radiusd fails after few hours

[floaty]

radiusd fails when receiving duplicated rad-requests from cisco-wlc-4404.
is there an possibility to monitor and restart the radiusd, when this problem occurs?

####################
rml-detail: couldn´t open file /var/log/radius/reply: too many open files

[imported_fulvio]

I will try to fix the problem in the next release.
In any case you should put the following line in the file /etc/rc.local

echo 4194304 > /proc/sys/fs/file-max

and then reboot the box.
This could be a temporary solution.

Regards
Fulvio

[floaty]

thank you, i will post the results for this workaround asap

[floaty]

filehandles increased … looks better now.

(the duplicated rad-requests from wlc440x seems to be normal behaviour …
if wpa and wpa2 with AES and TKIP are enabled the controller sends 4 requests with only wpa2 and AES only it is 1 request)

[floaty]

… is now running 4 days till it crashs

… i have tested with NTRADPing on my test-server (without 802.1x) … the open files are not increased per rad-auth-(acct)-request

[

root@foorad fs> cat file-nr
384 0 12083
root@foorad fs> cat file-nr
384 0 12083
root@foorad fs> cat file-nr
384 0 12083
root@foorad fs>

]

no idea … maybe an openssl-problem ?

a cron-job or timed-reboot-script would be nice as a workaround, but my scripting skills are very “misunderestimated” 🙂

[imported_fulvio]

Try to type the following comand from the console shell:

echo “( sleep 86400 ; reboot ) &” >> /etc/rc.local

and the reboot the system.
It will automatically reboot Zeroshell everyday.

Regards
Fulvio

[Author]

Posts