Open GoogleCodeExporter opened 8 years ago
Hello, this is a Fedora issue, so the bug should not be opened here but instead
in bugzilla.
Shellinabox in Fedora runs from systemd:
[root@3zpc0560 ~]# systemctl status shellinaboxd.service
shellinaboxd.service - Shell In A Box daemon
Loaded: loaded (/usr/lib/systemd/system/shellinaboxd.service; disabled)
Active: active (running) since Wed, 09 Jan 2013 11:21:36 +0100; 3s ago
Docs: man:shellinaboxd(1)
Main PID: 6351 (shellinaboxd)
CGroup: name=systemd:/system/shellinaboxd.service
├ 6351 /usr/sbin/shellinaboxd -u shellinabox -g shellinabox
--cert=/var/lib/shellinabox --port=4200 --disable-ssl-menu -s /:LOGIN
└ 6352 /usr/sbin/shellinaboxd -u shellinabox -g shellinabox
--cert=/var/lib/shellinabox --port=4200 --disable-ssl-menu -s /:LOGIN
[root@3zpc0560 ~]# ls -al /var/lib/shellinabox/
total 12
drwxr-x---. 2 shellinabox shellinabox 4096 Jan 9 11:21 .
drwxr-xr-x. 61 root root 4096 Jan 9 11:21 ..
-rw-------. 1 shellinabox shellinabox 2851 Jan 9 11:21 certificate.pem
If you want to run the daemon from another directory and not through systemd,
you have to pass the "--cert=" parameter to it to point to the certificate file.
"/sbin/shellinaboxd" is not a script but a binary, so there's no way to put a
"cd /tmp" anywhere.
Regards,
--Simone
Original comment by negativ...@gmail.com
on 9 Jan 2013 at 10:30
Hi!
This was an end-user issue - I had no idea systemctl knew about shellinaboxd.
Maybe a little documentation on the package would help - possibly as simple as
adding "This service is managed via systemctl" to the package description.
Thanks for your help - please consider this matter closed.
Original comment by reiber
on 9 Jan 2013 at 9:20
changing dir to tmp the shellinabox worked. But is this unsafe if other users
can peek into /tmp? Maybe shellinabox can create a folder in /tmp, apply secure
permissions, then create the certificate?
Original comment by TruSktr
on 16 Sep 2013 at 1:57
Running "systemctl start shellinaboxd.service" runs, but trying to connect to
it with https results in "(Error code: ssl_error_rx_record_too_long)".
Visiting http://localhost:4200 (no ssl) results in "Session closed.". I can
press the "Connect" button multiple times until finally it connects. It'll
refresh the page each time showing the "Session closed." message before finally
connecting after a random number of clicks on the Connect button.
Original comment by TruSktr
on 16 Sep 2013 at 2:05
[root@CENTOS7 lib]# systemctl status shellinaboxd.service ● shellinaboxd.service - Shell In A Box daemon Loaded: loaded (/usr/lib/systemd/system/shellinaboxd.service; enabled; vendor preset: disabled) Active: failed (Result: start-limit) since Fri 2016-05-20 03:07:04 IST; 2s ago Docs: man:shellinaboxd(1) Process: 1587 ExecStart=/usr/sbin/shellinaboxd -u $USER -g $GROUP --cert=${CERTDIR} --port=${PORT} $OPTS (code=exited, status=1/FAILURE) Main PID: 1587 (code=exited, status=1/FAILURE)
May 20 03:07:04 CENTOS7 systemd[1]: shellinaboxd.service: main process exited, code=exited, status=1/FAILURE May 20 03:07:04 CENTOS7 systemd[1]: Unit shellinaboxd.service entered failed state. May 20 03:07:04 CENTOS7 systemd[1]: shellinaboxd.service failed. May 20 03:07:04 CENTOS7 systemd[1]: shellinaboxd.service holdoff time over, scheduling restart. May 20 03:07:04 CENTOS7 systemd[1]: start request repeated too quickly for shellinaboxd.service May 20 03:07:04 CENTOS7 systemd[1]: Failed to start Shell In A Box daemon. May 20 03:07:04 CENTOS7 systemd[1]: Unit shellinaboxd.service entered failed state. May 20 03:07:04 CENTOS7 systemd[1]: shellinaboxd.service failed.
what about this pls sggest
Original issue reported on code.google.com by
reiber
on 31 Dec 2012 at 1:42