.0.0.1 TAG_NONE/200 0 CONNECT analytics.twitter.com:443 - HIER_DIRECT/199.59.149.200 - 1509690588.280 62 127.0.0.1 TCP_REFRESH_UNMODIFIED/200 38412 GET http://www.latimes.com/nation/la-na-vegas-shooting-sheriff-20171102-story.html - HIER_DIRECT/104.120.143.198 text/html <================== error is here 1509690588.356 220 127.0.0.1 TCP_MISS/200 960 GET https://partners.tremorhub.com/syncnoad? - HIER_DIRECT/34.228.123.38 text/xml 1509690588.366 304 127.0.0.1 TAG_NONE/200 0 CONNECT geo.moatads.com:443 - HIER_DIRECT/52.21.172.68 - 1509690588.374 303 127.0.0.1 TAG_NONE/200 0 CONNECT rtr.innovid.com:443 - HIER_DIRECT/13.58.208.14 - 1509690588.377 33 127.0.0.1 TCP_MISS/200 498 GET https://tribpubdfp745347008913.s.moatpixel.com/pixel.gif? - HIER_ If there are particulars and I attempt to recreate this problem are there any specific logging parms or settings that would help you understand this particular error or shed some light on it that I could set on my end. Jeff _______________________________________________ squid-users mailing list squid-users@xxxxxxxxxxxxxxxxxxxxx http://lists.squid-cache.org/listinfo/squid-users
