This release incorporate a new technology which allows a better integration with current linux distribution. This is the first release of Linuxconf with this technology. We are anticipating few adjustments. This technology change should have NO effect on current Linuxconf installation. Tester may upgrade without fear.
For a long time, we had no clue about how to make Linuxconf compatible with this technology, while preserving its added values. Now we know. After much exchanges with different linuxconf's users (especially Debian users), we have come up with a solution.
This solution has many impacts:
|/etc/inittab||Had to be replaced||No changes needed|
|first boot script
|/sbin/askrunlevel is appended||Same as before|
|Master boot script
|Totally replaced by a rc.M||Slightly modified|
Linuxconf picks the proper directory based on one optional line in /etc/conf.linuxconf. It goes like
#sysv Linuxconf network network nfsfs mountnet cron crond syslog syslog portmap portmap inet inetd lpd.init lpd named.init named amd amd sendmail sendmail nfs nfs keytable keytable
When Linuxconf is executing a scripts in the proper run-level directory, if it finds that there is a built-in equivalent or a dropin with the same name, it will use the built-in strategy or the dropin instead of the init script.
I am supplying some diff to help people test this new stuff on their system. These diff are done against RedHat 4.1 but are so simple, you should be able to copy the idea by hand.
*** rc.d.old/rc.sysinit Sun Dec 22 12:10:37 1996 --- rc.d/rc.sysinit Wed Jun 11 14:47:13 1997 *************** *** 187,189 **** --- 187,194 ---- # Feed entropy into the entropy pool /etc/rc.d/init.d/random start + if [ -x /sbin/askrunlevel ] ; then + echo Executing /sbin/askrunlevel + echo + /sbin/askrunlevel + fi *** rc.d.old/rc Mon Sep 9 14:43:00 1996 --- rc.d/rc Wed Jun 11 15:03:48 1997 *************** *** 24,29 **** --- 24,33 ---- # Is there an rc directory for this new runlevel? if [ -d /etc/rc.d/rc$runlevel.d ]; then + if [ $runlevel != 6 -a $runlevel != 1 -a -x /bin/netconf ] ; then + /bin/netconf --bootrc /etc/rc.d/rc$runlevel.d + exit + fi # First, run the KILL scripts. for i in /etc/rc.d/rc$runlevel.d/K*; do # Check if the script is there.
For now, the dialog has been enhanced so you can at least enter the path of this file.
This solve a problem very few people have seen. This was happening when you were moving one host from one domain to another for example.
if netconf --connect config then do_something netconf --disconnect config fi
This works for manual and on demand configuration. A success exit code is also returned when the link is already UP
In the PPP dialout dialog, there is now two new entries allowing you to enter a command to execute when the link is up and a command to execute when the link is down.
In the chat section, there is a new optional field called trigger. This is used for chat to ISP offering shell account. After you get a prompt, you must issue a command which I called the trigger.
The lock strategy used could be fold into the official pop3d daemon in no time. Please email me.
vpop3d has also a nice little feature. It clearly identify the domain it is serving in its greeting. The most common errors when setting up a virtual email domain are generally related to improper DNS setup and particularly improper reverse mapping. The problem is enhanced because of the lack of proper testing tools. An email client is a very bad testing tool as it reports too little information. There is one nice test tool which will show clearly what is going on: It is telnet. Here is an example of its use in this case. Lets create a virtual domain called toto.com. We have defined the domain in our DNS. We have defined mail.toto.com as the virtual server with proper reverse mapping. Then we do
$ telnet mail.toto.com pop-3 Trying 192.168.1.21... Connected to mail.toto.com. Escape character is '^]'. +OK Virtual toto.com POP3 Server (Version 1.004) ready. user jack +OK please send PASS command pass dummy +OK 0 messages ready for jack in /var/spool/vmail/toto.com//jack
We should see immediately the greeting of vpop3d announcing that it is performing for the toto.com domain. If this is not the case, we have something wrong with the DNS. If we are in (see proper greeting) and the user and pass command did succeed, all is fine and the normal POP client will be happy.
Complex routing had a bug when one wanted to redirect a complete domain to another domain. Some combination were generating improper sendmail.cf configuration.
It is possible to control the start user ID for vdomain. Linuxconf used to allocate them starting at 60,000. You can pick the one you want now. This was asked for people dealing with user quotas.
passwd (without argument) Change your own password passwd user_account Change interactivly this user account password passwd -h passwd --help Print this screen passwd -l user_account Lock the account passwd -P user_account Change the password from a pipe echo new_passwd | passwd -P user_account passwd -u user_account Unlock the account (Only available with shadow password)