Re: RAID1 Question....
> Greetings, > I have a machine that I would like to institue RAID1 on. Two 4 Gig > drives, > the primary is partitioned with a > /dev/hda1 / > /dev/hda2 swap > /dev/hda3 /var > /dev/hda4 /usr > > I have the secondary drive partitioned the same way. I would like to mirror > the partitions on the primary drive to those on the secondary drive. I have > made an mdtab file, > > # mdtab entry for /dev/md3 > /dev/md3raid1,8k,1,e46c7dcc /dev/hda3 /dev/hdb3 > > but can't seem to get it working. Anyone got this kind of stuff running? > Any help would be appricated. > > Anthony > You cannot make make a raid1 partition from an already existing filesystem. What i've done : 1. install a minimal system on / (64 MB). 2. install the package raid-tools (you might need to create a /etc/raid/raid5.conf file to allow final setup) and a raid1-aware kernel. 3. reboot 4. mdcreate /dev/md0 /dev/hda3 /dev/hdb3 5. mdcreate /dev/md1 /dev/hda4 /dev/hdb4 6. rm /etc/raid/raid5.conf and create a /etc/raid/raid1.conf: # Sample raid-1 configuration # /dev/md0 : /tmp raiddev /dev/md0 raid-level 1 nr-raid-disks 2 nr-spare-disks 0 device /dev/hda3 raid-disk 0 device /dev/hdb3 raid-disk 1 # # /dev/md1 : /var raiddev /dev/md1 raid-level 1 nr-raid-disks 2 nr-spare-disks 0 device /dev/hda4 raid-disk 0 device /dev/hdb4 raid-disk 1 # end of raid1.conf 7. mkraid /etc/raid/raid1.conf 8. mdadd -ar 9. mke2fs /dev/md0 10. mke2fs /dev/md1 11. Move /var to /dev/md0: mount /dev/md0 /mnt cd /var ; tar cf - . |(cd /mnt ; tar xvpSf -) mv /var /var- umount /mnt ; mount /dev/md0 /mnt rm -rf /var 12. Move /usr 13. sync; sync ; sync ; reboot. You need another machine to build your custom kernel( 64 MB is not enough for gcc+kernel-source). Pierre Blanchet.
Re: FW: md driver problem
> > Hi, Peter, > > It doesn't seen to be a floppy problem but the md-driver seems to hang. > > Ciao > Andreas Loosen > I think it's not an md problem but rather the next driver to be loaded in the kernel that used to be WD-7000 SCSI card driver. It take sometimes a long time (2-3mins) before it realize there's no WD-7000. If waiting a few minutes does not help, try to put Linux reserve=0x350 in order to tell the driver not to probe this adress (this might be one of the following: 0x0300, 0x0308, 0x0310, 0x0318, 0x0320, 0x0328, 0x0330, 0x0338, 0x0340, 0x0348, 0x0350, 0x0358, 0x0360, 0x0368, 0x0370, 0x0378, 0x0380, 0x0388, 0x0390, 0x0398, 0x03a0, 0x03a8, 0x03b0, 0x03b8, 0x03c0, 0x03c8, 0x03d0, 0x03d8, 0x03e0, 0x03e8, 0x03f0, 0x03f8 , sorry don't know exactly what's the good one) Pierre Blanchet. > > > -Original Message- > > From: Peter Iannarelli [mailto:[EMAIL PROTECTED] > > Sent: Wednesday, October 28, 1998 11:28 AM > > To: Rodrigo Moya > > Cc: Debian User List > > Subject: Re: md driver problem > > > > > > hello Rodrigo: > > > > I've never seen problems with a floppy before. Thats a good one. > > > > Try booting a DOS floppy if you can to see what happens. > > > > Try resetting the CMOS to the manufacturers defaults. You may also > > try removing the two hard drives on your second IDE chain for the install. > > You can reconnect and initialize them once your up and running. > > > > First - check your physical connections. You never know, something may > > have shaken loose. > > > > Assuming the FD controller is on the motherboard, you may want to > > go to the manufacturer web site as see if there is a firmware update > > > > > > > > Peter. > > > > > > > > Rodrigo Moya wrote: > > > > > Hi all! > > > > > > I've been trying to install Linux (Suse, Debian, RedHat...) in > > a Pentium MMX > > > 200 with 3 hard disks and a cdrom, but with either > > distribution, the boot-up > > > process hangs up in the same place (never mind whether started > > from DOS boot > > > disk, Linux boot disk, cdrom...). This is the output I get: > > > > > > > > > hda: Maxtor 90432D3, 4121MB w/256 cache, CHS=8374/16/63 > > > hdb: SAMSUNG WINNER-3 WN32543A(2.5GB), 2423MB w/109kB cache, > > > CHS=4924/16/63 > > > hdc: ST3290A, 249MB w/64kB cache, CHS=1001/15/34 > > > hdd: MATSHITA CR-585, ATAPI CDROM drive > > > ide0 at 0x1f0-0x1ft,0x3f6 on irq14 > > > ide1 at 0x170-0x177, 0x376 on irq 15 > > > no cont in shutdown > > > floppy0: FDC access conflict > > > Floppy drive(s): fd0 is 1.44M > > > trying to lock fdc while usage count = 0 > > > floppy0: no floppy controllers found > > > md driver 0.35 MAX_MD_DEV=4, MAX_REAL=8 > > > > > > Of course I've got a floppy disk controller. I've tried > > disabling it but the > > > results are the same. I would appreciate any help because now I've been > > > trying to install linux in this computer for almost six months > > and in the > > > meanwhile I have to use linux on a 386. > > > > > > Please help me > > > > > > Regards > > >
Re: Make Config Does Not Work!
> I look forward to hearing from someone who can offer help on making > "make config" work. > > I followed the instructions of > http://www.us.debian.org/doc/FAQ/debian-faq-11.html, "The Debian > GNU/Linux FAQ: Debian and the kernel." > > Installed: gcc, libc6-dev, bin86, binutils, and gawk. > Executed: dpkg --install kernel-package_VVV_all.deb [..snip..] You also need to install kernel-source*.deb. Pierre Blanchet
Re: scilab demos crash whole scilab
> > I have installed scilab-2.4 (binaries form ftp.inria.fr) > and when I start it, it works but when I choose from menu > demos scilab crashes with the following messages: > > [1] 410 > $ Scilab: warning, error event receieved: > X Error of failed request: BadPixmap (invalid Pixmap parameter) > Major opcode of failed request: 54 (X_FreePixmap) > Resource id in failed request: 0x0 > Serial number of failed request: 1920 > Current serial number in output stream: 1985 > > How can I get scilab demos runnig properly under debian2.0 ? You are possibly out of (X)ressources (colors,...). Try to kill netscape or any application that consumes a lot of colors (xv,gimp,...) before running scilab. > > > Jan Krupa
Re: machine hangs at boot
> > --twz1s1Hj1O0rHoT0 > Content-Type: text/plain; charset=us-ascii > Content-Transfer-Encoding: quoted-printable > > On Wed, Sep 02, 1998 at 11:08:21PM -0500, Azog wrote: > > > Please post your lilo.conf and your partition table. > > hda: FUJITSU MPB3064ATU, 6187MB w/0kB Cache, CHS=3D788/255/63, UDMA=20 > > lilo.conf: > > boot =3D /dev/hda2 > > delay =3D 40 > > compact > > vga =3D normal > > root =3D /dev/hda2 > > read-only > > prompt > > image =3D /vmlinuz-2.1.119 > > label =3D new > > image =3D /vmlinuz-2.1.117 > > label =3D linux > >=20 > > Partition Table(from fdisk): > > Disk /dev/hda: 255 heads, 63 sectors, 788 cylinders > > Units =3D cylinders of 16065 * 512 bytes > >=20 > >Device Boot BeginStart End Blocks Id System > >/dev/hda111 33 265041 82 Linux swap > >/dev/hda2 * 34 34 224 1534207+ 83 Linux native > >/dev/hda3 225 225 355 1052257+ 83 Linux native > >/dev/hda4 356 356 788 3478072+ 83 Linux native > Lilo is installed on the 2nd partition and this is the active one, looks ok. > The only reason I could think of is that the MBR is damaged/not installed. > If you have a DOS Bootdisk (>4.0) try starting from it and > give the command > fdisk/mbr > Then reboot again. > > Nils > > -- Try to put boot=/dev/hda (not /dev/hda2) in order to put LILO on the MBR. -- Pierre Blanchet.
Re: Last try: can't make rshd recognize +@netgroup
"KS" == Kevin Squire <[EMAIL PROTECTED]> writes:
KS> Hi there --
KS> I've tried asking this before, but I think I might have been asking the
KS> wrong question (or at least no one answered). I'll try one more time with
KS> different wording. On a hamm machine, I'm trying to get rshd to let me
KS> put [EMAIL PROTECTED] in hosts.equiv. If I list the local machines in the
KS> file, lets users in, but when I try to use the netgroup alias, it doesn't
KS> seem to perform the nis lookup. It worked fine before upgrading to hamm.
KS> I thought the problem was in nsswitch.conf, but that seems to be fine.
KS> If anyone has any ideas, I would really appreciate it--thanks.
KS> Kevin
KS> -
KS> Kevin Squire
KS> [EMAIL PROTECTED]
KS> Research Student
KS> Nakajima Research Laboratory
KS> Tokyo Institute of Technology
From /usr/doc/libc6/BUGS.gz
-
[ *] The rcmd() functions (more concrete: the handling of .rhosts) does
not support netgroups and wildcards (+).
-
There might be a solution since RedHat 5.0 seems to support
netgroups. You might take a look at PAM.
I've tried different PAM configurations, but to no avail.
Please, mail me if you have success in configuring pam.
Pierre Blanchet.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: XF86_S3V Problems
"TB" == Tim Buller <[EMAIL PROTECTED]> writes: TB> I have a hamm system with a STB Nitro 3D / 4MB video card, using the S3V TB> server from XFree86 3.3.2.1-1. The display is managed by xdm. TB> The problem is the about 75% of the time when I end a session (normal WM TB> is afterstep, but the problem happens with others as well), the machine TB> locks hard and vertical bars of color are displayed on the screen. The TB> only way to get it back is to hard boot; it does not respond to pings from TB> remote hosts, etc. TB> This does not happen when I use Accelerated X 4.1. Any ideas would be TB> appreciated... TB> Tim TB> TB> Tim Buller [EMAIL PROTECTED] TB> Systems Specialist Office: Snow Hall 643 TB> Department of Mathematics Voice: 785-864-7311 TB> University of Kansas, Lawrence, KS 66045 Fax: 785-864-5255 You may want to try the XF86_SVGA server instead. It works better for me. You do not need to change your XF86Config, just install the server and check /etc/X11/Xserver. Pierre Blanchet. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: non-interactive Debian installation or update
"HZ" == HOLGER ZIEROLD <[EMAIL PROTECTED]> writes: HZ> Hi, HZ> I'm looking for a way to automate an upgrade/update of already HZ> installed Debian systems. Since there are several packages that HZ> require user interaction this cannot be done over night or at remote HZ> sites without the intervention of a system administrator. HZ> Does the dpkg command allow to accept default values for the specified HZ> packages with or without a configuration file or is there any other HZ> way to do it? HZ> Thanks, HZ> Holger. [..sig snip..] First install the package you want with: tee package.install.log | dpkg -i package.deb (Your answers will be logged in the file package.install.log) -=- MIME -=- Hi, I'm looking for a way to automate an upgrade/update of already installed Debian systems. Since there are several packages that require user interaction this cannot be done over night or at remote sites without the intervention of a system administrator. Does the dpkg command allow to accept default values for the specified packages with or without a configuration file or is there any other way to do it? Thanks, Holger. === * HEWLETT-PACKARD GmbH Holger Zierold / **PS Telecom *** /_ _ ** eMail: [EMAIL PROTECTED] ** / / /_/ ** P.O. Box 1641 HP-Telnet: 706-1437 *** / *** 61286 Bad Homburg phone: +49 6172 16-1437 / ** Germany fax: +49 6172 16-1419 * === -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] On your other machines, just do : cat package.install.log |dpkg -i package.deb Here, i have a central machine where root can access all other machines using rsh : for i in 'list-of-machines' do echo -n $i ... rcp package.deb $i:/tmp/package.deb cat package.install.log | rsh $i "PATH=/usr/sbin:$PATH \ dpkg -i /tmp/package.deb" echo done. done Pierre Blanchet. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: How to build Debian Linux cluster?
"RC" == Rainer Clasen <[EMAIL PROTECTED]> writes: RC> Pierre Blanchet ([EMAIL PROTECTED]): >> Here, i have 85 PC to install. And the method i will try will be >> slighty different (note: it will only work in a network, with a >> local mirror): >> 3) Write a shell script, dpkg-record, something like : RC> Just out of curiosity: What speaks against using rdist? IIRC somebody RC> suggested this in a previose discussion on this topic. RC> Rainer Because rdist is just like cp. It will only distribute directories (actually files also, but it will be very difficult to rdist 10-20k files of a traditionnal Debian system). The idea behind dpkg-record/dpkg-replay is flexibility and simplicity. You can install whatever package wherever you want automatically (if you have installed this package already). Pierre Blanchet. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: How to build Debian Linux cluster?
"KLS" == Kenneth L Summers <[EMAIL PROTECTED]> writes:
>> Hi, Ken!
>>
>> I post this message again, as there were some problems with listserver.
KLS> I think it might have gotten lost at this end... we have a non-Linux mail
KLS> server :)
>> > component in the cluster within 2 hours. I have since found out
listening
>> > to this list that the dpkg utility helps to make this chore a lot
simpler,
>> > if you learn to use the command line interface (silly me).
>> >
>> Could you please give more details about that. Imagine I need to update
some
>> package (perl for example) on 6 nodes simultaneously, what do I do?
KLS> It was posted a week or so ago, but here's what they said:
KLS> a) Install one machine with all the packages you like.
KLS> b) Get the selection with
KLS> dpkg --get-selections > my.selection
KLS> c) Install the next machine
KLS> Quit dselect after you specified the access method
KLS> d) Add the selection from the other machine with
KLS> cat my.selection | dpkg --set-selections
KLS> e) Run dselect to install the files Or if you need batch mode
KLS> mount /pub/debian under /mnt and then
KLS> cd /mnt
KLS> dpkg -iGROEB hamm/hamm hamm/contrib hamm/non-free
KLS> I suppose this assumes you're installing hamm (the last command anyway).
KLS> What I had done is write a script that would take as input a status file
KLS> from the machine I want to clone, the status file of the newly installed
KLS> machine, and produced a status file that would cause dselect/install to
KLS> install or remove whatever was necessary. This kind of fiddling with the
KLS> system makes me nervous, though.
KLS> Hope that helps.
KLS> Ken
Here, i have 85 PC to install. And the method i will try will be
slighty different (note: it will only work in a network, with a
local mirror):
1) Install the first machine (just floppy install).
2) Run dselect and set up my selection.
3) Write a shell script, dpkg-record, something like :
for i in
do
tee $i.install.log | dpkg -i $i
done
($i.install.log is my answer when a package asks me)
4) Modify the dinstall script on the boot floppies to not run
dselect at reboot but something like :
mount local-mirror:/debian /mnt
echo > /root/.rhost
/mnt/dpkg-replay
where dpkg-replay is :
for i in
do
cat $i.install.log | dpkg -i $i
done
5) whenever i install a new package on the first machine, i'll
just :
on the first-machine: dpkg-record package.deb
and then rsh dpkg-replay package.deb
Well, i miss some glue, but you get the idea,
Pierre Blanchet.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Can't get a kernel to recognize my Asus PCI-SC875 scsi card
"GDM" == G Dale Miller <[EMAIL PROTECTED]> writes: GDM> I am trying to build a kernel with only the options I want now that I GDM> have hamm GDM> installed again. The boot floppy made from this install recognizes and GDM> uses the GDM> ASUS SC875 (Symbiosor NCR) scsi card that I have no problem but when I GDM> build my own kernel it won't recognize it. GDM> I am trying to build from the kernel-source-2.0.33 package in hamm GDM> and am choosing the ncr53c7xx,8xx option for my scsi card which I GDM> thought should work. Am I choosing the wrong card or missing some sort GDM> of option that is included in the boot disks? There are two drivers for the NCR 53c8xx SCSI card. You should not choose nc53c7xx,8xx but the BSD driver (ncr53c8xx), which possibly didn't show up or was grayed if you have selected nc53c7xx,8xx. It's faster and works very fine with the SC875. Pierre Blanchet. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Strange NIS behaviour
I'm running Debian 1.3.1 on 55 machines here, and they all have the
same problem:
$ps aux|grep yp
blanchet 5008 0.0 1.0 912 328 p0 S 14:28 0:00 grep yp
Ok, ypbind does not run (it crashes often when NIS server shuts down).
$ypwhich
can't clnt_call: Can't communicate with ypbind
Ok.
$ypcat passwd
[... my whole passwd map ...]
I don't understand why, ypbind is not running but NIS functions
still work ?
nis version:
2.20-1
libc version:
5.4.33-6
Thanks
Pierre Blanchet.
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: Debian-lite
"AC" == Alec Clews <[EMAIL PROTECTED]> writes: >> That's why the name Debian-lite makes me sick. It reminds me of >> those commercial products where you have to pay for a full >> version of the product. AC> OK. How about Debian-Desktop? (Debian-BBS, Debian-ISP etc) AC> Regards, Alec -- Alec Clews, <[EMAIL PROTECTED]>, PGP AC> keyid:48FA EB81 TCA Consulting Ltd Tel. 44-(0)171-415-8159 AC> Fax:44-(0)171-556-0022 New CIty Court, 20 St Thomas Street, AC> London, Britain, SE1 9SD == Usual AC> Disclaimers Apply = Personal and PGP key AC> http://www.earth.demon.co.uk/alec How about Deb-One (it's sound like debian, but built for one special task, or just for one user) ? Pierre Blanchet. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: nis problem...
>>>>> "Magic" == Magic <[EMAIL PROTECTED]> writes: Magic> Hi! I have some computers in the laboratory with RH 4.1 Magic> One of tem is a nis server, a the rest - clients. Now i Magic> would like to have one PC with Debian on my desk :) Magic> and... i've a problem... Magic> I instaled: nis-2.10-1 Magic> other: libc5 5.4.23-2 libgdbm1 1.7.3-19 netbase 2.10-1 Magic> In /etc/deafultdomain i have my nis domain... In Magic> /etc/yp.conf adreess of my nis server and in /etc/passwd Magic> i've add +:: Magic> /etc/init.d/nis stop ; /etc/init.d/nist start Magic> and... what can i do now? It's my first touch nis and i Magic> havn't idea. can u help me? thank's... Magic> what about /var/yp/binding ? Magic> Magic --- Magic> Artur Zaworski, Computer Center, Technical University of Magic> Gdansk, Poland. email: [EMAIL PROTECTED] Magic> http://www.pg.gda.pl/~magic phone(office): 47.27.88 Dzialac Magic> mozna dopoty, dopoki sie ma marzenia. Ok, downgrade to the libc in stable(5.4.20-?). I got the same problems, (but my machine _is_ a NIS server so ypserv works but not ypbind). Unstable is unstable ... :) Pierre Blanchet.
Building a Debian bootable CD-Rom ?
I've heard somewhere that 1.2 could be bootable from a CD.
Well, I have a CD-writer (under Win NT, no chance :-)) a
complete mirror of the Debian distribution.
But i've no experience writing a CD. I would appreciate any
help.
Thanks,
Pierre Blanchet.
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
