etch installer hanging on packages
Hi all, We've been using preseeds with etch since December for testing. Recently, something has broken these installs. I suspect the recent key switch on the repositories. The installer goes through all steps and does everything correctly, getting network, grabbing preseed.cfg, partitioning, etc. The problem comes at Select and install software progress bar -- it halts at 5%. Syslog on console 4 says that apt is trying to grab some packages (nice long list -- includes at, bc, exim4*, and python, just to name a few). Then it gives the following: Mar 2 09:15:07 in-target: WARNING: UNTRUSTED VERSIONS OF THE FOLLOWING PACKAGES WILL BE INSTALLED! Mar 2 09:15:07 in-target: Mar 2 09:15:07 in-target: Untrusted packages could compromise your system's security Mar 2 09:15:07 in-target: You should only preceed with the installation if you are certain that Mar 2 09:15:07 in-target: this is what you want to do. Mar 2 09:15:07 in-target: Mar 2 09:15:07 in-target: apt-utils coreutils apt libldap2 Mar 2 09:15:07 in-target: Do you want to ignore this warning and proceed anyway? Mar 2 09:15:07 in-target: To continue, enter Yes; to abort, enter No: So, there's a prompt... that I can't put Yes into b/c it's stdin is somewhere in installer-land. Three days ago, the list of packages was just libldap2 -- apt-utils, coreutils, and apt are new this morning. This happened with the etch RC1 netinst media, the 20070227-1 nightly and last weeks weekly. The dailys and weeklys I'm using are from here: http://www.debian.org/devel/debian-installer/ Anyone know what's going on? Any ideas or clues would be helpful. Thanks, Kevin - Kevin Sumner [EMAIL PROTECTED] (919) 962-6494 Assistant Unix Administrator Physics and Astronomy Networking Infrastructure and Computing University of North Carolina at Chapel Hill -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: SMBCLIENT PROBLEM
On Mon, Mar 13, 2000 at 01:34:40PM +0100, Vicente Torres wrote: I use smbmount-2.1.x to access Windows NT file systems from other computers in the net. Sometimes the connection gets lost and I get the following message when I try to list those directories: bash-2.02$ ls /net/XXX ls: /net/XXX: Input/output error bash-2.02$ If I try to mount again the net directories, smbmount-2.1.x says: Could not resolve mount point and the only solution I know is to restart my potato. After restart the connections are available again. How can I restore connections without restarting? As root you can # umount /net/XXX then smbmount will work fine for the next mount. As someone else has noted, the potato versions work much better (though I still get this problem occasionally). Kevin -- Dr Kevin Scott Philips Corporate Intellectual Property Cross Oak Lane, Redhill Tel: +44 1293 815281 Surrey RH1 5HA Fax: +44 1293 815060 UKE-mail: [EMAIL PROTECTED]
Re: UK Freeserve Connection
I've added some comments below from what I remember - I got a connection to freeserve working fine using pppconfig. On Thu, Dec 23, 1999 at 09:27:37PM +, Mike Norris wrote: O.K. still no joy obviously I'm doing something wrong still ... please comment on the following the pppconfig generated ... __ hidepassword noauth connect /usr/sbin/chat/ -v -f /etc/chatscripts/freeserve debug /dev/ttyS1 115200 defaultroute noipdefault user 'island-phase.fsnet.co.uk' I found that quotes around the user name cause pon to fail. remotename freeserve ippram freeserve usepeerdns ___ the debug gives this in /var/log/ppp.log pppd started by root, uid0 Using interface ppp0 Connect: ppp0 -- /dev/ttyS1 Remote message: request denied PAP authentication failed I used normal chatscript, ie no PAP/CHAP authentication. I've seen comments that PAP is claimed to be supported but doesn't work. The default prompts generated by pppconfig for the chatscript (e.g. ogin:, ssword:) seemed to work OK. LCP terminated by peer Connection terminated tcflush failed: Invalid argument Exit __ my /etc/chatscripts/freeserve is __ # ispauth PAP Don't use PAP (see above). # abortstring ABORT BUSY ABORT 'NO CARRIER' ABORT VOICE ABORT 'NO DIALTONE' ABORT 'NO DIALTONE' ABORT 'NO ANASWER' # modeminit ' ' ATZ # ispnumber OK-AT-OK ATDT08450796699 # ispconnect CONNECT \d\c # prelogin #ispname island-phase.fsnet.co.uk # isppassword my correct password in this line # postlogin __ Hope the above helps. Kevin -- Dr Kevin Scott Philips Corporate Intellectual Property Cross Oak Lane, Redhill Tel: +44 1293 815281 Surrey RH1 5HA Fax: +44 1293 815060 UKE-mail: [EMAIL PROTECTED]
Re: Using Computer Modern PS fonts with LaTeX?
On Mon, Dec 06, 1999 at 05:05:39PM -0800, Ron Hale-Evans wrote: That never happened to me, and I have not heard from anyone seeing this. It just uses TeX inside... Hmmm. Try pdflatex on http://www.nada.kth.se/~nv89-nun/offloading/mlg.tex.Z. Produces weird output, nothing like what TeX produces. Ron Seems normal to me, using pdflatex included in tetex 1.0-5 from potato. Kevin -- Dr Kevin Scott Philips Corporate Intellectual Property Cross Oak Lane, Redhill Tel: +44 1293 815281 Surrey RH1 5HA Fax: +44 1293 815060 UKE-mail: [EMAIL PROTECTED]
Problems with smbmount
I'm using smbmount-2.1.x (from smbfsx 2.0.3-1) in a slink system with kernel 2.2.9 and some potato stuff. (smbfs is compiled as a module, and CONFIG_SMB_WIN95 is not set.) I use it to mount shares from our NT domain, which generally works fine. However, I have a couple of problems: 1. Sometimes the connection gets dropped, so the mount point appears empty, the connection is still present according to df but cannot be unmounted. Re-mounting works but gives a duplicate entry in df output. A smbumount after this remounting removes all entries relating to that connection. 2. Very occasionally (like yesterday) the connection goes bad. It no longer appears in df output and if I try to look at or move the mount point I get a message like ls: L: Input/output error (where L is the mount point). I have not figured out a way to correct this problem without rebooting, which should only be for new hardware or kernel versions! Anyone else have these problems? Any solutions or pointers, especially regarding the second problem, greatly appreciated. Kevin
More hard disk woes
I have a system with a WD 2GB IDE drive that has started acting flaky - when I tried to upgrade from bo I got a thoroughly corrupted filesystem (files and directories becoming block special files that could not be deleted, lots of filesystem errors), so I re-installed hamm from scratch with a re-format of the disk. Unfortunately I have now started to lose files and directories again. Running e2fsck produced a lot of files in /lost+found. I would just go and replace the disk drive were it not for the fact that during my installation of hamm from CD-ROM the installation paused a number of times giving errors from the CD-ROM drive: IRQ timeout Status 0xd0 Although the installation seemed to work OK initially, this makes me wonder if I have a flaky IDE controller on the motherboard (in which case a new PC may be the better option given its age - Pentium 100 vintage). Does anyone have any idea whether the above points to controller or disk failure, or know how I could determine which was at fault? Thanks for any pointers, Kevin -- Dr Kevin Scott Philips Electronics UK Limited Patents and Trade Marks Department Tel: +44 1293 815281 Cross Oak Lane, Redhill Fax: +44 1293 815060 Surrey RH1 5HA, UK E-mail: [EMAIL PROTECTED]
Re: Easy X exit manager....
.xtlist.list.dopey \
-text dopey \
-padx 11 \
-width 7 \
-command {exec /usr/local/bin/tkxterm dopey }
button .xtlist.list.grumpy \
-text grumpy \
-padx 11 \
-width 7 \
-command {exec /usr/local/bin/tkxterm grumpy }
button .xtlist.list.happy \
-text happy \
-padx 10 \
-width 7 \
-command {exec /usr/local/bin/tkxterm happy }
pack append .xtlist.list \
.xtlist.list.sleepy {left expand fill} \
.xtlist.list.sneezy {left expand fill} \
.xtlist.list.bashful {left expand fill} \
.xtlist.list.dopey {left expand fill} \
.xtlist.list.grumpy {left expand fill} \
.xtlist.list.happy {left expand fill}
pack .xtlist.list -side top -fill x -expand true
}
===
It references a script /usr/local/bin/tkxterm, which does the following:
1. Stops xdm, and waits for the X server to die
2. Starts up as an xterm connected to the required machine (using XDMCP)
3. Restarts xdm when the xterm session finishes.
This script should be quite general. It contains:
===
#!/bin/sh
#
# Script called from tkmgr. It kills xdm and runs a remote X session
# on a given machine. When this terminates, xdm is restarted.
#
if [ $# != 1 ]
then
echo Usage: $0 machine
else
/etc/init.d/xdm stop
while [ -f /tmp/.X0-lock ]
do
sleep 1
done
/usr/bin/X11/X -query $1 -once
/etc/init.d/xdm start
fi
===
Hope this is all useful!
Kevin
--
Dr Kevin Scott
Philips Electronics UK Limited
Patents and Trade Marks Department Tel: +44 1293 815281
Cross Oak Lane, Redhill Fax: +44 1293 815060
Surrey RH1 5HA, UK E-mail: [EMAIL PROTECTED]
More hard disk woes (repeat)
(resent as smail configuration was dodgy - apologies if it appears twice) I have a system with a WD 2GB IDE drive that has started acting flaky - when I tried to upgrade from bo I got a thoroughly corrupted filesystem (files and directories becoming block special files that could not be deleted, lots of filesystem errors), so I re-installed hamm from scratch with a re-format of the disk. Unfortunately I have now started to lose files and directories again. Running e2fsck produced a lot of files in /lost+found. I would just go and replace the disk drive were it not for the fact that during my installation of hamm from CD-ROM the installation paused a number of times giving errors from the CD-ROM drive: IRQ timeout Status 0xd0 Although the installation seemed to work OK initially, this makes me wonder if I have a flaky IDE controller on the motherboard (in which case a new PC may be the better option given its age - Pentium 100 vintage). Does anyone have any idea whether the above points to controller or disk failure, or know how I could determine which was at fault? Thanks for any pointers, Kevin -- Dr Kevin Scott Philips Electronics UK Limited Patents and Trade Marks Department Tel: +44 1293 815281 Cross Oak Lane, Redhill Fax: +44 1293 815060 Surrey RH1 5HA, UK E-mail: [EMAIL PROTECTED]
Problems with 2.0 upgrade (hardware?)
I've just got the Cheap*bytes Debian 4 CD set (from the Linux emporium http://www.polo.demon.co.uk/emporium.html in the UK - ordered Thurs am, received Fri am!). Yesterday tried to upgrade my bo system. I ran into the problems that others have found with the cd_autoup.sh script, as well as the lack of a stable symbolic link on the cd, but managed to solve these (less elegantly than the posted solutions...). Well into the remainder of the upgrade I started getting e2fs error messages, and wound up losing about half of my home directory. This being Linux and Debian I'm inclined to assume hardware problems first, but I'm asking here in case anyone else has had the same problems. Although I ran script to log the upgrade it doesn't seem to log these error messages. What I now have is a rather full lost+found directory, from which I can salvage quite a bit, and a large number of files that have become block special files which I can do nothing to - can't rm them, and haven't managed to figure out how to turn them into a normal file in case there's something salvageable hidden there. Since I have these scattered around the file system too I suspect I need to do a wipe and reinstall, although the system is acting OK at the moment - I don't think all files installed properly as some files that should have been replaced had turned into block special files and therefore couldn't be overwritten. If anyone has come across these problems I'd like to know, especially if they have some solutions other than a complete re-install. I'd also like to know how to do something to those block special files! Thanks for any pointers, Kevin Dr Kevin Scott Philips Electronics UK Limited Patents and Trade Marks Department Tel: +44 1293 815281 Cross Oak Lane, Redhill Fax: +44 1293 815060 Surrey RH1 5HA, UK E-mail: [EMAIL PROTECTED]
Failing to print using lpr and nprint
I'm trying to set up lpr to print to a remote netware printer. Following the instructions in /usr/doc/HOWTO/Printing-HOWTO.gz I added the following entry to /etc/printcap: jeremy:\ :lp=/dev/null:sh:\ :sd=/var/spool/lpd/jeremy:\ :if=/var/spool/lpd/nprint-jeremy:\ :lf=/var/log/lp-errs: and created a script file /var/spool/lpd/nprint-jeremy (which is executable) simply containing: #! /bin/sh /usr/bin/nprint -S novell_fs_1 -n -q JEREMY - The nprint command works on its own, but when I try to print using lpr an infinite loop results. Messages appearing in /var/log/messages are repeats of: Feb 5 12:40:57 e310-pc1 lpd[2854]: restarting jeremy Feb 5 12:40:57 e310-pc1 lpd[2854]: jeremy: lock: Bad file number Feb 5 12:40:57 e310-pc1 lpd[2854]: jeremy: lock: Bad file number Feb 5 12:40:57 e310-pc1 lpd[2854]: restarting jeremy Feb 5 12:40:57 e310-pc1 lpd[2854]: jeremy: lock: Bad file number Feb 5 12:40:57 e310-pc1 lpd[2854]: jeremy: lock: Bad file number attaching strace to the appropriate pid (2854 here) shows that the problem seems to be related to trying to open the file errsa02854 (name generated by mktemp from errsXX) which does not exist. If instead of nprint in the script above I have something like cat /tmp/my_print_job the file is generated, which suggests nothing fundamentally wrong with my setup. I am running bo (with some libc6 stuff), 2.0.30, libc5 5.4.38-1, lpr 5.9-20. Any hints gratefully received! Thanks, Kevin -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: R.I.P
John, Did you by any chance defragment the win95 partition? We have a machine where doing this seems to upset the linux partition - best guess was that the win95 defragmenter had a record of the disk partitioning, but hadn't noticed that FIPS had shrunk the linux partition. We couldn't quite believe that it could be designed this way, but you never know - it would take too much time to track it down more precisely! If anyone else has had this occur it would be interesting. Kevin Dr Kevin Scott Philips Electronics UK Limited Patents and Trade Marks Department Tel: +44 1293 815281 Cross Oak Lane, Redhill Fax: +44 1293 815060 Surrey RH1 5HA, UK E-mail: [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
man gives segmentation fault
I am running Debian 1.3.1, and recently trying to run man gives a segmentation fault when run as a normal user (but not when run as root): % man man Segmentation fault xman and tkman work fine. I was fairly sure I hadn't played with any relevant settings - the executable has setuid bit set as I would expect: % ls -l /usr/bin/man -rwsr-xr-x 1 man root71204 May 21 10:40 /usr/bin/man testing man with strace also fails as a normal user: % strace man execve(/usr/bin/man, [man], [/* 29 vars */]) = 0 strace: exec: Operation not permitted I am rather puzzled, so any ideas gratefully received! Thanks, Kevin Dr Kevin Scott Philips Electronics UK Limited Patents and Trade Marks Department Tel: +44 1293 815281 Cross Oak Lane, Redhill Fax: +44 1293 815060 Surrey RH1 5HA, UK E-mail: [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
tk script to reboot etc from xdm login screen
\
-padx 11 \
-width 7 \
-command {exec /usr/local/bin/tkxterm dopey }
button .xtlist.list.grumpy \
-text grumpy \
-padx 11 \
-width 7 \
-command {exec /usr/local/bin/tkxterm grumpy }
button .xtlist.list.happy \
-text happy \
-padx 10 \
-width 7 \
-command {exec /usr/local/bin/tkxterm happy }
pack append .xtlist.list \
.xtlist.list.sleepy {left expand fill} \
.xtlist.list.sneezy {left expand fill} \
.xtlist.list.bashful {left expand fill} \
.xtlist.list.dopey {left expand fill} \
.xtlist.list.grumpy {left expand fill} \
.xtlist.list.happy {left expand fill}
pack .xtlist.list -side top -fill x -expand true
}
===
It references a script /usr/local/bin/tkxterm, which does the following:
1. Stops xdm, and waits for the X server to die
2. Starts up as an xterm connected to the required machine (using XDMCP)
3. Restarts xdm when the xterm session finishes.
This script should be quite general. It contains:
===
#!/bin/sh
#
# Script called from tkmgr. It kills xdm and runs a remote X session
# on a given machine. When this terminates, xdm is restarted.
#
if [ $# != 1 ]
then
echo Usage: $0 machine
else
/etc/init.d/xdm stop
while [ -f /tmp/.X0-lock ]
do
sleep 1
done
/usr/bin/X11/X -query $1 -once
/etc/init.d/xdm start
fi
===
Hope this is all useful!
Kevin
Dr Kevin Scott
Cordless Communications Group
Philips Research LaboratoriesTel: +44 1293 815281
Cross Oak Lane, Redhill Fax: +44 1293 815500
Surrey RH1 5HA, UK E-mail: [EMAIL PROTECTED]
Re: turning off computer
If you're using xdm, it is quite easy to add buttons to the login screen to enable the machine to be rebooted or shut down - a sample Tk script was posted to the list earlier this year, and we are currently using a modified version here. I can post more details if there is any interest... Kevin Dr Kevin Scott Cordless Communications Group Philips Research LaboratoriesTel: +44 1293 815281 Cross Oak Lane, Redhill Fax: +44 1293 815500 Surrey RH1 5HA, UK E-mail: [EMAIL PROTECTED]
Colour inkjet printers
I am currently considering the purchase of a colour inkjet printer, for text and graphics printing (including images from Photo-CD). I would be interested in experiences people have had in using them with Debian - the models I am currently interested in are the Epson Stylus 500 and Stylus Pro and the HP 820CXi and 870CXi. I presume the main issue is how well Ghostscript manages to drive the printers, and whether all the driver options can be accessed without running under Windows. Thanks for any comments, Kevin Dr Kevin Scott Cordless Communications Group Philips Research LaboratoriesTel: +44 1293 815281 Cross Oak Lane, Redhill Fax: +44 1293 815000 Surrey RH1 5HA, UK E-mail: [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
kernel-image-2.0.23_0 in Debian 1.1.14
I have just upgraded one machine to 1.1.14 and the 2.0.23 kernel, and am having some problems with the networking. In particular, running xdm on a remote server (using X -query host) seems to freeze, whereas it works OK when I boot up with the previous kernel version (2.0.6). Is this a problem that anyone else has encountered, or do I need to dig deeper myself? Thanks for any pointers, Kevin Dr Kevin Scott Cordless Communications Group Philips Research LaboratoriesTel: +44 1293 815281 Cross Oak Lane, Redhill Fax: +44 1293 815000 Surrey RH1 5HA, UK E-mail: [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
