Linux-Misc Digest #394, Volume #19 Wed, 10 Mar 99 03:13:13 EST
Contents:
Re: Is my hard drive dying? (Eric Lee Green)
Re: Can Linux use 36-bit Xeon addressing? (Tomasz Korycki)
Re: SuSE 6.0 feelings please. ("Robert C. Paulsen, Jr.")
chown: bug or feature (Ilya)
Re: Is my hard drive dying? (Ted Staberow)
Re: Ghostscript vs Canon BJ10e (Grant Taylor)
Re: Can Linux use 36-bit Xeon addressing? (Gianni Mariani)
Re: How does rpm check dependencies? (Micha� Kuratczyk)
Re: CPU 0.1% idle after X login by root ? ([EMAIL PROTECTED])
RedHat 5.2, Enscript and HP LJ 5L ("Cheng C. Yeh")
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Eric Lee Green)
Crossposted-To: linux.redhat.misc,comp.os.linux.hardware
Subject: Re: Is my hard drive dying?
Date: 10 Mar 1999 03:59:32 GMT
Reply-To: [EMAIL PROTECTED]
On Tue, 09 Mar 1999 23:19:50 GMT, Michael Riffle <[EMAIL PROTECTED]> wrote:
>I go to the datacenter, and the hard drive indicator light is on. But
>I can't pull up a login prompt or anything at the console. Screen is
>blank.
>
>I reboot, and fsck finds and fixes a bunch of things.
>
>I check /var/log/messages and see :
>
>kernel: EXT2-fs error (device 03:03): ext2_find_entry: bad entry in
>directory # 890265: rec_len % 4 != 0 - offset=252, inode=1869902965,
>rec_len=11886, name_len=26983
>
>I actually see this repeated a LOT in the /var/log/messages file.
>It's been going on for 3 days, I just didn't notice. I saw one
>timestamped right when the system locked up.
>
>Anyway, my question is: is my hard drive going kaput? I've actually
>seen file system corruption the last 3 or 4 times I've restarted the
>machine... but I always figured it was because it wasn't shut down
>properly.
The first thing I would check would be to see whether your fan has locked
up. I have seen drives freeze up like this because the power supply fan
went kaput and the hard drive literally baked to death.
If that's the case, the hard drive is probably on its last legs after all
that torture. If not... well... it still sounds like your hard drive
is on its last legs :-(. Sorry I couldn't give you any good news there :-(.
--
Eric Lee Green [EMAIL PROTECTED] http://www.linux-hw.com/~eric
"First they ignore you. Then they laugh at you. Then they fight you.
Then you win." -- Ghandi
------------------------------
From: Tomasz Korycki <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux,comp.os.linux.hardware
Subject: Re: Can Linux use 36-bit Xeon addressing?
Date: Tue, 09 Mar 1999 22:55:34 -0500
brian moore wrote:
>
> On Tue, 09 Mar 1999 15:02:24 -0500,
> Tomasz Korycki <[EMAIL PROTECTED]> wrote:
> >
> > That's called "courtesy", or "attribution". I'd hate for somebody not to
> > know who said what, if they stumbled on this thread only now....
>
> Attribution would be the first line of this message. The .sig is not
> attribution. (And proper news clients will automatically strip it from
> replies.)
Some clients, though, only put "brian moore wrote:". I do not control
what clients people use. I would, nonetheless, like them to see the
attribution retained.
>
> > Hey, You're quick on the draw! Now: what about everything ELSE I said?
> > Does it not warrant a reply? If so, just tell me what am I full of, I'll
> > go away. Not necessarily changing my mind, though. So, if You wish to
> > convince me, PLS reply to the whole thing, not just it's most irrelevant
> > part!
>
> Most of it was irrelevant nonsense. As relevant as your insisting you
> trimmed your posts followed by quoting the .sig.
>
I didn't insist I did trim "the texts", as You put so quaintly. I said I
usually did it. Usually does NOT mean always. I avoid the word. As for
irrelevant nonsense: see below.
> > BTW, I never had the (dubious, I infer from Your previous post)
> > pleasure to work on HP3000, but I've also never heard of RSTS. RSX-11M
> > yes, I still have nightmares, but not RSTS. What was (is) it?
>
> One of several OS's for the PDP-11. Far more popular than RSX-11 as I
> recall, or it certainly was around here (at least on non-Unix PDP-11s).
>
Thank You. I didn't know that.
Now, that's why I deem it discourteous to trim text: I don't think my
previous message was "irrelevant nonsense". You may beg to differ. But
if so, please point out _precisely_ what of what I said WAS, in fact,
nonsense, and please justify it. I like to learn new things, especially
if they contradict my heretofore held convictions. Nothing worse (for
me, anyway) than certainty one's right, when one isn't.
For Your convenience, I will quote my previous post verbatim. PLEASE
break at any point I write nonsense and tell me why is it so.
>>>>>> begin repost <<<<<<<<
brian moore wrote:
>
> On Tue, 09 Mar 1999 01:26:56 -0500,
> Tomasz Korycki <[EMAIL PROTECTED]> wrote:
> > brian moore wrote:
> > >
> > > > Yes. And Your point, as related to "the last holdout from basing their
> > > > systems on Unix concepts is Microsoft" bit? Mind You, if You look deep
> > > > enough into NT architecture, You'll see.... VMS!
> > >
> > > VMS is based on Unix?
> > >
> > > Very interesting news indeed.
> >
> > Never said it was. That was just an "aside". And the explicit answer to
> > my question? I am a little slow on the uptake, as I can't make it out
> > from Your reply....
>
> I didn't think it needed an explanation, since it's clear that IBM is a
> Unix vendor despite also selling OS/400. (Heck, they also sell Windows
> on Aptivas.) Your argument is a red herring. May as well claim that
> Safeway doesn't sell apples because they have milk.
>
Well, allright: IBM makes AIX, which I actually love immensely. But You
walk into almost any bank and You won't see too many AIX boxes (with
exception of SP-2's), or other **IX boxes. What You'll see will be
UNISYS XPM3800 series (love the -82!), S/390, Tandem (and NOT the SGI's
OEM'd boxes). Same in power companies. Same in baggage handling
facilities. And so on.... What does it mean? Not that *IX is worse OS,
just that it's not suitable for some tasks. Like OLTP. What that means,
in turn, is that some companies shoot for that niche. You shouldn't look
at IBM as a one company - they are not. Thay all feed off IBM's brand
recognition and size economies (common sales, etc), but that's pretty
much all. OK, You can call 1-800-IBM-SERV and get service for any IBM
product, be it Aptiva or S/390, but that does not mean You're dealing
with the same organisation in all those cases. In fact the AIX group
(and RS6k group) are completely seperated from, say, OS/2 group (yes, it
still exists). I ,don't care about their common name: they're a separate
company from the S/390 people. Or CPU design and fab people.
Now, that niche also has some specialised players: like UNISYS. They
make their OS2200 and market it quite succesfully, along with the HW it
supports. I looked: they don't see to be making any **IX OS, even though
they DO support their products on various flavours of it.
Correct me if I'm wrong: I thought the definition of "last" included
that there's NO other. Also, Your original statement did NOT include the
"major" quallifier. Which is good, as ho do You define "major".
> The last holdout is Microsoft: every other major OS vendor (including
> IBM, which was unthinkable in 1980, as was DEC) is dealing Unix.
>
See above about the "major" difficulty I have. Plus a vendor is someone
who sells, not necesserily makes, a product. So You seems to be
broadening the domain QUITE a bit... Was that intended?
> Apple, HP, IBM, DEC: all proprietary and wierd OS's of their own 15
> years sgo (remember HP-3000's? or RSTS?). Now with their major OS
> investments in Unix.
>
15 Years ago they weren't weird: they just were. Unix was than "weird"
(by the day's standards), and, generally speaking, a bet on the future
going in certain direction, with VERY little security or ways to
introduce it. That's why "the Enterprise" didn't take to it in
overwhelming numbers.
> The exception: Microsoft.
>
Not the only one, as I'm trying to show. BTW, You seem to be mellowing
Your original statement: it seemed to say MS was the _only_ exception.
Now we have addition of "major", domain broadening with "vendor" and
weakening withlack of "only". Am I to assume "Among companies selling
OSs, Microsoft is an exception in that it doesn't sell any unix-based
OS"? If that's the correct way to read "the last holdout from basing
their systems on Unix concepts is Microsoft", forgive me, as English is
not my mother tongue.
I would also suggest we move to e-mail (if You wish to continue, as
our discussion doesn't have a lot to do with this thread...
>>>>>>>> snip! <<<<<<<<
> Please trim texts when quoting.
>
I do, usually. Unless something seems to me relevant to the subject, or
to someone who might stumble upon a message and wish to know what it is
all about.
>>>>>>>>> end repost <<<<<<<<<<<
------------------------------
From: "Robert C. Paulsen, Jr." <[EMAIL PROTECTED]>
Subject: Re: SuSE 6.0 feelings please.
Date: Wed, 10 Mar 1999 00:18:39 -0600
Keith Davey wrote:
>
> I would be interested in hearing from anyone who has installed SuSE
> 6.0. What do you think of the product? How would you compare it to
> other
> distros like RedHat 5.2 and so forth.
>
> Keith Davey
> [EMAIL PROTECTED]
I recently installed both RH 5.2 and SuSE 6.0 with the intention of
comparing them -- I ended up with SuSE.
SuSE has a much better installation and configuration process -- the
YaST program in conjunction with the SuSEConfig program automate...
1. Initial installation
2. Future additions/deletions/updates
3. Integration of installed apps with the various supported GUIs by
keeping their menus synchronized with the set of apps actually installed
4. Configuration of the various applications and servers. So, for
example, auto dial (with diald) and IP Masquerading are configured for
you.
5. All this is done in a way that doesn't prevent you from tweaking
things on your own.
Of course RH has an automated install process as well, but it automates
far less. For example RH suggested I write down the list of installed
apps so that I could keep track of what was installed myself -- SuSE
makes this unnecessary by managing it all for you.
Configuration of Xwindows is simplified with SuSE's SaX program. It
recognized my monitor and set things up to give me a higher resolution
than either RH or Win98 could muster (still maintaining true-color). I
was able to use the XF86Config generated by SuSE with RH, but would
never have come up with as nice a configuration by myself or with only
RH's help.
SuSE also comes with a larger selection of applications. Just for
starters, RH doesn't have StarOffice, KDE, diald. There are many more.
SuSE comes with a bootable CD -- a huge advantage for those who's
hardware supports this.
--
Robert Paulsen http://paulsen.home.texas.net
If my return address contains "ZAP." please remove it. Sorry for the
inconvenience but the unsolicited email is getting out of control.
------------------------------
From: Ilya <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.development.system,linux.redhat.devel,alt.os.linux,linux.dev.newbie
Subject: chown: bug or feature
Date: Tue, 09 Mar 1999 22:09:35 -0800
I've just noticed that I can't cown my own files to other users under
Linux. Tried it under RedHat 5.2 and Slackware 3.5
Examining kernel sources I found in fs/inode.c, line 307 (function
inode_change_ok):
/*make sure a caller can chown*/
if ((attr->ia_valid & ATTR_UID) &&
(current->fsusid != inode->i_uid ||
attr->ia_uid != inode->i_uid) && !fsuser())
return -EPERM;
If my brain is still worth any thing, this means
"if you are trying to change UID and you are not currently owner of
inode ?OR? you are trying to change UID to something other then it
currently is, and you are not SU, then .....",
that is if you are trying to change owner user and you are not SU, you
are automatically denied access.
I thought it had to be
/*make sure a caller can chown*/
if ((attr->ia_valid & ATTR_UID) && attr->ia_uid != inode->i_uid &&
(current->fsusid != inode->i_uid || !fsuser())
return -EPERM;
"if you are trying to change UID to something other then it is now and
you are not owner of inode or SU, then deny access"
Is it a feature of linux,
or perhaps my understanding of meaning of structures is wrong?
inode -- inode we want to change
attr -- new attributes
current->fsuser -- UID of caller
if I am wrong about that, please correct me.
Thanks.
Ilya.
------------------------------
Date: Wed, 10 Mar 1999 00:26:40 -0600
From: Ted Staberow <[EMAIL PROTECTED]>
Reply-To: tstaber@no!spam.ibm.net
Crossposted-To: linux.redhat.misc,comp.os.linux.hardware
Subject: Re: Is my hard drive dying?
Hi Michael,
Another poster said to check your cooling fan. This is good advice.
A problem I have noticed with some drives is that Linux does not always
see the correct geometry. I have a Maxtor 6GB drive in one of my boxes
that has 826 cylinders, 255 heads, 63 sectors/track (in LBA mode). Linux
sees the correct size but the geometry is very different (16 heads). I
always use fdisk (not diskdruid) because of the (v)erify option. This
checks for mistakes in the partition settings before you commit them to
disk. Make sure that fdisk sees the same numbers that are reported by
your BIOS other wise you will have some strange problems. This is
especially likely if you create more than one partition on the drive. If
the numbers don't match, you can use the 'expert ' options in fdisk to
change the numbers of cylinders, heads, and sectors to match the nubers
in the BIOS. If this does not apply to your setup, you may just have a
bad drive.
--
Ted Staberow
Prairie Networking, Inc.
We Like GNU Ideas.
Michael Riffle wrote:
> Hoy,
>
> Today, I download, gunzip and untar about a 9 meg file. Right after
> it's done untarring, I get a prompt and then the system freezes. No
> pings, no nothing.
>
> I go to the datacenter, and the hard drive indicator light is on. But
> I can't pull up a login prompt or anything at the console. Screen is
> blank.
>
> I reboot, and fsck finds and fixes a bunch of things.
>
> I check /var/log/messages and see :
>
> kernel: EXT2-fs error (device 03:03): ext2_find_entry: bad entry in
> directory # 890265: rec_len % 4 != 0 - offset=252, inode=1869902965,
> rec_len=11886, name_len=26983
>
> I actually see this repeated a LOT in the /var/log/messages file.
> It's been going on for 3 days, I just didn't notice. I saw one
> timestamped right when the system locked up.
>
> Anyway, my question is: is my hard drive going kaput? I've actually
> seen file system corruption the last 3 or 4 times I've restarted the
> machine... but I always figured it was because it wasn't shut down
> properly.
>
> Any advice out there?
>
> Mike
> (don't respond to my email, it's spam proof. please post here.)
------------------------------
From: Grant Taylor <[EMAIL PROTECTED]>
Subject: Re: Ghostscript vs Canon BJ10e
Date: 10 Mar 1999 01:32:26 -0500
[EMAIL PROTECTED] (Matthew Palmer) writes:
> Here's the story - I've got a BJ10ex printer (yes, getting on a bit
> I'm using a command line I got from the Printing-HOWTO, which worked
> for the Epsons, but I can't get it to go with the BJ10. When I try
> printing something, the printer beeps, locks up, I have to
> power-cycle it, then when it comes up it prints part of the
> document, beeps, and so on.
Someone has reported to me that the bj10 works only in Epson mode. I
think there's a dip switch or something?
--
Grant Taylor - gtaylor@picante<dot>com - http://www.picante.com/~gtaylor/
Cellphone information: http://www.picante.com/~gtaylor/cell/
Libretto information: http://www.picante.com/~gtaylor/portable/
Linux Printing HOWTO: http://www.picante.com/~gtaylor/pht/
------------------------------
From: Gianni Mariani <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.hardware
Subject: Re: Can Linux use 36-bit Xeon addressing?
Date: Wed, 10 Mar 1999 06:48:55 GMT
Johan Kullstam wrote:
> Gianni Mariani <[EMAIL PROTECTED]> writes:
>
> > Johan Kullstam wrote:
> >
> > > John Hasler <[EMAIL PROTECTED]> writes:
> > >
> > > > Brian Moore writes:
> > > > > (Again, much of this is due to their non-compete clause that they signed
> > > > > when they sold Xenix off to SCO, so it's unlikely to change.)
> > > >
> > > > It'll change the moment they decide that buying out the agreement is a
> > > > worthwhile investment.
> > >
> > > yes, but as far as i can tell, microsoft are idealogically committed
> > > to destroying unix. i can see bill gates taking off his shoe and
> > > pounding it on the rostrum....
> > >
>
> > Microsoft doesn't even know how to spell Unix :(*) Who are you kidding :)(
>
> > Unix/Linux is alot of fun to use. But, until there is *lots* of cash
> > in it, MS won't take it seriously.
>
> MS do take unix seriously.
OK - how ?
> MS are committed to `windows everywhere'.
> that means, no other operating systems, e.g., unix. MS wouldn't even
> have to say `windows everywhere' is there goal if there were *no*
> competition.
NO, Windows everywhere is the way they see they will make the most
cash. Their committed to their shareholders, employees and customers.
That seems to best translate to Windows unless you have a better idea.
>
> > How do you go to your shareholders (of a $400billion company) and
> > tell them, BTW - we're dropping everything and going to Linux ? The
> > only ideology MS has is capitalism.
>
> why would they do that? every linux install is a rejection of
> windows.
Circular logic ...
>
>
> > When MS sees that it is able to make around $1billion/year in
> > software sales on Linux, you will see them porting stuff. Be
> > careful though, do you really want the "registry", "proc calls",
> > "GDI", "Direct-X" on Linux ? Gee, MS would love it,
>
> this is going off on a strange tangent...
Why ?
>
>
> > you get to support irate Win98 come Linux customers over the
> > comp.os.linux.misc newsgroup and they would have a field day
> > shipping Office for Linux. Oh, and then they would be the evil MS
> > because they support Linux and make a huge bunch of cash on Linux
> > apps but don't support the OS.
>
> > Tell you what: Make a proposal to MS (other than MS get lost) on how
> > you would make *lots* of cash selling Linux based products for MS.
>
> why? i really *do* want MS to *get* *lost*.
There are currently 400 billion reasons why they won't get lost in a hurry.
Live with it !
>
>
> --
> J o h a n K u l l s t a m
> [[EMAIL PROTECTED]]
> Don't Fear the Penguin!
------------------------------
From: [EMAIL PROTECTED] (Micha� Kuratczyk)
Subject: Re: How does rpm check dependencies?
Reply-To: [EMAIL PROTECTED]
Date: Wed, 10 Mar 1999 06:55:46 GMT
David M. Siegel wrote:
>I built the libraries in question from a source tarball. Is there some
>way to get rpm to install the libraries I built?
Yes, of course. RPM source package is a tar.gz file + spec file. Everything
what you have to do is write spec file. If you want to learn this you can
download "Maximum RPM" from ftp.rpm.org or mail me - I can explain you how
to do this :)
--
Micha� Kuratczyk
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.networking
Subject: Re: CPU 0.1% idle after X login by root ?
Date: Wed, 10 Mar 1999 06:49:46 GMT
In article <79oddu$7hs$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (Tom Holub) wrote:
> In article <[EMAIL PROTECTED]>,
> Y W Wong <[EMAIL PROTECTED]> wrote:
> )
> )Why the control-panel take 98% of the CPU resources even I have already
> )logout
> )from a remote X-Term ? ( Resource cannot release )
> )Is it a bug of Linux xdm ?
>
> It's a bug in control-panel; if you log in as root and get the default
> xdm setup (which includes control-panel) and then log out, control-panel
> doesn't handle the exit signal correctly and sometimes goes into
> an infinite loop.
>
> Workarounds include not using control-panel, killing control-panel
> when it gets stuck, exiting control-panel before you log out, or
> fixing control-panel.
> -Tom
>
Yep, Tom is right, it's all control-panel's fault.
There are lots of ways you can deal with this - after all, you can always kill
just the process :)
Here is what I did about it, maybe you'll also find it helpful:
1. In my xdm-config file (/etc/X11/xdm/xdm-config) I changed
DisplayManager._0.reset: to
DisplayManager._0.reset: /etc/X11/xdm/Xreset
Now when I shut down my window manager, xdm executes
/etc/X11/xdm/Xreset
2. Next I created /etc/X11/xdm/Xreset and put 'killall control-panel' in it.
Here is the whole file:
#!/bin/sh
#
# Xreset
#
# This program is run as root after the session ends
#
/usr/X11R6/lib/xdm/TakeConsole
#kill the crazy control-panel
killall control-panel
exit 0
That's it! Now I don't have to remember to kill the process every time I log
out.
Alex
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
Date: Tue, 09 Mar 1999 21:20:40 -0500
From: "Cheng C. Yeh" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware,comp.os.linux.setup
Subject: RedHat 5.2, Enscript and HP LJ 5L
Hi,
I was wondering if anyone had ran into this problem and found a
solution. I have a perfectly good running system based on Red Hat 5.2
with SMP, 2.0.36 kernel; recently upgraded to 2.2.2 kernel.
I have an HP LaserJet 5L printer connected to /dev/lp0 and I am using
the LaserJet 4/5/6 printer filter. Everthing works fine, I can print
with lpr and stuff. However when I tries to print with enscript, the
printer doesn't eject the page. I can see that data are being sent to
the printer because the light is flashing but no matter how many pages I
print, I don't get any output. After print job is complete, I press the
form feed button and all I get is one page, the first page.
Has anyone else ran into this probem?
Thanks,
Cheng
------------------------------
** 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 (and comp.os.linux.misc) via:
Internet: [EMAIL PROTECTED]
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
******************************
