I was having a problem getting nodogsplash started with Kamikaze r13193.
Note, I don't really know what I am doing, but that has never stopped me :-)
Two problems in /etc/init.d/nodogsplash.init, one after the other:
1) The first was this test in start():
if $0 test_module && $WD_DIR/nodogsplash $OPTIONS; then
This resulted in the error:
alix : Starting nodogsplash ...
alix : /etc/rc.common: .: line 74: test_module: not found
alix : FAILED: nodogsplash exited with non 0 status
I fixed that with:
if test_module && $WD_DIR/nodogsplash $OPTIONS; then
This allowed me to reach the next error. nodogsplash start failed randomly, about 4 out of 5 times, with log errors like:
(fw_iptables.c:86) Nonzero exit status 100 from command: iptables -t mangle -A ndsOUT -s 192.168.58.249 \
-m mac --mac-source 00:09:6B:86:02:24 -j MARK --set-mark 0x100
(and no, I don't know what status 100 means - it seems to come from an enum in the iptables
source code, which seems an odd way to reproduceably number errors to inexperienced me -
I expect other people will see different error status numbers).
The errors in my log file were different every time, and iptables -L did not show any of the chains set
up. When it worked, I saw four "ndsXXX" chains.
Flailing randomly, I added these lines before the if statement in /etc/nodogsplash/nodogsplash.conf:
### added KHL
iptables -t nat -N ndsOUT
iptables -t filter -N ndsRTR
iptables -t filter -N ndsNET
iptables -t filter -N ndsAUT
iptables -t mangle -N ndsTRU
###----
if test_module && $WD_DIR/nodogsplash $OPTIONS; then
That seems to start it up every time. I'm guessing that iptables needs more time to set up those chains
than iptables_do_command() in fw_iptables.c in nodogsplash is allowing. But then, when I fool around
with software I guess a lot.
Any ideas of what I should /really/ be doing? Is there some way that a clue-challenged semi-newbie
should be reporting this to the folks that care?
Keith