Red Hat Bugzilla – Bug 18221
linuxconf-web does not work
Last modified: 2008-05-01 11:37:59 EDT
Just installed RH 7.0 and can't get linuxconf access via HTTP. Works
fine in text and in GUI mode.
I try to activate it via linuxconf text mode
-> Control Service Activity
linuxconf is set to "manual". I set it to "automatic" and that
registers. I try to "start" it next at the same spot and it returns
Executing: /etc/rc.d/init.d/linuxconf start
> Running Linuxconf hooks: [ OK ]
but when I hit the "accept" button and go back a level, linuxconf is
I tried getting linuxconf-web to run also, which was set at
"automatic". I tried changing it to "automatic" and I get a:
Executing: /sbin/chkconfig linuxconf-web: reset
* error reading information on service linuxconf-web:: No such
file or directory
* return 1
! Switch service linuxconf-web: to Automatic
I tried to "start" it next and I get a:
Executing: /etc/rc.d/init.d/linuxconf-web: start
* sh: /etc/rc.d/init.d/linuxconf-web:: No such file or
* return 127
I don't even see a '/etc/rc.d/init.d/linuxconf-web" file in the RH 7.0
try in /etc/xinet.d/linuxconf-web
to change the line "disable = yes" to "disable = no"
Changing the "/etc/xidentd.d/linuxweb-conf" doesn't do a thing.
I'm also having problems with linuxconf --http. I cannot get it to run on the
command line or in a browser. When I do a linuxconf --http --debug, I get a
segment fault(core dumped).
I also have the problem
I set disabled to no in /etc/xinetd.d/linuxconf-web
I use telnet localhost 98 to test
each time i type a line I'm disconnected unless I have another telnet open
If I use linuxconf --http from a command line
linuxconf automagically runs in debug mode and open a socket on port 8000, where
I can successfully connect
I've traced the problem to dialog/html.cc in html_get
the command ok=cmd->listen(timeout*60) never returns (this function is defined
by the way, the correct syntax is linuxconf --http --debug 0|1
it is also possible to use --name or --port
I've send a little patch to the maintainer, Jack Gelinas, so that it no longer
coredumps for this
trying to access linuxconf-web trough port 8000 doesn't work for me...
is suppossed to work in this broken release ?
amazing that nalin hasn't tryied linuxconf-web before the release...
to connect to port 8000 in debug mode, you have to start linuxconf --http in a
shell before (so you completely bypass xinetd)
you can do a netstat -a to show the open port
for the linuxconf-web problem, it seems that indeed xinetd is bugged
it should handle, as inetd does, the control socket to linuxconf but instead it
handles only the data part !
Oct 13 12:49:32 orpheus xinetd: START: linuxconf pid=7222
Oct 13 12:49:32 orpheus xinetd: EXIT: linuxconf signal=11 pid=7222
Unless I'm mistaken it appears linuxconf is segfaulting?
I added the following to the linuxconf-web config to generate that:
log_type = SYSLOG daemon
log_on_success += EXIT
We've fixed this in xinetd, coming as an errata soonish.
The updated xinetd does not fix the problem.
I still can't use linuxconf-web - when starting it from the command
line /sbin/linuxconf --http I only receive an "ending html".
the updated xinetd works for me :-)
I can connect to http://localhost:98
I've set disabled= no in /etc/xinetd.d/linuxconf-web
I'm using linuxconf 1.21r6 (compiled from source)
It ididn't work with the orginal RPM.
I updated to 1.21r1 and it works now
The updated xinetd package did not fix the problem.
Now I get:
Oct 23 11:38:53 orpheus xinetd: warning: can't get client address: Invalid
Oct 23 11:38:53 orpheus xinetd: refused connect from 0.0.0.0
Oct 23 11:38:53 orpheus xinetd: FAIL: linuxconf libwrap from=0.0.0.0
Oct 23 11:38:53 orpheus xinetd: file descriptor of service linuxconf has
Oct 23 11:38:53 orpheus xinetd: select reported EBADF but no bad file
descriptors were found
After that it just refuses connections from anywhere.
When I install linuxconf 1.21r8 (and devel) and Xinetd update,
Got a error message in /var/log/messages:
Oct 30 10:24:16 mail xinetd: file descriptor of service linuxconf has been
Oct 30 10:24:16 mail xinetd: select reported EBADF but no bad file descript
ors were found
After install linuxconf-lib-1.21r8-1.i386.rpm and linuxconf-util-1.21r8-
1.i386.rpm, it seems like working fine so far. Linuxconf-web was down every 15
to 20 minute, now it has been running for an hour, and no error at all.
Unfortunately when linuxconf dies (time-out or kill), xinetd occur EBADF and
after that it rejects any access to port 98.
I also tested this with linuxconf 1.22-1 and xinetd 126.96.36.199pre12-3.
Actually I use inetd from RH6.2 for linuxconf-web and xinetd for other services.
How do we reopen this? It is obviously not resolved?!
Nov 30 17:18:38 smail xinetd: warning: can't get client address: Invalid
Nov 30 17:18:38 smail xinetd: START: linuxconf pid=31684 from=0.0.0.0
Nov 30 17:18:39 smail xinetd: EXIT: linuxconf status=255 pid=31684
Nov 30 17:18:39 smail xinetd: file descriptor of service linuxconf has
Nov 30 17:18:39 smail xinetd: select reported EBADF but no bad file
descriptors were found
NOTE: I'm not calling it from 0.0.0.0?!
there's a fix for xinetd which solves the problem.
Still stuffed for me.
None of the above fixes solve any of the problems.
Still broken. Posts to Jacques Gelinas also give no Joy. Linuxconf will work
from the local machine but never as web interface. Linuxconf looks for a
linuxconf-web script when initializing the web interface from the command line
linuxconf. This script does not exist. When booted, RH7 initiales a listening
and waiting port 98. This port dies unceremoniously as soon as I attempt to
access it. I am running DHCP, Samba (SWAT does work fine), print server,
ipchains. Everything works fine except linuxconf-web. The error message when I
attempt to rouse linuxconf over the local net is:
Jan 4 10:05:25 grippo xinetd: warning: can't get client address: Invalid
Jan 4 10:05:25 grippo xinetd: refused connect from <no address>
Jan 4 10:05:25 grippo xinetd: linuxconf service was deactivated because
Jan 4 10:10:00 grippo CROND: (root) CMD ( /sbin/rmmod -as)
The 'invalid address' repeats 10 times before it terminates the service.
my netstat -nr is:
Kernel IP routing table
Destination Gateway Genmask Flags MSS Window irtt Iface
255.255.255.255 0.0.0.0 255.255.255.255 UH 0 0 0 eth0
192.168.1.0 0.0.0.0 255.255.255.0 U 0 0 0 eth0
188.8.131.52 0.0.0.0 255.255.252.0 U 0 0 0 eth1
127.0.0.0 0.0.0.0 255.0.0.0 U 0 0 0 lo
0.0.0.0 184.108.40.206 0.0.0.0 UG 0 0 0 eth1
I have hosts.deny set to all:all and hosts allow to ALL:
192.168.1.0/255.255.255.0. I have not set any explicit routes and have
installed RH7 clean with only using the firewall howto to get things running.
works perfectly for me :
RH7 + all errata
xinetd rpm with fix from linuxconf web site (see url above) because errata from
RH contains only a partial fix
you have to set disable= no
check that the server path is good (else you have a warning in
restart xinetd : /etc/init.d/xinetd restart
I use linuxconf 1.23
This is still not resolved. It works fine as long as you don't have an active
/etc/hosts.deny file, but when hosts.deny is active, the following appears
Jan 7 19:57:44 linux xinetd: warning: can't get client address: Invalid
Jan 7 19:57:44 linux xinetd: refused connect from <no address>
To fix the above error if you are using a hosts.deny/hosts.allow file, you can
to the hosts.allow file.
This makes it work.
Thanks for the hosts.allow comment. I knew I wasn't a total idiot. I feel this
should be fixed a little different as this may pose a security risk. It sounds
like it is an xinetd problem...
The update xnetd(xinetd-220.127.116.11pre11-1.i386) stil not work
When i start or reload xinetd "linuxconf-web"
the service stay up for 1 thought 3 minutes...and down after...
i have upgraded packages linuxconf and xinetd..
and trouble continues...
Go to http://www.solucorp.qc.ca/linuxconf and install the xinetd which has
the "wait" patch xinetd-18.104.22.168pre11-1wait.i386.rpm cause the xinetd from
RedHat doesn't fix it.
What is the secret combination of versions to get this to work. Im running
Linuxconf 1.25 - xinetd 22.214.171.124pre11-1wait and I still get the error
Executing: /etc/rc.d/init.d/linuxconf-web: start
* sh: /etc/rc.d/init.d/linuxconf-web:: No such file or directory
* return 127
Does anyone have this working??? Thanks
The bug in linuxconf-web was fixed only when i upgraded xinetd to
xinetd-126.96.36.199pre11-1.i386 and remove line disable = yes from linuxconf-web
and configure the linuxconf network access. After this restart xinetd.conf.
And it should be worked.
*** Bug 14986 has been marked as a duplicate of this bug. ***
*** Bug 33608 has been marked as a duplicate of this bug. ***
I upgraded to xinetd-188.8.131.52pre14, the linuxconf now start (via xinetd, the old
xinetd from RH 7.0 simply bomb out) but nothing more happened.
linuxconf --http --debug still gives Segmentation fault.
Manual connectin to port 98 by telnet I can see that linuxconf write back
nothing after a passed to it GET / HTTP/1.0 (connection was closed after
I tryed to upgrade to linuxconf-1.24r2-10. I had to update more packages
(mount, util-linux, pam, etc) and now linuxconf do not close the open
connection but gives nothing to browser. When trying WWW acces through WWW
client I can see that xinetd kill linuxconf after 48 sec but there is still
nothing usable on the screen.
My DNS and other services are Ok, I have pernament acces to the Internet and I
have no other troubles with anything.
*** This bug has been marked as a duplicate of 23560 ***