1.2.3-RELEASE – TCP 80 suddenly stops for clients
After spending 30 minutes writing the problem description below, I found the problem and solution: /var/squid/cache had consumed all available hard drive space! So, for anyone else reading this post because port 80 stops randomly, please check the obvious (a full hard drive), which leads to an easy solution (#killall squid, #rm /var/squid/cache/swap.state, #reboot, #crontab -e * 12 * * * /usr/local/sbin/squid -k rotate). I'll review this post for performance tweaks (http://doc.pfsense.org/index.php/Squid_Package_Tuning) and report back if I see any ongoing issues.
System: 1.2.3-RELEASE, Name is pfsense1. Dell DXP051. Two aftermarket NICS: LAN (192.168.0.0 class C) and WAN (x.x.x.x) as router and transparent proxy for small office environment in USA.
Packages: Lightsquid, squid
Problem: "tcpdump -i vr0" shows (and employees confirm) that port 80 works fine, then stops functioning. Other protocols working normally (including TCP 25, 23, 443(?))
Workaround: We activated (turned on) pfsense2, an cold swap (i.e. offline) pfsense configured identically to the primary, online now.
History: This problem started yesterday after months of working perfectly.
Known Problems: After rebooting pfsense, I got this problem (http://forum.pfsense.org/index.php?topic=35253.0). The recommended solution (I deleted squid and Lightsquid, then reinstalled them) fixed the reboot slowness. Then, port 80 worked perfectly for about 1 hour, then the problem started again.