Situation improving. Ben suggested that something like this:
yum reinstall $( rpm -qa --qf '%{name}\n' )
...would be one way to refresh the on-disk images along with the
corresponding MD5 sums in the RPM database, but some kind of RPM
dependency hell (maybe just in my situation?) made
Have you done straight (non-hashed) content-comparison of any of
these files? Are they actually gratuitously different in content,
or are they just truncated on one system? MD5sums are effectively
dependant on file-size
The files were the same size but differed in various locations
I'm looking at some supposedly identical CentOS5.3 systems that are
behaving strangely and while grasping at straws I generated lists of
the MD5 sums of all the files on the root partitions and I'm seeing
differences in the on-disk images of things like /sbin/mount and
/lib64/libblkid.so.1.0 that
I'm running an rpm --verify --all pass on those machines right now and
it's showing quite a few indications of unexpected differences based
on the info recorded in the RPM database. Ben is right; that's a very
nice feature of RPM. When I captured the output in a file and then said
things like
It pains me to say anything that appears to cut ComCast any slack
because I have no love for them whatsoever but, FWIW, I'm seeing
essentially the same traceroute output reported by Kenta:
e521:~ 395--- traceroute linuxquestions.org | lineup
traceroute to linuxquestions.org
DTVZ wrote:
I thought you guys might have some input on this re: usability and
features... any suggestions for improvement would be appreciated.
http://sensatronics.com/index.php/demos/senturion-demo.html
___
gnhlug-discuss mailing list
The web site and phone number Google finds are both no longer
in service.
http://www.lacarretamex.com/
I'll be out of town around then but I still have GNHLUG
emails from mid 1995 so I'll be with y'all in spirit...
___
gnhlug-discuss mailing
Eric Raymond can be a PITA but in this case his recommendations (or,
more to the point, the principles behind them) are refined and spot-on;
they're applicable in pretty much any situation involving interactions
among multiple (smart) parties, not just online mailing lists...
On this busy morning I've only had time to glance at some docs for SVM
(Secure Virtual Machine) support but it does appear that in some cases
external hardware (in the form of a TPM - the dread Trusted Platform
Module) can be involved in the prep and execution of the Secure Loader
and,
I did not want to eat up people's time with this thread.
This thread is interesting and something that I've been meaning to
learn more about so I was pleased to have an excuse to dig an old
CPU manual out of my desk midden.
___
gnhlug-discuss
There was a study published a couple years back that showed
enabling the VT instructions can result in lower performance
Heh. The x86 instruction set offers some fancy instructions that are
supposed to help you implement an OS by doing (in one swell foop) some
fairly involved stuff like
Seriously:
http://www.unitednuclear.com/magnets.htm
...and you'll have some fun, if dangerous, toys to play with, too.
I don't know that they're officially represented as being appropriate
for your specific purpose but some of them are so dangerously powerful
you can find pictures on the
I have fairly deep OS-level experience (including some Virtual Machine
work) but I confess that I'm not up on the very latest VM technology
so to further the discussion let me ask something that may also have
occurred to others:
What is it in the nature of VM support in these processors (or
So if VM support is enabled by flipping some bit(s) in some CPU
Control Register(s) I'd assume that a VM-capable OS could flip those
bits as well as any BIOS code. I suppose it's possible that the CPU
might first insist on seeing a certain logic level on a certain input
pin before
As an experiment, can you link a helloWorld-style object (that needs
the symbols in question) against the libs in question?
In other words, arrange for that memcpy_cell(void*,blah,blah) to be
unresolved in your helloWorld object and then link it against the object
or lib you think should be
What flavor are the libs in question? If you're generating x86_64
objects you can't link against i686 libs and vice versa, etc, etc...
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
In just the few minutes I had to actually look at that pack2.c I found
several scary sequences of code, so if you're staying your current course
(instead of adopting some of the other recommendations offered here)
you might want to keep on looking or else write your own from scratch.
For
Any suggestions to look for a 2^53 type problem?
Well, just for fun, how about going back to basics - what does this
little program generate on all the systems in question?
#include stdio.h
main( int argc,
char *argv[] )
{
printf( sizeof( double) %2u\n, sizeof(
Ooops - I forgot about the void * ...
Any suggestions to look for a 2^53 type problem?
Well, just for fun, how about going back to basics - what does this
little program generate on all the systems in question?
#include stdio.h
main( int argc,
char *argv[] )
{
printf( sizeof(
FWIW: I work at a site where we share a building with an ISP/colo vendor
with 3 big diesel generators (each in its own shipping-container sized
enclosure and so powerful they require an automobile-sized resistive
load to dump their juice into during testing) out front plus at least one
slightly
Most of our HP systems (all recent models like xw8600 and z800) refuse
to eject their optical media when the drive button is pushed after a
warm reboot following a rescue/install session booted from that drive.
Have you reported this to HP as a BIOS bug?
Yes, working that with them now.
Most of our HP systems (all recent models like xw8600 and z800) refuse
to eject their optical media when the drive button is pushed after a
warm reboot following a rescue/install session booted from that drive.
Other systems (eg. Dells) behave as we'd like when booted from those
same discs so
Is there some trick (maybe some kernel commandline option or some program
executed during shutdown) that will leave the drive willing to eject the
media (unlock it?) without us having to power-cycle these machines?
Why won't the paperclip trick (manual release) won't work?
Because that
You could do like that character in Cryptonomicon (a good read, BTW) who
was imprisoned in what he assumed was a TEMPEST-instrumented jail with
his laptop, so he rigged it surreptitiously to do I/O via Morse code... ;-
___
gnhlug-discuss mailing
I wonder if the CentOS (Community ENTerprise Operating System)
founders knew about this or if it's just, like, ya know - kosmic:
e521:~/codeGen 601--- dict cento
1 definition found
From The Collaborative International Dictionary of English v.0.48 [gcide]:
Cento \Cento\, n.;
I was going to write up a description of my traceroute investigations
into ComCast's DNS hijacking when I found a very similar writeup here:
http://slashdot.org/submission/1052907/Comcast-Hijacking-DNS-wMicrosofts-Help
...with add'l info here:
At least you can opt-out now via a form presented on the page.
Grumble... well the actual opt-out page is here:
https://dns-opt-out.comcast.net/
...and just for a bit of ironic fun I wondered what would happen
if tried the www. version of that hostname, thus:
oh, wait. That would be better implemented as a USB device,
wouldn't it? That way your smells would be mobile, and the
device could be of arbitrary size.
Weirder people than y'all (yes, they exist) are way ahead of you...
http://www.google.com/search?q=usb+aroma+therapy
Anybody know anything beyond what's mentioned in this open letter
to CentOS's Lance Davis signed by a number of key CentOS players?
http://www.centos.org/
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
Anybody know anything beyond what's mentioned in this open letter
to CentOS's Lance Davis signed by a number of key CentOS players?
http://www.centos.org/
Ah. Some further info here:
http://www.h-online.com/open/Growing-unrest-within-the-CentOS-project--/news/113889
Yes, I'm a space junkie! :-D
I've liked the Moon Machines series:
http://www.youtube.com/results?search_query=moon+machines
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
Anybody have any experience with the Phoronix test suite?
http://www.phoronix-test-suite.com/
I've seen some praise for it on the WWW but so far my impression is
not good. I've pulled both the 2.0.0b2 and the 1.8.1 tarballs (they also
offer a .deb but no RPM) onto a fairly standard
I use and like the User Agent Switcher extension for Firefox but
I don't think it'll help in this case. I say that because I find
that Firefox on Linux can talk to my employer's Exchange server if I
instruct it to ID itself truthfully, but if I tell it to lie and ID
itself as Internet Exploder
YMMV with an agent switcher. If the site uses the agent to
determine what code to feed (javascript, ActiveX) like it's supposed
to, it won't work. If it uses the agent to say we don't work on
your browser even though it's generic HTML, it won't work.
We presume you meant that last one to
Another example of cool content that's only viewable if you're
willing to jump through a bunch of irritating proprietary hoops:
http://tierneylab.blogs.nytimes.com/2009/07/15/gates-puts-feynman-lectures-online/
___
gnhlug-discuss mailing list
I confess that I'm happy to use free wireless on occasion but I worry
that if I make my own AP similarly available then somebody is going to
use it to post kiddy-porn or make threats and it'll be traced to my IP
address and people with guns will then make me spend a lot of time
money trying to
Hey! cool - if this FUD approach is so effective maybe we can
use it to rid the world of some other scourges. Like what if we
very coyly insinuated that there *might* be one or two flaws in
Microsoft Windows[...]
It hasn't worked agains MS yet...
Right - that was my (possibly
I'm not sure how widespread it is, but I know that ANHosting
(MidPhase) is blocking it entirely. And they've got no ETA for
when they'll put it back so far. I guess they're waiting for
details and patches about the exploit to be released... ugh.
HostGator has disabled OpenSSH support
I am running into a disk space issue on an older server. I'd like
to do a tune2fs -m 1 (or maybe 0) to get rid of most, if not all of
the reserved block space on the partition that is close to full.
The disk is actually an iSCSI volume mounted from an EqualLogic
array, and then exported
Does anyone know of a tool that can automatically update symbolic
links when moving files around on a filesystem, so as to maintain
symlink consistency?
rsync has options to deal with symbolic links (and hard links) however
you'd like it to
I think the complication he's dealing with is that
I wrote:
His problem is that he's moving relative symlinks (or their
referents) around such that the symlinks no longer point at the
right thing and he wants something that detects that and recreates
those symlinks such that they still point at the right things when
the dust has settled.
I notice that there is no swap listed. Umm, how does one add swap to a
nfs based system?
NFS swap of course ;-)
Have any good references on NFS swap?
To see if swap is even going to help you you might:
# Create an empty 1Gb file
dd if=/dev/zero of=/someNFSdirectory/mySwapFile
I wrote:
If no joy just delete that swapFile,
Yikes! I hope it was obvious but I forgot to say that you should:
swapoff /someNFSdirectory/mySwapFile
...before deleting it.
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
I had a look at /sbin/rescan-scsi-bus.sh - it's not the most elegant
bash script I've seen but it apparently works - cool. I've known about
the add-single-device trick but it's clunky and there have been times in
the past where it was easier to just reboot to get the kernel to notice
a new
[ Not Linux-specific but likely of interest to some on this channel ]
A blow-by-blow description of how a secure WWW session gets setup:
http://www.moserware.com/2009/06/first-few-milliseconds-of-https.html
___
gnhlug-discuss mailing list
Two possibilities come to mind: grepping for : funcName and
storing funcName in $_. Grepping for : funcName might be an easy
way to find where a function is defined if the source is spread
out among many files and none of the files contain any comments.
Plausible, though that idiom isn't
There are several ways to get GCC to supply the info in question along
with a blizzard of other stuff:
- Ask it to be generally verbose by adding the '-v' flag to the command line.
Maybe not enough info for your purposes but often useful
nevertheless.
- Ask it to mention each
http://www.facebook.com/home.php#/photo.php?pid=30037422id=1223007473
Facebook intrusively demand that I create an account and provide
contact info even if all I want to do is view the image you've
invited me to see. I have no plans to create a Facebook account
so I wonder if there's another
I'm surprised they let through HACKER with all of the criminal
rep that name has unjustly acquired.
I have no idea how DMV personnel decide what's acceptable on vanity
plates but I saw 455H01E not long ago so I guess they haven't
perfected their filtering techniques.
My car has my official
Heh. It's funny now, looking back on it, but I experienced
several minutes of panic this morning as I connected to a remote
system via SSH and discovered that I was unexpectedly able to
connect back to the originating system at will without mentioning
any password. It was definitely a WTF
My Debian system's tar identifies itself as tar (GNU tar) 1.22 and
the output generated with --help includes the following excerpt where
that --no-overwrite-dir option sounds like what you wanted:
Overwrite control:
-k, --keep-old-files don't replace existing files when extracting
I want to search a text file for a few (alphabetic) words which
must be near each other, but not necessarily on the same line.
grep is pretty much line oriented and although it's possible to script
elaborate workarounds involving transfers back and forth between the
pattern space and the hold
Offered are some random computer components, including: half a dozen mice
(some USB and some PS/2), PCI 10/100 enet adapters graphics adapters,
a parallel printer cable, several IDE CDROM drives, etc. In North
Chelmsford near Drum Hill. Please take the whole lot if interested.
Thx,
--M
Data point [...] it was asserted the girl's spotless discipline
record was not proof she was a good student, but only that she had
not been caught violating school rules. Consider the implication
Considered. The implication is that the presumed innocent dictum
was ignored in that case and
one of the things that I am enthusiastic about: set -e
I like that one, too, and I also like the way bash allows
you to trap on error conditions, thusly:
#!/bin/bash
function errHandler() {
echo 'errHandler activated'
}
trap errHandler ERR
echo Ready for first
the basic echoing of commands that I know is out there some where
would make my scripts way more professional. Anyone know that?
Not quite sure what you're referring to but between bash's -x and -v
options you will probably get as much chatty output as you could ever
want. If you remember
Is there any way to stop init partway through, so I could at
least see if unreasonable amounts of memory seem to have been
used, or look for other information?
I've occasionally had to engage in this sort of hackery when debugging
b0rken init logic and such. In the script of your choice,
Thanks for the responses.
Tom Buskey wrote:
Shutdown cleanly so your system doesn't have to fsck.
OMG! And all this time we've been instructing our customers to just pull
the AC plug from the wall when they're finished using the systems... ;-
Yes, of course, clean shutdowns are to be
He is going to copy a disk image to another machine and see if
things work there.
That'll be a good test but it'd be easier (circumstances permitting)
to just move the physical drive temporarily to another machine.
You don't even care if that other machine can completely boot
all the way up
our customers do find themselves occasionally needing to (re)start
systems in time-critical situations
Or they *think* they need an emergency reboot. This isn't windows.
Ahh, but I love the way Ctl-Alt-Del gives That Fresh Feeling(tm)!
With few exceptions our customers are Windows users
I don't think it was ever considered a design feature that one
could/should interrupt fsck.
Indeed, I concluded after my first (shallow) investigation that it
couldn't possibly work. Just for starters, IIRC, the kernel's console
driver doesn't even do job control or signal delivery. It's a
In certain time-critical situations it is desirable that we be able to
interrupt fsck as it tries to preen certain huge filesystems. Yes, we
know that interrupting fsck is not good sysadmin hygiene and we generally
discourage such behavior, but when a machine is being (re)booted in a
crisis
FWIW:
http://apcmag.com/new-worm-can-infect-home-modemrouters.htm
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Now, don't everybody all crowd in at once trying to be the lucky
person who snags this freebie, but I'm offering a US Robotics
Sportster Vi external FAX/modem unit, pretty much new in box
(most items still in shrinkwrap) to the first taker...
___
Free - never used - box still shrinkwrapped:
http://www.usr.com/support/product-template.asp?prod=3cxm756
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Both items have been claimed.
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
This sounds similar to the problems you'd see if you did this:
mdadm --create --level=1 --raid-devices=2 /dev/md0 /dev/sda1 /dev/sdb1
mkfs.ext3 -L HiMom /dev/md0
...and you then prepared an fstab line that referenced that device by
its label, thus:
LABEL=HiMom / ext3 defaults 1 1
I wrote:
The reference to that label in the fstab confuses the kernel
because of the (apparently) duplicate labels it finds in the
mirrored partitions.
Come to think of it, I'm not sure it's kernel code that's getting
confused. Maybe it's the mount command or some library code -
whatever;
feed rollers on some scanners and printers. Basically, they
get glazed and cannot pull the paper. Cleaning the feed rollers
helps sometimes. Typically I use alcohol to clean them and then,
if I'm still having a problem, a very, very mild abrasive
I'll second that and as an aside I'll
find . -mtime 60
Doesn't mtime's arg mean that many days ago? So you're asking find
to mention files 60 days old, I think. Is that what you intended?
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
find . -mtime 60
Doesn't mtime's arg mean that many days ago? So you're asking find
to mention files 60 days old, I think. Is that what you intended?
I actually should have said, So you're asking find to mention files
last modified 60 days ago
Anybody have a contact at Google or T-Mobile that
might be useful in acquiring (purchase/loan/gift)
multiple live cell-capable Android handsets for app
development? I know somebody who says he's tried
via the normal channels and gotten no love...
___
In light of the occasional questions here about sed and awk this
article seems worth a mention:
http://www.osnews.com/story/21004/Awk_and_Sed_One-Liners_Explained
...as it in turn mentions these:
http://www.pement.org/awk/awk1line.txt
You will see the solution as soon as you hit Send
Rich Hall (author of Sniglets) defines the Onosecond as, the interval of
time after you strike the Enter key before you realize what you've just done...
___
gnhlug-discuss mailing list
http://www.1234567890day.com/
http://coolepochcountdown.com/
An NPR radio program (Day to Day) discussing the event was
broadcast on Friday and this followup program will apparently
be heard today at 3pm:
Unix: An Operating System, Not A Timekeeper
by Steve Proffitt
Day to Day,
http://www.1234567890day.com/
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Some easily digestible writeups about memory management,
CPU caches, the boot process, etc:
http://duartes.org/gustavo/blog/
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
But if you change the macro to:
#define b0rken(x) ({return x;})
you will find that the compiler likes it again.
I mistakenly interpreted your statement that the two constructions
are equivalent:
What I see confufsing is:
do { ... } while(0);
[...]
{ ... } would be
The bottom line is that in the section of code you presented,
what was important was to establish a block so a variable could
be defined. Adding the do statement while(0) is just adding
some extraneous code that would be most probably optimized out,
but even if it is not, it is in an error
I'm not sure if kernel printfs are enabled in production kernels.
I forget how they are configured, so it is possible that no-one
will ever see the message.
They're enabled - that's what drew my attention to that code
in the first place. The 32bit version of libraw1394 apparently
uses
OK - I'm seeing stuff like this the following in some kernel
syscall handling code and it's making my brain hurt, so I hope
somebody can explain it:
.
.
.
static int mt_ioctl_trans(unsigned int fd, unsigned int cmd, unsigned long arg)
{
mm_segment_t old_fs =
There is no such thing as Uninitialized static. All static
variables in C are initialialized by default according to the
C standard.
In the case of an int, it is initialized to 0. In the code below,
it is printing only the first 20 times mt_ioctl_trans() is called
with an invalid
Jarod Wilson wrote:
Nope, in the kernel, all statics are initialized to zero
Yes. Right. Sheesh, thanks a bunch, guys but I get the CompSci101
stuff (I've *written* compilers and kernels) I just ask questions
like these in public to keep discussion flowing, and I remarked
about that
What I see confufsing is:
do { ... } while(0);
What this means is to go through the loop once. You need a leading
curly so you can set up counter as a local variable as variable
names are block scope. { ... } would be equivalent to above.
They're definitely not equivalent - that's
Ben wrote:
sed s/.*:\([[:xdigit:]]*\)\\.*/\1/
That looks good to me, though I assume he meant to show that
expression in single quotes. Also, I can't remember if those
character class notations count as Extended Regular Expressions
but, if so, some versions of sed might want something like
[ Dang! I was naively hoping I'd be able to post this before too
many folks had a chance to respond to my message from last night. ]
Executive Summary: I've changed my mind (sort of)
I shot from the hip after seeing that article about Facebook within
minutes of seeing that other GNHLUG
It just so happens that Facebook was where I first noticed the
issue of the UNIX time event coming up, and I forwarded that link
in case others might be interested in the source.
OK, I understand that (now) but it wasn't obvious from your message,
which appeared to be an invitation directly
Offered is an IBM CDPD Cellular Modem (P/N 04H7387) PCMCIA adapter
with two anntenna+cable assemblies with mounting brackets.
Believed to be working when put into storage back in 2004 but
the batteries (four AA in each antenna module) leaked a bit so
the condition of the two anntenna modules
Facebook tells me I need to sign up to use their service. I'm not
really thrilled with the idea
Me, neither - Facebook would seem to be a poor choice as a venue for
conducting GNHLUG business as (it seems to me that) Facebook is not
really in keeping with much of what the GNHLUG is about:
Jarod Wilson wrote:
[...] which firewire is this? The (really crappy) old ieee1394
stack, or the newer firewire stack? RHEL5.2 shipped w/the newer
one as a still somewhat immature tech preview. 5.3 will be much
improved (but no clue if its improved in any way that would help
you out).
postfix/smtp[11991]: 3C4A1918124: to=michael.odonn...@comcast.net,
relay=smtp.comcast.net[76.96.62.117]:587, delay=0.39,
delays=0.01/0.02/0.33/0.04, dsn=5.1.0, status=bounced (host
smtp.comcast.net[76.96.62.117] said: 550 5.1.0 m...@e521 sender rejected :
invalid sender domain (in
BTW, I forward what very little spam I get to
missed-s...@comcast.net, as well as s...@uce.gov.
H, it may have just been coincidence but my nastygram
came immediately after I sent some SPAM to that missed-spam
address, so I wondered if there wasn't some connection.
I understand the UCE
Paul Lussier writes:
set up sftp in a chroot environment [...] Centos 5...
If anyone knows of any gotchas or tricks I could pass over to him
I STFW thus:
http://www.google.com/search?q=sftp+chroot+centos
...and got this:
http://librenix.com/?inode=11442
...which, being entitled
OK - this is at least a minor case of WTF - why is it that
the list of active file descriptors shown when {t,}csh is
running shows no descriptor lower than 15 ?!?
e521:~ 711--- /bin/bash
e521:~ 513--- echo My PID is $$, contents of my /proc/$$/fd
follow... ; ls -l
Because on startup /bin/csh dups stdin/out/err to higher fds...
[...]
this means /bin/csh uses FDs 16,17,18 as it's stdin/out/err instead
of 0/1/2.
[...]
when running a program: after clone(), it just dup()s back to
stdin/out/err before exec():
Cool - thanks for the
We're trying to run a 32bit app on 64bit CentOS5.2 (RHEL5) systems and
the 32bit version of a certain library is dying horribly because it
disagrees with the kernel module it's interfacing with about the layouts
of various structs. I can't believe I'm the first person to deal with
this sort of
I know of a few compilers that support a structure packing pragma
that can be used to control how the structure is packed/padded.
Googling gcc pragma pack yielded the following link which seems
to explain it fairly well.
http://gcc.gnu.org/onlinedocs/gcc/Structure_002dPacking-Pragmas.html
Googling gcc pragma pack yielded the following link which seems
to explain it fairly well.
http://gcc.gnu.org/onlinedocs/gcc/Structure_002dPacking-Pragmas.html
Ha! I was reading that very section in a freshly downloaded PDF
Dang - it looks like those pragma directives are for solving
I have no clue offhand what's up, but of curiosity, which firewire
is this? The (really crappy) old ieee1394 stack, or the newer
firewire stack?
The app in question is a 32bit-compiled app that links with
libraw1394 so I guess that means it's the (really crappy) old
ieee1394 stack. We
I note that RHEL3 kernels seem to be unhappy (griping about
max inode count or some such) when asked to mount a v1.39
filesystem created using RHEL5 but the RHEL5 kernels mount
the older filesystems without complaint. We're in the process
of moving some users forward from RHEL3 to RHEL5 so I
201 - 300 of 944 matches
Mail list logo
