rhaynes

Hi, i have used drakbackup extensively with 2006 PP, now with 2007 PP I have some odd behaviour. After I configure drakbackup and run an initial backup the config file gets overwritten with a default file. Since I have a cron job to run drakbackup this is most annoying.

Anyone have a fix?

Thanks,

R Haynes

Hi, I now have mandriva 2006 powerpack and mandrake 10.0 official installed on one machine. 10.0 is still my work machine until I get 2006 configured properly.

I have a home partition on my 2006 "side". While in 2006 I mount my 10.0 /home partition to /mnt/home and I want to copy all the files (except the hidden files) to my 2006 /home/username partition.

As root I mount the old partition and then copy * to /home/xyz.

As I found out then all the files and directories (some 7GB) are now controlled by root.

I then tried to tar all the files and move the tar file and untar. This worked but now all the files are owned by "501".

How can I copy everything and keep ownership and group info from the old partition?

Thanks,

R Haynes

Hi, here is the situation:

I originally had mandrake 10.0 installed on several partitions with /

(and hence /boot) was on /dev/hde6 with other partitions on

/dev/hde5--9. When I installed

mandrake 2006 in place of winxp the installer indicated that partitions

were being renumbered 5 -->6, 6-->7 etc... and the 2006 installer

called the partitions /dev/sda(n) instead of /dev/hde(n). I left all

of the mandrake partitions alone --- except for the renaming by the

2006 installer/partitioning tool.

The installed went fine, I installed a bootloader from the 2006 in the

master boot record.

I am able to boot 2006 fine. Now I am trying to configure lilo to

boot the old mandrake distribution. The question is what changes do

I have to make to the lilo.conf file (from 2006) to boot the old

distribution. Do I have make changes to the fstab file (from the 10.0

distribution)? Should I change references to partitions on the

mandrake 10.0 side to refer to /dev/sda(n) instead of /dev/hde(n)? What

about the lilo.conf file (in 2006)? When I add a boot entry for

mandrake 10.0 should i say root=/dev/hde7 or /dev/sda7? Should I copy

the kernel images (from 10.0) over to the /boot on the 2006 side? Or

can I somehow refer to the kernel images on the /dev/hde7 (sda7)

partition?

This has me quite confused ---as you can tell from the message above.

Any help would be greatly appreciated.

Sincerely,

R Haynes

Hi, thanks for your replies. So Mandriva will handle the renumbering fine, but will I have to boot 10.0 with the rescue disk (passing root=/dev/newslash --- since the parition number has changed) and then edit lilo.conf and fstab? It won't do this automatically will it?

Thanks,

R Haynes

Hi, I currently have a dual boot win xp and mandrake 10.0 official

distribution. Win xp is on the first partition.

What I would like to do is to delete the win xp partition and create

partitions (using that space) for testing of mandrake 2006.

The 2006 installer easily allows me to delete the win xp partition.

But when I begin to add partitions for 2006 it gives me a message saying

that partitions will be renumbered.

My plan was to install 2006 but not the bootloader (I use lilo in mandrake

10.0). Then I would just boot to mandrake 10.0 and edit the lilo.conf

file to add an extry for the 2006 boot.

But if the partitions get renumbered during the 2006 install will I

be able to boot to the old distribution?

What is the best way to go about this?

Thanks,

R Haynes

Hi, just wondering if anyone has xen working with mandrake 10.0 ?? If so would you mind

posting any problems you had with the install and confiig??

Thanks,

R Haynes

Hi, As detailed in the hardware section of this board I have been having intermittant

problems with a usb printer (hp laserjet 1300). I am running mandrake 10.0 official

with the stock 2.6.3-7smp kernel. The printer installed correctly and ran after the initial install. It has died a few times since.

Printerdrake recognizes the printer on /usb/lp1 it identifies the correct model and

recommends a driver. But the test pages or nothing else prints.

The first hint of something mysterious happens when I go into the cups www admin

page and the message says trying to connect to printer (retrying in 30 seconds) on

/dev/usb/lp0. Somehow the device number has change. Trying to change it back

to lp1 in the cups admin page doesn't seem to work. Jobs head to the print Q and

stay there.

A google search indicates a possible usb problem when multiple (in my case two)

usb devices are present. Something to do with device addressing.

My question is if anyone knows which kernel version (if any) addresses this issue.

And if so ... will it work without breaking anything else on Mandrake 10.0 official?

What is the highest kernel version one should try to use on 10.0?

Thanks,

R Haynes

Hi, thanks for the reply... if I ssh in and then su to root and run the job, logout out as

root, log out as normal user and then log back in the job is still running!!!

Any thoughts?

R Haynes

Hi, I have a f77 program that I want to run. I log into my workstation from home

via ssh and run the executable by

./a.out > runspecs &

the job then begins to run but when I log out and log back in the job has died (yes

I have tried running the job at the terminal and all is fine...).

the file redirection captures output that is normally printed to the screen.

Any ideas?

I am running mandrake 10.0 official kernel 2.6.3-7

Thanks,

R Haynes

AFter digging a little more some people seem to think that this is a more general

kernel issues and the way the kernel assigns addresses to devices.

Anyone have a fix or kernel patch for this?

R Haynes

Hi, thanks for the link. I will try when I get to work on Monday. Why isn't this configured in Mandrake by default with printerdrake.??? Why does it work sometimes and not others?

Again thanks for you help.

R Haynes

Hi, running the linux hplip install from HP and then installing the printer from the cups www admin page did the trick.

Thanks,

R Haynes

----------------------------------------------

Hi, I am running mandrake 10.0 official. I have a hp 1300 laserjet printing

which installed and worked fine until recently until I tried to get a local

network printer working.

I deleted the printer definition from

/etc/cups/printers.conf and used printerdrake to reinstall.

I searched for printers and it found the hp 1300 printer on /dev/usb/lp1 (it is a usb printer) and I chose the recommended driver (one of the choices under

the laserjet 1300 list). The test files would not print.

I exited the printerdrake and tried to print a file. Nothing happened.

lpq says the jobs are in the Q.

If I run the cups www admin tool it says the printer is not connected, "retrying in 30 seconds...) I have tried to manually change the device

to the various usb devices without any luck. I changed the Loglevel

in the cups.conf file to debug2 and the error_log file again said printer is

not connected.

I do remember seeing the device listed on the cups www admin tool when the printer was working as something like /dev/usb/hp%%%1300%%% or something like that. I certainly didn't enter

this device.

I tried to delete cups ( to see if that is the problem) but then a lpq returns a "can not connect

to server" message (or something like that). If I delete cups what do I need to get printing working?

Does anyone have any suggestions?

Thanks,

R Haynes

when using -l you don't need to include the lib in the libname:

 

f77 -O -I/usr/lib -pg -lliblapack -o drive driver.o odesolve.o \

 

should become:

 

f77 -O -I/usr/lib -pg -llapack -o drive driver.o odesolve.o \

<{POST_SNAPBACK}>

Hi, thanks for your reply... unfortunately that doesn't work either.... is there anything I have to do to make the system "aware" that lapack exists??

R Haynes

Hi, this problem has now been solved by also installing lapack3-devel

Thanks,

R Haynes

Hi, I am trying to link lapack with some fortran code. I have done urpmi liblapack3 and the docs, everything installed fine. When I run my make file I get the following message:

[rhaynes@rhaynes gaussian]$ make

f77 -O -I/usr/lib -pg -lliblapack -o drive driver.o odesolve.o \

mynewton_nor.o xfdjac.o xfdjac1.o

/usr/bin/ld: cannot find -lliblapack

collect2: ld returned 1 exit status

make: *** [drive] Error 1

I am assuming it can't find liblapack. Here is my makefile

drive: driver.o odesolve.o mynewton_nor.o xfdjac.o xfdjac1.o

f77 -O -I/usr/lib -pg -lliblapack -o drive driver.o odesolve.o \

mynewton_nor.o xfdjac.o xfdjac1.o

driver.o: driver.f

f77 -c -O -pg driver.f

odesolve.o: odesolve.f

f77 -c -O -pg odesolve.f

mynewton_nor.o: mynewton_nor.f

f77 -c -O -pg mynewton_nor.f

xfdjac.o: xfdjac.f

f77 -c -O -pg xfdjac.f

xfdjac1.o: xfdjac1.f

f77 -c -O -pg xfdjac1.f

This worked on a old redhat system. Any suggestions?

Thanks,

R Haynes

Hi & happy new year to all,

After some recent power outages and associated problems with my server at work I am about to order a APC UPS system. One which came recommended was the

BX1000 model.

Does anyone here have any experience getting this to work with linux? What will I need to do? Does the software for linux work out of the box? Do I need some other software to make this work?

Any comments or suggestions would be greatly appreciated.

Thanks,

R Haynes

Hi, as mentioned in my previous post I have to try and do some system recovery on a machine that experienced a hard shutdown due to a power outage.

It seems that / (my /dev/hde6) has been messed up. Assuming the other partitions can be mounted is it possible to run the installation disk and just have the stuff from / reinstalled.

My other partitions (which hopefully are intact) are swap,home,usr,var

I would like to have /var,/home left intact (/var contains web server and database stuff...)

I have a backup but of course now I can think of several things which weren't backed up (for example a relatively smalled database whose directory I forgot to add to the drakbackup list...).

Thanks,

R Haynes

Thanks for the reply... I am pretty sure I used ext3 filesystems... and I just ran the command that was indicated by the system when the automatic check failed, it said to run fsck manually without the -a or -p options. Hopefully it would detect the filesystem type and indicate the corresponding command to run.

Thanks,

R Haynes

Hi, after a recent power outage my mandrake 10.0 official system did not come up nicely. It wanted to do a system check on /dev/hde6 (which is my / partition). It then complained and said that I would have to run fsck manually on the partition. From the rescue prompt I did:

fsck /dev/hde6

and I answered yes to all the fixes it suggested, some stuff was copied to /lost+found.

I then rebooted but got a kernel panic it suggested that I boot with init=

I went through all of this during a storm in November and ended up having to do a reinstall. Before I make things worse I am looking for some advice.

What should I do next? Should run CD1 and go into rescue mode .... and try to mount the partitions? Should I run fsck on all the partitions? The last time this happened I wasn't able to mount the partitions (it complained about superblocks...).

The machine is at work and I hope to get in tomorrow to begin to salvage.

All my filesystems are ext2 or ext3.

Any suggestions as to how to proceed? How about filesystem choice to avoid this in the future. I have had redhat systems in the past which seem much more robust when faced with a hard shutdown.

Thanks,

R Haynes

[moved from Installing Mandrake by spinynorman]

Hi, how do you know if the Cd's were being read by a DVD reader?? My CD/DVD drive (LG super mutli) may be doing this too. I was able to install 10.0 no problem but 10.1 would not work with some ramdisk error.

Thanks,

R Haynes

Hi, I would like to install mandrake 10.1 along side 10.0 for testing purposes.

I would like to be able to choose which distribution I boot too. The only partition I need to share would be /home. Is this difficult to set up with the existing install program? Any issues to be aware of?

Thanks,

R Haynes