1.2-RELEASE becomes unstable, CP and GUI not loding
-
Hey long time i havent been here…
I got small problem, my 1.2-RELEASE box become unstable, symptoms are CP doesnt work, barly loads the page is page is loaded form doesnt get submitted ot rakes 30-80 minutes to submit. Also I can not open any GUI pages, takes 30 min to open any page. When I reboot the device(phisical power unplug) everything becomes ok for day or so then same problem. I thought thats a hardware problemm so i arranged ac in the rooom, that didnt help... that I disapled the CP and it seamd to help but I dont know (i did it just now, will see the results)And yes the one thing, if users where authenticated before CP crashes, the internet connection works fine, dhcp works fine... to me it lookd like internel webserver crash (miniHTTPD or whatever PF is running) Am i right? How can I fix it, and more what could cause this problem? I hope you can help me out, tnx!
-
Sounds like hardware or corrupt data to me. We have used rc4 for months with no issues and have 1 release in use for the last couple months with no issues. How long has the system been up? We do some tweaks and reset every now and then but it just sounds like corrupt data or bad hardware on your end.
-
Well before the first system crash the box was up since the release of 1.2 what that 6 month, and then after forsyt issue of "webserver" carsh it was up for 3 days and crashed again. … technically i can make reset code to run each week but i'm afraid if it resets itself and it doesnt boot up in the middle of the night I will have to go creazy getting the tech out there and fixing it :(
-
Ok. The software seems fine, we have had 1.2 release up on one of ours for this long "51 days, 00:17". almost 2 months, and 1.2rc4 on all the rest for even longer. We run CP on all of them. It still seems to be a hardware issue. Maybe even a hard drive issue. If it is a hard drive that is dying, then there is possibly some bad sectors on that drive, and resetting is a temporary fix. Could possibly be memory too, but I'd suspect hard drive and moving parts 1st. If the sectors that are bad or going bad contain any CP data, this could be causing your problems. It sounds like it "eventually" works, therefore sounds like some type of read error on the drive. Hope you get this fixed, because rebooting or a script written to reboot is just a work around and not a fix.
-
Are you exhausting the state table maybe? Check the front page of the web interface when this occurs, or just increase the state table size and see what happens.
-
i think i may have the same problem as him… first it would hang every 4-5 days, and now it's becoming more frequent, every day it seems to do this and i have to restart, the router just seems to freeze, i can't access the webGUI, i can't SSH in, and i have the same version as you. the states don't go above 1,900/10,000, so my states table doesn't fill up, but maybe my problem is different, thought i'd atleast reply to let you know you may not be the only1 with the problem.
-
The web server in pfSense is lighttpd in pfSense 1.2 it is configured by default for 1 maybe 2 connections only. This is okay for just standard firewall config on a limited resource box. This is bad for captive portal especially if you have more than one person connecting to captive portal at the same time.
pfSense 1.3 addresses this issue when captive portal is enabled the lighttpd is given more resources so it can handle more connections.
If you want to manually add these changes now to your pfSense firewall take a look at the following links.
This prepares the pfSense built in web server for more concurrent traffic.
http://forum.pfsense.org/index.php/topic,8861.msg50280.html#msg50280This helps optimize PHP so it doesn't hold web server resources for a long period of time.
http://forum.pfsense.org/index.php/topic,8878.0.htmlA workaround for now that will automatically remove the lock file after it is older than 3 minutes.
http://forum.pfsense.org/index.php/topic,8152.msg57899.html#msg57899After making these changes I have supported over 130 people behind captive portal with no further problems.