Updated check-debian-cd

2000-10-03 Thread Francesco Potorti`
check-debian-cd is a shell script for quick checking of Debian CD images
after  having burnt them,  and for  checking of  Debian CD  images after
having rsync'ed them.  It is released under GNU GPL.  Simple usage:

   check-debian-cd
looks in /dev/cdrom and cheks the disc.  Or:

  check-debian-cd binary-i386-1_NONUS.iso
checks the image.  You can give any number of images as arguments.

Checking is  done using  the md5sum program,  and comparing  the results
with an  internal database of  sums.  The program  automatically detects
the  image, because  the internal  database contains  signatures  of the
Debian CDs.   If the  signature is not  found in the  internal database,
check-debian-cd  looks in  the  current and  parent  directories for  an
MD5SUMS file (or  one can be provided with an  option), and compares the
hash generated by md5sum with all those found in the MD5SUMS files.

Comments  and  bug  reports  welcome.   The  latest  version  is  always
available  at  ftp://fly.cnuce.cnr.it/pub/software/unix/check-debian-cd.
Please note that I am not on debian-user.



Updated check-debian-cd

2000-09-27 Thread Francesco Potorti`
I made an update to my check-debian-cd script.  Now it knows about every
Debian CD in the r0 and 0a  series, and about the binary i386 slink CDs.
It also emits different return  codes depending on the failure mode, and
can usefully  be used to  check for  a whole series  of iso images  in a
single run.

I  suggest  it to  be  added  to the  pseudo-image  toolkit,  and to  be
mentioned in the FAQ as an easy way of checking a CD after having burned
it.  To check any Debian CD, just do:
  check-debian-cd
it will automatically recognise the CD and check the MD5 hash.

To check a series of images:
  check-debian-cd *.iso

It can also be run in quiet mode from scripts.

Please note that I do not subscribe to either debian-cd or debian-user.

===File /var/mirrors/potato/check-debian-cd=
#! /bin/bash

## Made September, 2000 by Francesco Potortì [EMAIL PROTECTED]
## current version is 1.3
## This program is released under the GNU public license, whichever version
##
## BUGS:
## Does not work when file names given as arguments contain spaces

isodefault=/dev/cdrom
quiet=
cdversion=2.2.0

while getopts :qv option; do
case $option in
q)
quiet=yes ;;
v)
echo $cdversion; exit 0 ;;
*)  
cat -EOF
A program for checking Debian CDs and image files.
It reads /dev/cdrom or the given files and compares their MD5 
hash with
those contained in an internal database of Debian CDs.

usage: $(basename $0) [options] [files...]
   Options:
   -q: run quietly, do not print informative messages
   -v: print last known version of CD images (currently 
$cdversion)
EOF
exit 10
esac
done

bs=2048

## List of known Debian cd image disks: length in $bs, md5sum, id string
## the list is complete as of September, 2000, apart from the slink sources
#
#Id- 312412 c2c046ed739d78d56b1c5b34202268d1 'Debian-i386 2.1 Disc 1  '
#Id- 265072 f126f97fb20f3e1dd4868ae1d3309d24 'Debian-i386 2.1 Disc 2  '
#Id- 315877 b0d41dc89753ed4af3927d2446e8c7fe 'Debian 2.2 r0 Src-1 '
#Id- 326892 3e6272d23a5f6a2f3c8ce92841eca119 'Debian 2.2 r0 Src-1_NONUS   '
#Id- 323574 c04d100965361437e0f970d781153274 'Debian 2.2 r0 Src-2 '
#Id- 321204 a8142b47bce78bc9678ba2dcf52984ba 'Debian 2.2 r0 Src-3 '
#Id- 316351 b5951d4325de81eeaa8514ec78ecab9d 'Debian 2.2 0a alpha Bin-1   '
#Id- 322398 c83fe57ba0f4dc615efcaed19dc166cc 'Debian 2.2 0a alpha Bin-1_NONUS '
#Id- 327489 3b24410ea4fb25e455f2855407d9cb0f 'Debian 2.2 0a alpha Bin-2   '
#Id- 258784 fde444ae21f8ea4f18909d4a7a5acd91 'Debian 2.2 0a alpha Bin-3   '
#Id- 304639 9e757c6cd952efb030383fd6dd327a0f 'Debian 2.2 r0 arm Bin-1 '
#Id- 304977 35bc41583a999c43f33021e0ae4c5716 'Debian 2.2 r0 arm Bin-1_NONUS   '
#Id- 317057 fff543d36e1e313944c463a24dfb4307 'Debian 2.2 r0 arm Bin-2 '
#Id- 50704  4a4f07b7c7130c39f74340e05260a2be 'Debian 2.2 r0 arm Bin-3 '
#Id- 321743 c4435b6a6b8dd91ea51bd69bfb8642df 'Debian 2.2 r0 i386 Bin-1'
#Id- 327271 aa81c0848d199e9317735a1af0a2e4cc 'Debian 2.2 r0 i386 Bin-1_NONUS  '
#Id- 328594 be04cd6d17159d66978ad227b26ed17d 'Debian 2.2 r0 i386 Bin-2'
#Id- 257243 13a1509ea29e4891004ca1c397b316a6 'Debian 2.2 r0 i386 Bin-3'
#Id- 322735 83b0c5b40ca5b31db973d2a28bcc50af 'Debian 2.2 r0 m68k Bin-1'
#Id- 329160 6ccad6848468e7eccb9a8550215cddc9 'Debian 2.2 r0 m68k Bin-1_NONUS  '
#Id- 330903 a73b2ea406e0c493361236ebffb5ebdb 'Debian 2.2 r0 m68k Bin-2'
#Id- 111328 5a8c7778d4e18ddd622df8c8f5d4f3e9 'Debian 2.2 r0 m68k Bin-3'
#Id- 319500 c4435b6a6b8dd91ea51bd69bfb8642df 'Debian 2.2 r0 ppc Bin-1 '
#Id- 324505 aa81c0848d199e9317735a1af0a2e4cc 'Debian 2.2 r0 ppc Bin-1_NONUS   '
#Id- 329991 be04cd6d17159d66978ad227b26ed17d 'Debian 2.2 r0 ppc Bin-2 '
#Id- 141048 13a1509ea29e4891004ca1c397b316a6 'Debian 2.2 r0 ppc Bin-3 '
#Id- 326020 b8eca28f7eb4c96ce7a6f74efbe2f9bf 'Debian 2.2 0a sparc Bin-1   '
#Id- 331717 e5065a6bd3bdaf44bd2f00951c119fb3 'Debian 2.2 0a sparc Bin-1_NONUS '
#Id- 326370 acce0bffbd85edf52f9a446da3c2cb56 'Debian 2.2 0a sparc Bin-2   '
#Id- 153459 25e78e58aac0f229a6d7d60bdc9849ae 'Debian 2.2 0a sparc Bin-3   '

shift $((OPTIND-1))
case $# in
0)  # no arguments
iso=$isodefault ;;
1)  # the image file or device
iso=$1 ;;
*)  # more than one file: recursive call
options=
test $quiet  options=$options -q
exitcode=0
for arg; do
$0 $options $arg
ec=$?
test $ec -gt $exitcode  exitcode=$ec  test $quiet  break
done
exit $exitcode
esac

if [ ! -r $iso ]; then
echo Cannot read file $iso 2
exit 1
fi

set - $(grep 

No guest account from Mac using netatalk

1997-12-01 Thread Francesco Potorti`
Hi,

I installed netatalk and it works well, but I cannot login as guest
user from the Mac, the Guest radio button remains grayed.

What am I doing wrong?

Thanks
Francesco


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .


Bogomips halved!

1997-11-28 Thread Francesco Potorti`
Hi, 

I am using a 200MHz K6, which I now overclocked at 233 MHz.  After
having done that, I checked the bogomips number, and I discovered in
/var/log/messages that, after having been at 400 since ever (as
expected, for a K6), it had dropped at 200 in the latest reboot and,
after having overclocked it, went at 233.  If I run the bogomips
program, it gives 466, but /proc/cpuinfo still contains 233.

How can it be?  Could that be a source of problems for timings inside
the kernel?

Thanks for any info
Francesco


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .


Strange MS mouse behaviour: defective mouse or serial driver?

1997-11-26 Thread Francesco Potorti`
Hi,

I am experiencing a strange problem with a mouse used with gpm.
I connect a Microsoft mouse to /dev/ttyS0, then issue:

killall gpm; gpm -m /dev/ttyS0 -t bare

so far, so good: gpm displays its cursor, which moves regularly.  Then
I issue:

killall gpm; gpm -m /dev/ttyS1 -t bare
killall gpm; gpm -m /dev/ttyS0 -t bare

and the mouse stops working.  If I pull it out and connect it again,
it starts working.  This is perfectly reproducible.

An older MS mouse does not exhibit this behaviour.  That is, it does
not stop working when I change the gpm port and switch it back.  Is
the newer mouse broken?  Should I ask for a replacement or it is
normal?  And why ever?

Thanks for reading

Francesco


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .


Experience with initrd and loop devices (script included)

1997-11-26 Thread Francesco Potorti`
I started building an initrd image for use with loadlin and md (I
wanto to put everything under raid-0).  As first step, I needed the
image of an ext2 (or whatever) file system to mount via loopback
device, in order to put into it all the necessary modules and so on.

First, I realised that, if any docs exist on how to use the loopback
device, they are well hidden.  Hints are found in the mount man page
and in the Bootdisk-HOWTO, but nothing general enough.

Indeed, everyone assumes that you have an image and want to mount it,
but nothing is said about what to do to *create* an image from
scratch.  Well, the Bootdisk-HOWTO says how to do, but it does not
work...  It says that you just need to do

dd if=/dev/zero of=/tmp/fsfile bs=1k count=nnn
mount -o loop -t ext2 /tmp/fsfile /mnt

and voila`.  But mount refuses to mount, because it does not find a
superblock.  So I tried to mkfs.ext2 on the file, but to no avail:
apparently mkfs is not able to work on loop devices.  Moreover, the
mkfs.ext2 man page is missing (I filed a bug report).

Well, if you cannot suggest a better way, here is a small script that
creates an ext2 file systrem image from scratch.  One caveat: it uses
a ram disk, so you must have enough free memory to hold all of it.

Improvements welcome.

---
#!/bin/sh

# Create file IMG of size SIZE KB, which is the image of an ext2
# file system, using ramdisk device RAM, and mount it at MNT

IMG=/boot/initrd.img
SIZE=200
RAM=/dev/ramdisk
MNT=/mnt/initrd
LABEL='initrd image'

set -ex # exit on all errors

mkfs.ext2 -L $LABEL -m 0 -v -b 1024 $RAM $SIZE
mount $RAM $MNT
mount $RAM $MNT -o remount,ro
dd if=$RAM of=$IMG bs=1024 count=$SIZE
umount $MNT
mount -t ext2 -o loop $IMG $MNT


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .


strange MS mouse behaviour: defective mouse or serial driver?

1997-11-25 Thread Francesco Potorti`
Hi,

I am experiencing a strange problem with a mouse used with gpm.
I connect a Microsoft mouse to /dev/ttyS0, then issue:

killall gpm; gpm -m /dev/ttyS0 -t bare

so far, so good: gpm displays its cursor, which moves regularly.  Then
I issue:

killall gpm; gpm -m /dev/ttyS1 -t bare
killall gpm; gpm -m /dev/ttyS0 -t bare

and the mouse stops working.  If I pull it out and connect it again,
it starts working.  This is perfectly reproducible.

An older MS mouse does not exhibit this behaviour.  That is, it does
not stop working when I change the gpm port and switch it back.  Is
the newer mouse broken?  Should I ask for a replacement or it is
normal?  And why ever?

Thanks for reading

Francesco


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .


Re: tar, cpio, or cp?

1997-11-25 Thread Francesco Potorti`
Peter S Galbraith wrote:
   tar cf /path | ( cd other/path ; tar xvf -)
   
[EMAIL PROTECTED] (Stephen Zander) writes:
   cd /path; find . | cpio -dump /other/path
   
Why not 
   cp -a /path /other/path

Are there real differences between the three?  What about symlinks,
hard links, empty dirs, devices?  I know, I could read the manuals,
but after a first look none of the three seems to do everything right,
that's why I am asking.

Francesco


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .


How do people make backups on Linux?

1997-11-25 Thread Francesco Potorti`
Hi,

I am installing a system with a QIC tape.  I would like to know how do
people on this list do their backups.  I am switching from an ATT
environment where I used bru, and am very satisfied with it.  I know
it is available for Linux, too, but I was wondering if you can suggest
free alternatives.

I find taper way too unstable, and top/afio don't seem too
professional (I'd like to hear your opinions about this).  Also, are
there good experiences about using cpio or tar?

Thanks
Francesco


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .


Re: How do people make backups on Linux?

1997-11-25 Thread Francesco Potorti`
people on this list do their backups.  I am switching from an ATT
environment where I used bru, and am very satisfied with it.  I know
it is available for Linux, too, but I was wondering if you can suggest
free alternatives.
   
   dump  restore

I considered dump, but it does not compress, so I won't use it.  I am
thinking about improving tob to verify tapes.  I'd also love to hear
more comments on this topic, and people's experience.


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .


2.0.31: no more than 8 swap partitions, memory leaks?

1997-11-10 Thread Francesco Potorti`
Hi,

I am testing a new installation of 2.0.31 on the latetest stable
Debian.  Two things to notice up to now.

I left a torture test run on disks (three Bonnie processes running in
parallel).  All was okay, apparently.  While doing so, I tried to run
mirror on the Debian ftp site, and mirror grew in memory up to 100 MB
or so (exhausting all virtual memory).  At that point, every program I
tried to execute (ls or ps, for example) dumped core.  That's not nice
behaviour.  What's worst, memory remained in use after having managed
to kill mirror.  With an idle system, memory occupancy reported by
free was 50 MB higher than just after reboot.  Is 2.0.31 known to
suffer from memory leaks and bad behaviour when exhausting virtual
memory?  That would be very bad for a Unix kernel.  What is the latest
really stable kernel?  2.0.27?

Another thing is that swapon does not let me add more than 8 swap
partitions.  I read in the FAQ that up to 16 were allowed, but that
number is not reported in the swapon(8) nor in the swapon(2)
manpages.

Moreover, swapon -s will fail, telling that no /proc/swaps is
present.  That's true, but shouldn't have it been created by the
Debian installation?  How can I create it?

Thanks to everyone.


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .


Re: 2.1.30 and 3c900 module: Transmitter access conflict

1997-11-10 Thread Francesco Potorti`
[EMAIL PROTECTED] (Ferenc Kiraly) writes:
   
The subject says it all: I am installing a Debian 1.3.1 system on
a brand new PC with a 3c900 Combo board.
   
   I use that card with the vortex driver and it works fine ... I
   think the 3c59x module is the wrong one.

Thanks for you answer.  I don't know about the vortex driver.  Anyway,
the 3c59x is okay for that card.  Its author, Donald Becker, told me
that indeed the autoprobe for AUI does not work, because of apparently
insufficient demand for that feature.  However, I have as yet no
explanations for some (apparently innocuous) error messages appearing
at and shortly after module iniatialisation.  The card seems to work
okay.

I brought the discussions on the [EMAIL PROTECTED]
list.

Again, thanks to all that have helped.
 -=- MIME -=- 

--Boundary_(ID_Fy0Nmukjup0C7yYi9Ndneg)
Content-type: text/plain

This report relates to a message you sent with the following header fields:

  Message-id: [EMAIL PROTECTED]
  Date: Mon, 10 Nov 1997 10:29:06 +0100
  From: Francesco Potorti` [EMAIL PROTECTED]
  To: [EMAIL PROTECTED] (Ferenc Kiraly)
  Subject: Re: 2.1.30 and 3c900 module: Transmitter access conflict

Your message cannot be delivered to the following recipients:

  Recipient address: [EMAIL PROTECTED]
  Reason: Not found in directory


--Boundary_(ID_Fy0Nmukjup0C7yYi9Ndneg)
Content-type: message/DELIVERY-STATUS

Original-envelope-id: [EMAIL PROTECTED]
Reporting-MTA: dns; mailsrv.cnuce.cnr.it

Action: failed
Status: 5.1.1 (Not found in directory)
Original-recipient: rfc822;[EMAIL PROTECTED]
Final-recipient: rfc822;[EMAIL PROTECTED]

--Boundary_(ID_Fy0Nmukjup0C7yYi9Ndneg)
Content-type: MESSAGE/RFC822

Return-path: [EMAIL PROTECTED]
Received: from DIRECTORY-DAEMON by mailsrv.cnuce.cnr.it (PMDF V5.1-9 #16550)
 id [EMAIL PROTECTED]; Mon, 10 Nov 1997 10:27:13 MET
Received: from fly.cnuce.cnr.it ([131.114.192.86])
 by mailsrv.cnuce.cnr.it (PMDF V5.1-9 #16550)
 with SMTP id [EMAIL PROTECTED] for
 [EMAIL PROTECTED]; Mon, 10 Nov 1997 10:27:11 MET
Received: by fly.cnuce.cnr.it (Smail3.1.26.7 #3) id m0xUq9b-0001LgC; Mon,
 10 Nov 1997 10:29 +0100 (MET)
Date: Mon, 10 Nov 1997 10:29:06 +0100
From: Francesco Potorti` [EMAIL PROTECTED]
Subject: Re: 2.1.30 and 3c900 module: Transmitter access conflict
In-reply-to: [EMAIL PROTECTED]'s message of 6 Nov 97 09:00:10 GMT
Sender: [EMAIL PROTECTED] (Francesco Potorti`)
Sender: [EMAIL PROTECTED]
To: [EMAIL PROTECTED] (Ferenc Kiraly)
Cc: [EMAIL PROTECTED]
Message-id: [EMAIL PROTECTED]
MIME-version: 1.0
X-Mailer: Gnus v5.3/Emacs 19.34
Content-type: TEXT/PLAIN
Lines: 20
References: [EMAIL PROTECTED]

[EMAIL PROTECTED] (Ferenc Kiraly) writes:
   
The subject says it all: I am installing a Debian 1.3.1 system on
a brand new PC with a 3c900 Combo board.
   
   I use that card with the vortex driver and it works fine ... I
   think the 3c59x module is the wrong one.

Thanks for you answer.  I don't know about the vortex driver.  Anyway,
the 3c59x is okay for that card.  Its author, Donald Becker, told me
that indeed the autoprobe for AUI does not work, because of apparently
insufficient demand for that feature.  However, I have as yet no
explanations for some (apparently innocuous) error messages appearing
at and shortly after module iniatialisation.  The card seems to work
okay.

I brought the discussions on the [EMAIL PROTECTED]
list.

Again, thanks to all that have helped.

--Boundary_(ID_Fy0Nmukjup0C7yYi9Ndneg)--


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .


2.1.30 and 3c900 module: Transmitter access conflict

1997-11-05 Thread Francesco Potorti`
Hi,

The subject says it all:  I am installing a Debian 1.3.1 system on a
brand new PC with a 3c900 Combo board.  Since I have a Buslogic
Flashpoint SCSI card, I used the disks under the special/ directory,
which use 2.1.30.  Loading the 3c59x module seems to work, because the
module correctly recognises the board.  It then prints the following
messages:

3c59x.c:v0.30-all 12/23/96 [EMAIL PROTECTED]
  3Com EtherLink III: Overriding PCI latency timer (CFLT) setting of 32, new 
value is 255.
loading device 'eth0'...
eth0: 3Com 3c900 Boomerang 10Mbps/Combo at 0xd400, 00:60:08:12:4d:ef, IRQ 12
  Internal config register is 10302d8, transceivers 0xe138.
  8K word-wide RAM 3:5 Rx:Tx split, autoselect/10baseT interface.

Which is okay (I think), apart from the fact that the used interface
is really the AUI.  Then I run /etc/init.d/network, obtaining these
messages:

eth0: initial media type 10baseT.
eth0: vortex_open() InternalConfig 010302d8.
eth0: vortex_open() irq 12 media status 8800.
eth0: Media selection timer tick happened, 10baseT.
eth0: Media 10baseT is has no link beat, 80c0.
eth0: Media selection failed, now trying 10base2 port.
eth0: Media selection timer finished, 10base2.
eth0: Media selection timer tick happened, 10base2.
eth0: Media 10base2 is has no indication, 4810.
eth0: Media selection timer finished, 10base2.

I would have expected a word about the AUI interface, but there is
not.  Anyway, if I now try a ping to a numerical IP address, I get
sequences of:

eth0: Transmitter access conflict.

and, needless to say, no response to the ping is received (so I assume
nothing is getting out of the card).

Anyone has any idea of what's happening?  Donald, are you listening?
Thanks to all

Francesco


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .


Re: 2.1.30 and 3c900 module: Transmitter access conflict

1997-11-05 Thread Francesco Potorti`
3c59x.c:v0.30-all 12/23/96 [EMAIL PROTECTED]
 ^
   This is incredibly old. Check the Ethernet-HOWTO for the URL of the
   site where you can download a version (0.46something) which works
   fine.

Thanks.  This is the version that comes with 2.0.30.  Maybe the kernel
maintainers should be made aware of this.  I'll download the latest
ASAP.

Francesco


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .


Re: 2.1.30 and 3c900 module: Transmitter access conflict

1997-11-05 Thread Francesco Potorti`
   On Wed, 5 Nov 1997, Philippe Troin wrote:
   
   I had a similar set-up where I needed to use the AUI interface.  I
   found that the autoselect mechanism got fooled and I had to set the
   media type manually at boot time (or at module load time).  I think
   it was options=1, but you ought to check the doc's for sure.  (I
   would, but my system is not accessible at the moment.)

I had tried `insmod 3c59x options=1', which I guessed by looking at
the sources (what docs?), but I was fooled into thinking that didn't
work because the load message said autoselect/10baseT interface.
However, I just tried again and, after running ifconfig, I get

eth0: media override to transceiver 1 (10Mbs AUI)

and indeed it works.  Thanks to all for your help (I'll get the latest
version of the driver nonetheless).


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .


Where is the StarOffice debianized package?

1997-10-13 Thread Francesco Potorti`
Hi,

I read some weeks ago about a StarOffice debianized package that was
in non-free, but I couldn't find it on ftp.debian.org, either in
non-free or hamm/non-free.

What happened to it?  Is it so well hidden?  I also made searches in
the Packages file for staroffice, star, office without results.

Thanks to all


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .


A place for data?

1997-09-08 Thread Francesco Potorti`
Where would you put trees of data that do not belong to any user, but
to a project, for example?  /home, /usr/local or what else?


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . Trouble? 
e-mail to [EMAIL PROTECTED] .


Re: SCSI Host Adapter (+ Re: 2 CPU servers)

1997-09-08 Thread Francesco Potorti`
I am resending this, this time to the mailing list.  I suspect that
the list-group gateway is not bidirectional, is that true?

[EMAIL PROTECTED] (Jeff Noxon) writes:
   
   Note that the non-CPU FlashPoint models are the only ones available
   with twin-bus configurations.  BusLogic is marketing these at the
   server segment.  If the dedicated CPU was a big performance win, I
   don't think they would do that.

Everyone seems to think that buslogic BT-948 or BT-958 is the best
choice for Linux (apart from DPT cards).  I asked Leonard Zubkoff and
he told me that the flashpoint cards are slightly faster than the
multimaster series, except for very high disk loads and high interrupt
latency, but normally flashpoint is okay.

What I said in another previous post of mine that I think got lost, is
that even though the latest buslogic driver by Zubkoff supports the
flashpoint series, Debian 1.3.1 still includes only support for the
multimaster.  This means, for example, that you cannot install debian
1.3.1 from scratch on a box with only a flashpoint board (and non EIDE
disk, for example).

If you think going buslogic, the following are very recommended
readings:
http://www.dandelion.com/Linux/Quantum.html
http://www.dandelion.com/Linux/BusLogic.html

   If you have any other questions about the BusLogic SCSI adapters,
   I suggest reading drivers/scsi/README.BusLogic in the kernel
   source tree.  And check out the web page listed there, as well as
   http://www.buslogic.com.

Also, the SCSI-HOWTO and README.Buslogic in debian 1.3.1 are outdated
with respect to the info in the references I gave above.  Beware.


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . Trouble? 
e-mail to [EMAIL PROTECTED] .