I experience the same issue.
Found out that it has to do with an increase in the number of open IP connections over time. Router does not perform any more when it gets close to the max (in my case 4096, and yes, even when i set the timeouts very low).
I don't know how this happens, router runs well for some time and then almost hangs suddenly. In all cases this is due to a high number of open IP connections.
A reboot solves this (for a couple of days then).
No, i do not have P2P running. _________________ WRT54GS v4.0 with DD-WRT v23 SP1
WRT54GL v1.1 with DD-WRT v23 SP2
WRT350N v1.0 with DD-WRT v24 mega
La Fonera with FreeWLAN Addons 0.9.0
La Fonera with OpenWRT LEGEND
Although I am *not* using ssh or https because i have the micro version. For the last month, using a telnet session, I have watched the httpd process gobble up more and more memory until the router reboots itself. Usually takes a week or so. Just before the router reboots, CPU is extremely high. However, also note that memory used by the httpd process *only* increases when I have a browser(IE or FF) session open to the router. Even after killing the browser session httpd does not release the memory.
As a side note, after the last reboot, I purposely did not open a browser sesssion to the router and memory used by the httpd process is consistently hovering around 400k without ever creeping up.
I usually have only 1 client and that client is not doing anything strenous to the router... my wii occasionaly connecting to do updates and such
Since my last post above on "Wed Oct 10, 2007 4:34 pm" I have not had a lockup or any problems at all. But, please read below.
One thing I did notice. On occasion, I did use a browser to take a look at things on the router. During this time i used a telnet session to monitor the httpd process and its memory. After I closed the browser session, as noted before, none of the consumed memory was released. However, i then used the telnet session to kill the httpd process(killall httpd) and let watchdog restart it. Used memory was then released and things went back to normal.
Hi mates, i´ve a linksys Wrt-54gl with DD-WRT v23 SP2STD and i´ve the same problem, i´ve activate the https option and if i use remote access , httpd process overload 100% the cpu of the router. Do have any solution for the bug, actually ?
Im having the same problem here. I am not running p2p but I do have a web server, game server, and a voip server running that all together use a couple thousand connections at any given time. I have my max connections set to 4096 for this purpose. But I have noticed extremely high cpu usage, and looking at top I can see httpd using 75%-100%. The router will lock up a couple times a week, and it won't function again until it gets a hard reboot. Is there any solution yet?
Hi mates, i´ve a linksys Wrt-54gl with DD-WRT v23 SP2STD and i´ve the same problem, i´ve activate the https option and if i use remote access , httpd process overload 100% the cpu of the router. Do have any solution for the bug, actually ?
Thanks and sorry for my english.
Hi mates, i fix the problem with the new version dd-wrt v24std.
OK all......svn 9910 build has this problem fixed. I am testing for Eko and BS right now. All the 100% CPU issues are gone.
I hope to talk with BS and Eko this morning....but just be aware....it will be fixed in the next TNG release.
[EDIT] - and before anyone asks....I can not provide 9910-micro. The devs have to make the call as to when it is released....sorry
redhawk
I am going back to have the same problem with the dd-wrt 24std sp1 and Linksys WRT-54GL V1.1, my https wan access is cut every few days . All proccess working ok, but https go up until 98% of average of cpu charge. I only can fix it restarting.
I am going back to have the same problem with the dd-wrt 24std sp1 and Linksys WRT-54GL V1.1, my https wan access is cut every few days . The rest of the proccess working ok, but https go up until 98% of average of cpu charge. I only can fix it restarting.
