Linux-Misc Digest #184, Volume #27 Wed, 21 Feb 01 07:13:02 EST
Contents:
renaming 500+ files based on contents of an existing text file (Glitch)
Re: Kernel 2.4 / UIDs larger than 65535 (Karl Beckers - Sun Germany - Systems
Engineer)
Re: Reading CD-volumelabel? (Giulio Orsero)
Re: tar archive error (John Thompson)
Re: What is this? (John Thompson)
writing to tape (DLT 8000) (Christoph Kukulies)
/bin/sh (Christopher Albert)
Re: Compile problems (Eric Ho)
Re: What is this? ("rc")
Re: HOW DO I KILL THIS PROCESS? (Jean-David Beyer)
Re: enlightenment iconbox (Jean-David Beyer)
Re: <Q> using lpr to number pages (Jean-David Beyer)
Re: writing to tape (DLT 8000) (Jean-David Beyer)
Re: /bin/sh (Jean-David Beyer)
Re: Compile problems ("Peter T. Breuer")
Re: /bin/sh ("Peter T. Breuer")
Re: fastest SCSI CDROM drive (Robert Heller)
compiling glibc and gcc ("Mathias Rodenstein")
----------------------------------------------------------------------------
Date: Wed, 21 Feb 2001 03:06:49 -0500
From: Glitch <[EMAIL PROTECTED]>
Subject: renaming 500+ files based on contents of an existing text file
hello,
this is more of a curiousity question, although it would be nice if it
would be possible.
I have a collection of viruses. I use f-prot to scan them in order to
find out their correct names. I then rename the existing zip archive to
that of the name of the virus f-prot siad it was infected with(which is
the virus itself).
Problem is i have about 1000 viruses and doing it by hand is tedious.
Is it possible to make a shell script to read the output of the text
file generated by f-prot and then search for the correct zip archive on
the hard drive and have the shell script automatically rename teh zip
file to the name of the virus f-prot found it be infected by?
Here is what the output of f-prot looks like:
C:\ZIP2\VIRII\UNIXLO~1.SH Infection: Unix/LoveLetter
C:\ZIP2\VIRII\VBSFRE~1.VBS Infection: VBS/FreeLinks.A
C:\ZIP2\VIRII\NEWVIRII\ASCIIV.ZIP->VIRII/GLITCH/ASCIIV.COM Infection:
Ascii.613.unknown?
C:\ZIP2\VIRII\NEWVIRII\CEREBR~1.ZIP->cerebrus.zip->CEREBRUS.EXE
Infection: W95/Cerebrus.1482
I'd have the shell script find anything after "Infection:" and make that
the filename for the file listed on that same line.
Cerebrus.zip would be renamed to Cerebus.1482.zip
Is this possible and difficult to accomplish?
This would be a good project i can learn how to do shell scripting on if
it isnt too difficult.
Thanks
------------------------------
From: Karl Beckers - Sun Germany - Systems Engineer <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux
Subject: Re: Kernel 2.4 / UIDs larger than 65535
Date: Wed, 21 Feb 2001 11:00:58 +0100
This is a multi-part message in MIME format.
==============7535F77CAA2ADEE345064DF1
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Well,
thanks for the comment, but errm ... I have SuSE 7.1
installed with glib 2.2.7.
Any other ideas?
Karl.
Adam Gorecki wrote:
> Karl Beckers - Sun Germany - Systems Engineer wrote:
> >
> > Hi everybody,
> >
> > heard that the new Kernel 2.4 can handle uids larger than
> > 65535 ... need that for NFS mounting off Solaris ...
> >
> > While the systems seems to accept the uid chown doesn't
> > seem to allow me to chown that user any files.
> >
> > Seems to be about uid_t defined as int in GNU libc?!?
> >
> > Any chances to get this to work?
> > Been trying rebuilding fileutils ... should I defined uid_t
> > as unsigned log someplace?
> >
> > Any help appreciated,
> >
> > Karl.
> >
> Hi,
>
> I think you need glib 2.2.0 for this issue.
> See on suse's homepage for description of new 7.1 package.
>
> Kind regards,
>
> Adam
>
> --
> Ford-Werke AG \ Adam Gorecki \ MD/DME-29
> Henry-Ford-Str. 1 \ D-50735 Koeln
> +++++++++++++++++++++++++++++++++++++++++++++++++++++
> eMail private Uni Duisburg :[EMAIL PROTECTED]
> eMail business Ford-Werke AG:[EMAIL PROTECTED]
==============7535F77CAA2ADEE345064DF1
Content-Type: text/x-vcard; charset=us-ascii;
name="karl.beckers.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Karl Beckers - Sun Germany - Systems Engineer
Content-Disposition: attachment;
filename="karl.beckers.vcf"
begin:vcard
n:Beckers;Karl
tel;fax:+49 6103 752 299
tel;work:+49 6103 752 0
x-mozilla-html:FALSE
org:<P><A HREF="http://www.Sun.COM/"><Img Border=0
Src="http://www.sun.com/smrc/brand-elem/downloads/dotcomelement-blu.gif"></A></P>
adr:;;Amperestra�e 6;Langen;;63225;Germany
version:2.1
email;internet:[EMAIL PROTECTED]
title:Systems Engineer - Sales Support
x-mozilla-cpt:;0
fn:Karl H. Beckers
end:vcard
==============7535F77CAA2ADEE345064DF1==
------------------------------
From: Giulio Orsero <[EMAIL PROTECTED]>
Subject: Re: Reading CD-volumelabel?
Date: Wed, 21 Feb 2001 11:22:33 +0100
Michael Schaarwaechter <[EMAIL PROTECTED]> ha
scritto:
>fred smith wrote:
>> Dances With Crows <[EMAIL PROTECTED]> wrote:
>
>Thanks a lot for your help. isoinfo does what I want, but one must use
>the newest version which has a -V tag for read volumename. Mounted or
>unmounted doesn't matter because it uses the device directly.
>
If you have an old isoinfo you can do
isoinfo -d -i <device>
and you get the volume id together with other stuff.
--
[EMAIL PROTECTED]
------------------------------
From: John Thompson <[EMAIL PROTECTED]>
Subject: Re: tar archive error
Date: Tue, 20 Feb 2001 19:39:08 -0600
Wolfgang Batrla wrote:
> in order to clarify a misunderstanding about the tar
> archive:
> the file exists on a piece of quarter inch cassette tape. i
> did not download anything. so even if there is an end of
> fie, the information should still be there in one form or an
> other.
If the file resides on the part of the tape before the error you
probably will be able to recover it. If it's after the error,
all bets are off, especially if the data was compressed before it
was written to the tape. That's the main reason I switched from
tar to afio. I had been happily using tar for some time, no
problems, but once a crucial file needed to be restored and the
archive had an error. Fortunately I had an older version of the
file on an earlier backup, but at theat point I switched to afio
for my system backups.
--
-John ([EMAIL PROTECTED])
------------------------------
From: John Thompson <[EMAIL PROTECTED]>
Subject: Re: What is this?
Date: Tue, 20 Feb 2001 19:45:41 -0600
rc wrote:
> I keep finding this in my maillog.
>
> Feb 19 02:34:17 lnx sendmail[763]: CAA00763: <[EMAIL PROTECTED]>... User
> unknown
> Feb 19 02:34:17 lnx sendmail[763]: CAA00763: <[EMAIL PROTECTED]>... User unknown
> Feb 19 02:34:17 lnx sendmail[763]: CAA00763: <[EMAIL PROTECTED]>... User
> unknown
[clip...]
It looks like somebody's relaying mail (probably spam) off your
machine. You only see the attempts that failed here; god only
knows how much was successfully delivered.
I'd take the box off-line and check it thoroughly to make sure
nothing else has been compromised. Before you put it back on
line make sure you've configured sendmail to reject relaying
except from your local network and set up a firewall to help
deter future onslaughts.
--
-John ([EMAIL PROTECTED])
------------------------------
From: Christoph Kukulies <[EMAIL PROTECTED]>
Subject: writing to tape (DLT 8000)
Date: 21 Feb 2001 10:40:05 GMT
Since tapeware fails miserably on my RH 6.1 machine (SMP, nfs, yp, amd)
I have decided to use my own backup strategy.
I'm thinking of writing compressed tarballs to a DLT drive one of all
users in the network per day. Steered by a timestamp I put in some
directory on the central machine that has the tape drive. The user
dirs are automounted in /home/<host>/user. So everything is accessible
locally on the machine doing the backup.
First off, I can see only /dev/st0. Where is the nrst0?
How does one write to DLT? Is this different from 'normal' tapes?
I'm thinking of writing these tarballs as files to tape.
To recover data from a certain file I have to keep track of the
file numbers on tape and skip to the appropriate file.
How can I safely position to the end of tape for writing
new backups in the case that a recovery from backup has occured?
--
Chris Christoph P. U. Kukulies [EMAIL PROTECTED]
------------------------------
From: Christopher Albert <[EMAIL PROTECTED]>
Subject: /bin/sh
Date: Wed, 21 Feb 2001 12:17:47 +0100
Is there a linux distribution, for which
/bin/sh is NOT a symlink to bash?
Chris
------------------------------
From: Eric Ho <[EMAIL PROTECTED]>
Subject: Re: Compile problems
Date: 21 Feb 2001 11:22:23 GMT
Then what should I do, this is a fresh install of the Slackware.
Best Regards,
Eric Ho
D. Stimits <[EMAIL PROTECTED]> wrote:
> Eric Ho wrote:
>>
>> Hi,
>>
>> I have just installed Slackware 7.1 (actually Slackware-current),
>> and also upgraded the kernel to 2.4.1.
>> Everything is ok, except my Sound Blaster 16 doesn't work anymore,
>> and all of my K programs don't compile anymore. For example,
>> when try to compile my favorite ICQ program (KICQ), it gives the
>> following errors :
>> bash-2.04# make
>> make all-recursive
>> make[1]: Entering directory `/usr/src/kicq-0.3.0'
>> Making all in kicq
>> make[2]: Entering directory `/usr/src/kicq-0.3.0/kicq'
>> Making all in pics
>> make[3]: Entering directory `/usr/src/kicq-0.3.0/kicq/pics'
>> make[3]: Nothing to be done for `all'.
> This "Nothing to be done for `all'" leads me to believe there is stale
> code still there, compiled with your old system. Also be sure the
> directory "/usr/include/kde/" exists, and that the user doing the
> compile can cd there, and read files (those are the kde files it claims
> below as missing).
>> make[3]: Leaving directory `/usr/src/kicq-0.3.0/kicq/pics'
>> c++ -DHAVE_CONFIG_H -I. -I. -I.. -I/usr/local/include -I/opt/kde/include
>-I/usr/lib/qt/include -O2 -Wall -c main.cpp
>> In file included from main.cpp:8:
>> mainwindow.h:8: ktopwidget.h: No such file or directory
>> In file included from preferences.h:12,
>> from mainwindow.h:21,
>> from main.cpp:8:
>> prefuser.h:8: kintegerline.h: No such file or directory
>> In file included from preferences.h:13,
>> from mainwindow.h:21,
>> from main.cpp:8:
>> prefserver.h:8: kintegerline.h: No such file or directory
>> In file included from preferences.h:16,
>> from mainwindow.h:21,
>> from main.cpp:8:
>> prefproxy.h:8: kintegerline.h: No such file or directory
>> In file included from mainwindow.h:22,
>> from main.cpp:8:
>> findwindow.h:9: ktablistbox.h: No such file or directory
>> In file included from findwindow.h:16,
>> from mainwindow.h:22,
>> from main.cpp:8:
>> uinwidget.h:9: kintegerline.h: No such file or directory
>> make[2]: *** [main.o] Error 1
>> make[2]: Leaving directory `/usr/src/kicq-0.3.0/kicq'
>> make[1]: *** [all-recursive] Error 1
>> make[1]: Leaving directory `/usr/src/kicq-0.3.0'
>> make: *** [all-recursive-am] Error 2
>> bash-2.04#
>>
>> Can someone tell me what I did wrong or what I can do to fix the
>> problem ?
>>
>> Best Regards,
>> Eric Ho
------------------------------
From: "rc" <[EMAIL PROTECTED]>
Subject: Re: What is this?
Date: Wed, 21 Feb 2001 11:42:24 GMT
My domain is ccsn.com
I havbe had hackers in before. I rebuilt the system and turned a bunch of
stuff off. I don;t think the a-hole was able to relay anything. I did some
tests and I think it's safe. I wish I know how to block idiots like that
even more.
--
Robert
"Rod Smith" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> [Posted and mailed]
>
> In article <gPsk6.33877$[EMAIL PROTECTED]>,
> "rc" <[EMAIL PROTECTED]> writes:
> > Feb 19 02:34:17 lnx sendmail[763]: CAA00763: <[EMAIL PROTECTED]>... User
> > unknown
>
> [more like this snipped.]
>
> If your system is the mail server for ccsn.com, it looks like somebody
> is trying to spam random addresses or is fishing for usernames, possibly
> as a prelude to crack your system. If your system has no connection to
> ccsn.com, it looks like somebody's trying to use your system as a mail
> relay for spam. This has apparently been unsuccessful, but I'm puzzled
> as to the quantity of messages -- you'd think the spammer software would
> give up after the first attempt or two. It might be that some are
> getting through. You may want to check the following URL for spam
> prevention information:
>
> http://mail-abuse.org/tsi/
>
> > I keep finding this in my maillog. I also find this in my messages log:
> >
> > Feb 19 22:54:14 lnx rpc.statd[329]: gethostbyname error for
>
> [stuff that displays VERY strangely deleted]
>
> This looks like an attempt to exploit some sort of buffer overrun bug.
> If successful, rpc.statd would execute the code sent by the cracker,
> presumably giving the cracker access to your system. As a general rule,
> such attacks, if successful, don't appear in the log, but I can't
> promise that's true in this particular case (I don't know what bug your
> cracker was trying to exploit).
>
> IIRC, rpc.statd is related to NFS. This normally should not be
> accessible to the Internet. You can block access to the relevant ports
> by setting up ipchains (or iptables, on 2.4.x kernels) rules, or you can
> set up a firewall to protect your entire network. You'd be well advised
> to do this now. In fact, you should probably do at least minimal
> checking to be sure your system hasn't already been compromised. If
> you'r using an RPM-based system, "rpm -Va" will produce a list of files
> that have changed since installation. This list will probably be long
> because of legitimate changes, but changes to binaries or configuration
> files you know you haven't touched should raise a red flag.
> Unfortunately, the RPM database itself can be compromised, so a clean
> bill of health from this test doesn't necessarily mean you're safe.
>
> Unfortunately, network security is a complex topic. You may want to
> read up on it, at least minimally. There are Web sites like
> http://www.cert.org and http://ciac.llnl.gov that provide general
> information. There's a Linux Administrator's Security Guide at
> http://www.linuxdoc.org/LDP/lasg/. There are many books on Linux and
> UNIX security (I list some at
> http://www.rodsbooks.com/books/books-network.html).
>
> > Feb 20 03:11:55 lnx pumpd[251]: renewed lease for interface eth0
>
> This is normal if your network uses DHCP and your system is a DHCP
> client. DHCP "leases" IP addresses from a central DHCP server, which
> keeps track of which ones are in use. DHCP requires periodic lease
> "renewal," to keep the list of used IP addresses reasonably up-to-date.
> This message just indicates that pumpd (one of several DHCP clients for
> Linux) has successfully renewed its lease.
>
> --
> Rod Smith, [EMAIL PROTECTED]
> http://www.rodsbooks.com
> Author of books on Linux & multi-OS configuration
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Subject: Re: HOW DO I KILL THIS PROCESS?
Date: Wed, 21 Feb 2001 06:43:54 -0500
Tyler Larson wrote:
>
> >>So how to I get it off without rebooting?
> > [-]
> > You can't .. end of story,
> > Juergen
>
> It's kinda off topic, but as far as killing resistant processes:
>
> I heard that sometimes (particularly when a process hangs waiting on
> a network request), sending an INT (2) or QUIT (3) signal will often
> work even when the KILL signal fails.
>
> Anyone have any comments on that? Does it really work? Are there any
> other little known tricks to killing unwanted processes?
>
I very much doubt that a kill -2 or kill -3 would work when a kill
-9 fails. What is happening is that (this may be an
oversimplification) no interrupts will be delivered to the process
until the IO operation completes (as determined by examining the
STATUS flag of the process and observing that it is "D" as shown by
"top"). And the IO is not going to complete (I assume because an
interrupt from the device or controller was lost somehow).
In this case, I do not even know what device has failed to deliver
the interrupt (if it failed, and not just that the software got it
and lost it), though I cannot imagine it was other than a disk drive
or scsi controller that runs the disk drives, since the program in
question was rpm that I would not expect to be doing IO elsewhere.
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 6:35am up 14:09, 3 users, load average: 2.07, 2.12, 2.07
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Subject: Re: enlightenment iconbox
Date: Wed, 21 Feb 2001 06:48:26 -0500
Dan Martins wrote:
>
> There may be a more proper group to post this in, but i couldn't find
> one, so..
>
> I started using enlightenment as my window manager yesterday. I was
> messing around and closed my iconbox, now i'm wondering how to start it
> up again. I looked through all the options and couldn't find anything..
>
> Any tips?
Not from me. I have been running GNOME/Enlightenment since 1999 and
do not know what an iconbox is. Could you enlighten me?
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 6:45am up 14:19, 3 users, load average: 2.05, 2.12, 2.09
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Subject: Re: <Q> using lpr to number pages
Date: Wed, 21 Feb 2001 06:50:16 -0500
[EMAIL PROTECTED] wrote:
>
> Hi. is there a paramter I can pass to lpr to have it number the pages
> when I print? I am tired of having to hand number documents every time
> I print a web page from netscape :-)
>
I do not know that that is a proper function of lpr, but often an
earlier tool, such as pr, can do it. As in
pr -n filelist | lpr
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 6:45am up 14:19, 3 users, load average: 2.05, 2.12, 2.09
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Subject: Re: writing to tape (DLT 8000)
Date: Wed, 21 Feb 2001 07:00:26 -0500
Christoph Kukulies wrote:
>
> Since tapeware fails miserably on my RH 6.1 machine (SMP, nfs, yp, amd)
> I have decided to use my own backup strategy.
>
> I'm thinking of writing compressed tarballs to a DLT drive one of all
> users in the network per day.
I suggest cpio or afio (?) instead of tar, so it is easier to
recover from tape errors.
> Steered by a timestamp I put in some
> directory on the central machine that has the tape drive. The user
> dirs are automounted in /home/<host>/user. So everything is accessible
> locally on the machine doing the backup.
>
> First off, I can see only /dev/st0. Where is the nrst0?
There should be. Unless your tape drive cannot operate in
non-rewinding mode for some reason. I cannot imagine such a reason,
but one never knows.
>
> How does one write to DLT? Is this different from 'normal' tapes?
You can use cp, cpio, etc. I.e., it is "just another file", there is
no file system there.
>
> I'm thinking of writing these tarballs as files to tape.
As I said, I suggest avoiding tar.
>
> To recover data from a certain file I have to keep track of the
> file numbers on tape and skip to the appropriate file.
>
> How can I safely position to the end of tape for writing
> new backups in the case that a recovery from backup has occured?
>
I suggest writing only one backup to each tape. In the old days,
when tape cost was an issue, you might have been able to justify
writing a full backup and a few incrementals to the same tape. But
for my DDS-2 drive, 8 Gigabyte tapes are only about $8.00 each. And
with the disk capacity machines have these days, it is all you can
do to get a full backup onto a single tape as it is.
If you have, say, a full backup and a few incrementals on the same
tape, and the tape goes bad, you lose everything (just to save
$8.00). If you put your backups on separate tapes, you lose only one
day's work (which is bad enough). Do not save money by writing more
than one backup on a tape.
I have a separate tape for Monday, Tuesday, ..., Saturday. I have 5
tapes for Sunday (first Sunday, second Sunday, ...) each month, and
I cycle through these in the obvious way. That way I can always get
a backup less than a week old. I can get the any week for a month. I
also make a new monthly tape the first of each month and those go to
the safe deposit box at the bank. So I can ALWAYS get to any month I
might need (unless the bank suffers a disaster).
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 6:50am up 14:24, 3 users, load average: 2.01, 2.07, 2.08
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Subject: Re: /bin/sh
Date: Wed, 21 Feb 2001 07:01:47 -0500
Christopher Albert wrote:
>
> Is there a linux distribution, for which
> /bin/sh is NOT a symlink to bash?
>
Who knows? But you could always make it a symlink to something else,
if you want. (This might mess up some scripts somewhere, though, I
guess. I would worry most about the ones in /etc/rc.d/init.d if
running a Red Hat system.)
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 7:00am up 14:34, 3 users, load average: 2.01, 2.05, 2.07
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: Compile problems
Date: Wed, 21 Feb 2001 12:42:16 +0100
Eric Ho <[EMAIL PROTECTED]> wrote:
> Then what should I do, this is a fresh install of the Slackware.
What are you talking about?
> D. Stimits <[EMAIL PROTECTED]> wrote:
>> Eric Ho wrote:
>>> when try to compile my favorite ICQ program (KICQ), it gives the
>>> following errors :
>>> bash-2.04# make
>>> make all-recursive
>>> make[1]: Entering directory `/usr/src/kicq-0.3.0'
>>> Making all in kicq
>>> make[2]: Entering directory `/usr/src/kicq-0.3.0/kicq'
>>> Making all in pics
>>> make[3]: Entering directory `/usr/src/kicq-0.3.0/kicq/pics'
>>> make[3]: Nothing to be done for `all'.
>> This "Nothing to be done for `all'" leads me to believe there is stale
>> code still there, compiled with your old system. Also be sure the
Indeed. Just so.
>> directory "/usr/include/kde/" exists, and that the user doing the
>> compile can cd there, and read files (those are the kde files it claims
>> below as missing).
Fine advice. Please follow it! You might start by reading the README
that comes with the source you downloaded. If you didn't download it
(and I seriously doubt that!), then download it.
Peter
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: /bin/sh
Date: Wed, 21 Feb 2001 12:43:40 +0100
Christopher Albert <[EMAIL PROTECTED]> wrote:
> Is there a linux distribution, for which
> /bin/sh is NOT a symlink to bash?
I doubt it. Why would you need one? Bash is sh, when invoked as sh.
Symlink it to something else if you like! Make sure it is something
compatible, though, or you'll have to rewrite the init scripts and
everything :-).
Peter
------------------------------
From: Robert Heller <[EMAIL PROTECTED]>
Subject: Re: fastest SCSI CDROM drive
Date: 21 Feb 2001 06:04:54 -0600
[EMAIL PROTECTED] (B'ichela),
In a message on Tue, 20 Feb 2001 20:18:44 -0500, wrote :
B> On 20 Feb 2001 01:09:27 -0600, Robert Heller <[EMAIL PROTECTED]> wrote:
B> >2) CDROM drives are not like hard disks and don't use the 'Wide/Fast'
B> >interface (either ATA/Dma66 or SCSI-3). Thus the continuous data
B> >throughput is limited. Yes, while a 72x IDE CDROM can *burst* a
B> >*small* file at 72x (after several *seconds* spent spinning up), it
B> >cannot sustain this for long, so the larger the file the *slower* the
B> >*average* transfer rate, down to about 4x for IDE. SCSI CDROM drives
B> >will bottom out at about 8x (the SCSI bus is better than PIO IDE).
B> >Faster drives can possibly be worse -- spin up (several seconds), start
B> >bursting, flood bus, spin *down* (several seconds), wait for bus to
B> >clear, spin up again (several *more* seconds), start bursting again,
B> >flood bus again, spin down *again*, wait for bus to clear, etc.
B> While I am not exactly bragging. My Hitachi CDR-3750 seems
B> faster than the newer IDE drives out there. its a Caddy based SCSI
B> Single Speed unit. This drive is also better built than the free
B> Tatung 24x IDE drive I got! In playing Audio disks. this
B> drive rarely skips. Yet my cheap Tatung is worse than a record player
B> with a bad stylus! Any tiny blemishes or scratchs on the cds (that
B> happens in data too) will send the drive on a skipping fit! whereas my
B> single speeder ignores them! The caddies themselves are a godsend! No
B> need to take the disks out of the cases! Just slide the caddy in! This
B> provides dust protection as well as fast loading. I want a few more of
B> these hitachi CDR-3750s anyone know where I can get more units? 7
B> years ago I paid $99 for it. I really like the drive enough to buy a
B> few more!
B> I have never heard it speed up/slowdown. I believe mine is
B> always spinning. (I never checked). This drive keeps steady pace with
B> the Scsi-2 buss here. Where as the "CRAPSTER" always is doing the
B> "speed Juggle". When it starts it sounds like a turbine. My Hitachi is
B> wisper quiet.
It will be spinning up/down, just that since it is single speed, there
is not as much acceleration time. Also, like you said -- a single
speed drive won't have the bus-flooding problem a 'faster' drive will
have. Also, since it is slower, it probably has a longer spin-up
timeout (how long before automatic spin down when inactive) -- because
the slower speed motor uses less power, does not heat up as much and
the bearings are under less stress.
B>
B> --
B>
B> B'ichela
B>
B>
--
\/
Robert Heller ||InterNet: [EMAIL PROTECTED]
http://vis-www.cs.umass.edu/~heller || [EMAIL PROTECTED]
http://www.deepsoft.com /\FidoNet: 1:321/153
Posted Via Nuthinbutnews Premium Usenet Newsgroup Services
==========================================================
** SPEED ** RETENTION ** COMPLETION **
==========================================================
http://www.nuthinbutnews.com
------------------------------
From: "Mathias Rodenstein" <[EMAIL PROTECTED]>
Subject: compiling glibc and gcc
Date: Wed, 21 Feb 2001 13:08:44 +0100
hello out there,
i could really use some helpe on this since i have been trying to do it for
quite some time. can anyone tell me which steps i have to follow to make my
system work completely after compiling glibc (2.2.2)...i recently downloaded
it and installed on a debian potato r2...then i tried to recompile gcc
(2.95.2 and 2.95.3.test3). both times i get undefined references: saying
something like /usr/local/lib/libc.so.6: undefined reference ...
@glibc2.2...so what do i have to do and where do i have to start...i would
really appreciate help on this.
TiA, M
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list by posting to comp.os.linux.misc.
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Misc Digest
******************************
