After being disappointed by the CentOS5.2 LiveCD
kit I'm pleased to report that the Gentoo-based
SystemRescueCD does not suck. I found it remarkably
easy to customize to our needs and the same CD
(well, DVD in our case) can boot either a 32- or
64-bit kernel (particularly useful in our case)
In a previous thread I asked for recommendations and
somebody said:
For customizing CentOS, you might want to look at Revisor,
which is a tool for creating custom installers and live media
for Fedora, which also has CentOS support.
Revisor: http://revisor.fedoraunity.org/
...so I wanted
find . -maxdepth 1 -type d | sed -e s/..// | while read f
do
echo /usr/bin/7z a -t7z -mx=0 ${f}.7z ${f}
done
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
This has been discussed here in the past so good info can
probably be found in the archives (RTFA! ;- ) and IIRC
the recommendations usually come down to:
- Connect the drive to a different system, either directly
or using a USB converter, to determine if the problems lie
with the drive
One of the GNHLUG members runs a computer store:
http://www.justworksnh.com/blog/
I don't know if it meets your requirements but there's a
thread about it in the April archives on the GNHLUG server.
___
gnhlug-discuss mailing list
Does anyone know if it is possible to add user-created
custom tags to RPMs when creating them?
I don't quite understand what you're asking for but you
can cause rpm to generate a list of all possible tags thus:
rpm --querytags
...some of which can (I believe) be specified in .spec
file
Please don't transmit email encoded as HTML - thanks!
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
I'm working on a product that's currently based on RHEL3 and
whose installer is based on the venerable Timo's Rescue CD kit.
We've ported our product to a more current version of Linux and
when I went to port the installer I discovered that development
on Timo's kit stopped after 2004 and it
The CentOS5.2 LiveCD-creator kit looked promising until we found
bugs that prevent it from running on our bleeding-edge hardware.
I'm curious. Got time to provide any details? In particular, why
would the regular CentOS install disc work for you when the live
disc would not? Aren't they
Note to self: Avoid CentOS live discs.
...or at least be careful: I don't claim that *all* discs
created using that LiveCD-creator kit are b0rken and (to your
point in a previous msg) the standard installation CDs and DVDs
do seem to work just fine. But the ones I generated in-house
WTF is /sbin/loader ?
It executes early during startup when booting from an ISO. The
only stuff I can find online about this apparently pivotal item
is lots of other people also asking questions, but no answers.
Ah. I used strings on the /sbin/loader file from the ISO
and divined that
Since some of those parititions might be (as they are in our case)
components of a software RAID and since the act of mounting them
causes uncoordinated modfications to the mirrors behind MD's back,
we're less than pleased with this helpful behavior.
[...]
Any idea how to recover those
http://www.collegehumor.com/video:1886349
ObLinux: the very last line...
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Note, only valid in bash, not sh.
Yah, and it doesn't matter if you use #!/bin/bash at the top
if you're running the script with sh foo.sh. It took me five
minutes to figure that out just now. D'oh.
That process substitution facility is a favorite of mine but
it isn't POSIX, and IIRC
http://www.buy.com/prod/cavalry-1tb-dual-interface-usb-2-0-esata-external-hard-drive/q/loc/101/205986373.html
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
FWIW, NetAPP's WAFL and Sun's ZFS are COW file systems. COW makes
it easy(ier?) to do snapshots.
I (think I) get how a COW approach is supposed to work; when
you decide to take a snapshot you throw The Big Switch and
from then on all writes to blocks on the device in question
(where the
We may want to characterize some client/server systems and
applications (destined to be geographically distant from each
other) in the face of various LAN/WAN faults/conditions like
dropped/duplicate packets and varyious end-to-end throughput
rates, so I'm wondering if anybody here has a pointer
It sounds like you are looking for a network impairment generator.
Sounds like a perfect description of ComCast... ;-
If you are looking for F/OSS running on Linux, then take a look
at NISTNet ( http://snad.ncsl.nist.gov/nistnet/). It is beta,
and has been for a very long time, but it
You say you were able to mount that system's disk
on another system? In that case I think it might
also have been possible accomplish it thus:
chroot rootMountPointHere
passwd root
...at which point the passwd utility would have
allowed you to enter a new password and updated
all the
minicom works great and is not by nature modem-specific,
though the default config does have lots of modem init strings
all helpfully rigged to spew a bunch of stuff down the wire
to confuse hell out of whatever it is you're talking to.
Hand editing of minicom config files is generally not
cheap-N-nasty
Here (IIRC) is how I turn a machine into a serial-port server:
stty 115200 raw /dev/ttyS1
socket -l -s /dev/ttyS1 /dev/ttyS1 21
...after which anybody who telnets to port on that
machine is automatically connected to /dev/ttyS1.
/cheap-N-nasty
I just noticed that for any of my email folders Thunderbird will
report an URL of the form:
imap://[EMAIL PROTECTED]/pathName
...and I wonder if there's a Linux tool like wget that I can use
to pull email folders and such from an Exchange server, hopefully
into a local hierarchy whose
I share the author's grudging admiration for the
sophistication of this Windows malware:
http://www.theregister.co.uk/2008/08/22/anatomy_of_a_hack/print.html
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
dd methods really don't work that well, at least not wth a suspect
drive.
dd generally works with healthy drives but I've not had great
luck with it when even one sector is bad, so I've been using
dd_rescue (Debian pkg is named ddrescue, RPM equiv unknown)
which has features intended for such
Another thing to check is /etc/localtime which should be
(either a symlink to or a copy of) a timezone data file
such as are typically found under /usr/share/zoneinfo,
like maybe /usr/share/zoneinfo/right/SystemV/EST5EDT
___
gnhlug-discuss mailing
http://rivercoolcool.spaces.live.com/blog/cns!D6F05428A2B8CB48!1570.entry
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
It looks like the partition of interest might
be /dev/sdb2 so to mount it read only, try this:
mkdir/tmp/recovery
mount -oro /dev/sdb2 /tmp/recovery
cd /tmp/recovery
ls -la
...etc, etc.
___
VFS: Can't find ext3 filesystem on dev sdb2
Do we know that there was ever an ext3 filesystem there?
Was it maybe something like Reiser or xfs?
The fact that LVM was in play may indeed be a factor here
(or it might simply be the case that the superblock is scrogged)
but my LVM-fu is poor. If
I LVM'd the bigger disk (300GB) onto the base system disk (80GB)
Uh-ohh. As I said, my LVM-fu is poor but if the result of
doing that was to create a block device whose aggregated size
was ~380Gb and you then built an ext3 filesystem in that device
from which all but the first 80Gb are now
create a block device whose aggregated size was ~380Gb and you
then built an ext3 filesystem in that device from which all but
the first 80Gb are now MIA, I suspect you'll not be able to recover
very much... :-/
I think your understanding may not be correct. The 300 GB drive
is fine. I
I've found CONFIG_X86_TSC=y in my kernel config file, but
how do I verify that's the correct parameter or not?
grep -lr CONFIG_X86_TSC /usr/src/linux
Yes, CONFIG_X86_TSC is probably what you're looking for.
If you want to have the option presented to you when you say
make menuconfig
But where did this patch come from? Are you sure it's right?
In particular, I'm looking at the bool and default lines...
I created the patch. Yes, I'm sure it's right insofar as
it has the (narrowly defined) intended effect of providing
control over the config option in question where none
the amount of memory reported is not equal to the amount of memory
physcicall in the system. Since this is a known kernel bug
Which kernels/platforms/situations does this bug show up in?
Can you point us to a description somewhere?
___
I'm definitely looking forward to giving the newly free ESX 3i
a try but, FWIW, I found this guy's experience of interest:
http://blogs.zdnet.com/perlow/?p=9192
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
I'm trying to persuade an RHEL3.9 PAE kernel running on a machine
with 4GB RAM to let me create a ~1Gb ramdisk and the results
are, um, vexatious.
I start the kernel with the requisite ramdisk_size=nKb where I've
tried various values of n like 100 and 1048576. After the
kernel boots I say:
OK, the ramdisk story is looking a bit better now.
First, the tmpfs trick is definitely cool and worth remembering
but isn't what we're looking for because some of what makes it
cool also makes it unsuitable for our purposes. In particular,
tmpfs has close ties to the kernel's buffer cache so
shotsInTheDark
- Copy /bin/bash from another system into /tmp and test
that one? Maybe also some of the shared libs it uses...?
- Temporarily put all startup scripts bash might be using
somewhere it can't find them? If problem goes away,
reintroduce scripts one by one...
-
This is just a nit but I've never understood the meaning/usage
of the $_ variable in bash. I did RTFM and it says:
When bash invokes an external command, the variable $_ is set to
the full file name of the command and passed to that command in its
environment.
...which seems to describe
The filesystem code sometimes forbids hard links for directories;
I think ext3 is one example where it's not allowed.
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
OK - I now believe all observed behaviors are explainable
if that passage from TFM:
When bash invokes an external command, the variable
$_ is set to the full file name of the command
and passed to that command in its environment.
...is augmented thus:
Internally, bash uses $_ as
The _ variable appears to be overloaded... with different meanings
in different circumstances. Qutoing the bash man page:
_ At shell startup, set to the absolute file name of the shell
or shell script being executed as passed in the argument list.
Subsequently, expands to the last
For added fun, try this series of commands:
echo foo
echo /$_/
And given what's been learned (ie. that $_ and !$ are
equivalent in this context) it should now be no surprise
that the same result is seen with echo /!$/
___
gnhlug-discuss mailing
Full disclosure: I hate Comcast.
You're not alone. The Worst Company In America competition
at consumerist.com is now down to CountryWide vs. ComCast:
http://tinyurl.com/65wcns
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
Does the GUI:
http://localhost:631/printers/
...provide you control over the option in question?
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
If it were me I'd first dd the whole flash drive (assuming
it'd be practical, size-wise) into a file and operate on that
file via loopback rather than on the original, just in case...
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
ps auxww truncates command at 4096 characters,
[developers] want to see the entire thing...
If they know the PID in question this might work:
xargs -0 /proc/$PID/cmdline
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
names of my two sata hard drives had been interchanged.
[...] I had apparently bumped one of the sata data cables
enough to get a bad contact
Ah. Yes, that can be a real joybuzzer since the SATA drives
are presented via the SCSI midlayer and thus use the /dev/sd*
naming scheme. If your
If they know the PID in question this might work:
xargs -0 /proc/$PID/cmdline
Apparently /proc/$PID/cmdline only holds 4096 characters as well
(on RHEL5 at least).
Now that you mention it I think that might be a hard
system limit. Can you demonstrate that the command line
as
In kernels built from the kernel.org 2.6.18 kernel sources
you can control whether and where the User mode VM is split
between User/Kernel. Config directives like CONFIG_VMSPLIT_3G
come into play and are visible in the config files.
But when I look in the config files on a machine that just
had
excerpt from arch/i386/Kconfig:
--
choice
depends on EXPERIMENTAL !X86_PAE
prompt Memory split if EMBEDDED
default VMSPLIT_3G
--
Note the !X86_PAE part. At least above, you mentioned looking in a PAE
kernel config. :)
The default kernel appears to have PAE configured
/proc/meminfo is incorrect. I'm trying to find a way to
determine whether the amount reported is correct
Ick. You might be able to write (or find) something that'll use
DMI data to compute a total from everything currently marked as
being a Memory Device (type 17). The dmidecode tool can
it might be enough to compare with the output of free(1)
According to an strace I just ran on free:
.
.
.
open(/proc/meminfo, O_RDONLY) = 3
lseek(3, 0, SEEK_SET) = 0
read(3, MemTotal: 2010264 kB\nMemFre..., 1023) = 728
.
my first guess would be that you mistakenly pointed K3b at your
hard drive instead of the DVD drive.
I'd think the damage would be much worse if that were the case,
so my guess is that some b0rken kernel code (or maybe a HW
problem) corrupted some memory that happened to correspond to
I would definitely recommned booting from external media (DVD)
and checking *all* the filesystems. There may be undetected
corruption elsewhere on your disk. And you can't check the root
filesystem of the running system.
Good advice. Note that you can sometimes manage to fsck even
the
Thank you, Ben, for speaking up. It sucks to have to
play cop/babysitter but things were getting totally
out of hand here and our normally good S/N ratio is
definitely worth defending.
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
I tried the specified command from three different sites and
they all gave essentially identical responses:
aaa.bbb.ccc.ddd is POOR: 26 queries in 3.1 seconds from 1 ports with std
dev 0.00
That aaa.bbb.ccc.ddd address seems to be the (possibly NAT'd) IP
addr that the target site sees
I have been tasked with some clustering work, and I have run into a
few snags. Is anyone familiar with the RHEL 5 clustering suite?
You should find someone who used to work for MCLX whooh, wait.
Never mind. ;)
Yah, one might reasonably wonder if he isn't thinking that the
software in
And, FWIW, I've got a couple dozen WRT-54G[S,L,v4,v3,v2]* units
in the field and the only one that's failed had taken a lightning
strike. My oldest one is now 5 years old, so at $50-ish, I owe
them nothing.
Closer to $41 today if you can stomach a mail-in rebate:
I guess those $500 cables are OK if you're on a budget,
but us true audiophiles know these $7899.95 beauties are
the ones to have:
http://www.wildwestelectronics.net/g-oval-10-10.html
___
gnhlug-discuss mailing list
Maybe a $PATH problem? the new version probably got installed
someplace that's not mentioned in your $PATH, or in a place
that's mentioned after the old version...
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
I'm still running a fairly ancient version of Firefox because
many of my favorite extensions never (well, I haven't checked
in a while...) got ported forward to the newer versions.
I'm particularly fond of the bookmark sync extension by
Torisugari, but in light of stuff like Weave it may be
Brilliant Pebbles is a unique room system tuning device for
audio systems. Large Brilliant Pebbles now comes as a thin clear
plastic bag that contains various minerals/stones. Other sizes
range from Mikro to Extra Large, and applications vary from size
to size. A number of
It's now OK to download. The site incorrectly diagnosed my
system as needing the Windows tarball instead of the Linux one
(not an auspicious way to start!) so I had to go hunting for
the correct one. In case anybody else enjoys a similar SNAFU
here's the URL I ended up using:
You're the second person to phrase it that way, does the machine
know that it's supposed to route traffic. Showing my ignorance
I will ask how do I make sure it knows to do this. I thought
running the route command did that, but apparently there's more.
I believe the route command is used
I'm trying to understand why a majority of people on this list
find it acceptable to /recommend/ hardware only supported by
proprietary drivers
Hey, cut us some slack, Jack. It would be more accurate for you to
say, majority of postings I've read since (A) most of the traffic
on this list
Yah, if only projects like the Open Graphics board were
farther along:
http://lists.duskglow.com/open-graphics/2008-April/011376.html
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
Hmmm, your Windows requirement probably rules Glimpse out
unless Cygwin can somehow be used, but I'll mention it
anyway: Glimpse (and agrep, the Approximate grep) is a tool
that operates by pre-generating a database from a given file
repository and then allows fast searches using that database.
If I understand the requirements all you need to do is feed
the appropriate ulimit command to the same instance of bash
that ultimately executes the program you're trying to run,
right? One simple approach sketched in pseudo-C would be:
system( /bin/bash -c 'yourUlimitCommandHere ;
Is there a way to prevent them from being able to manually
change certain environment variables
Basically, no. The environment is just a bunch of NULL
terminated strings and there's nothing sacred or special
about any of them. The various flavors of the exec() syscall
just blindly accept
Oh, by the way, if you think that the second word in the acronym
really was fine, then I have a bridge that I'd like to sell you.
Is it compatible with the CPE in question?
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
With RHEL, you eliminate support when you put in a kernel that
isn't one of theirs.
Well, it's true that RHAT will not support just any random
kernel that you've overlaid on top of an RHEL distribution,
but even as a Debian devotee I respect their reasons for
adopting that stance - they're
I recently volunteered to troubleshoot a neighbor's laptop
and found it to be quite the monkey-puzzle - after removing
as many screws as I could find I could only get the various
subassemblies to wiggle slightly WRT each other but couldn't
convince myself that any of them could be separated
DTVZ wrote:
[a rather complete list of points re: internal pwrgnd planes]
Yah, it's been years since I was a bench tech but I'm pretty sure
internal pwrgnd planes are canonical these days. IIRC (and the
smart money is betting that I don't) the only drawback is maybe
something about oddball
I currently store my email in maildir format, I use fetchmail
to retrieve it, procmail to filter it and mutt to read it
My situation is similar enough to yours (I'm totally addicted
to maildir format but use EXMH instead of mutt) that I hope to
be able to follow your investigations here on
According to the Definitions, and goals section here:
http://www.inter7.com/courierimap/README.maildirquota.html
...you shouldn't have to care:
Maildir++ and Maildir shall be completely interchangeable.
A Maildir++ client will be able to use a standard Maildir,
automatically
Heh. The phrase PeeCee architecture is an oxymoron if ever
there was one, but it's a boon to those who make a living
selling hax that work around its deficiencies. One that
comes to mind apropos this thread is a card called the PC
Weasel (or maybe the Real Weasel?) that you could install in
STFW, Luke...
http://www.google.com/search?q=Bluetooth+USB+Wireless
http://castle.pricewatch.com/s/search.asp?s=bluetooth
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
I had a few lines of C Sharp in a flash programming utility
written for linux, (beats all I've seen to date).
My brain couldn't quite lock on to whatever it was you said
there - care to elaborate?
___
gnhlug-discuss mailing list
Working with a embedded platform using a TI part and
Monta Vista Pro embedded linux. A programming utility
was offered for this platform Flasher.tar.gz.
Aha. Yes, that makes rather more sense than anything I came
up with while I was stuck on the meaning of flash, as the
other person had
Did you STFW?
The first handful of hits here looked very promising:
http://www.google.com/search?q=RL5c476+linux
Particularly this one:
http://hardware4linux.info/component/16545/
___
gnhlug-discuss mailing list
In case nobody offers an actual terminal of the type desired
I'll bet there's an emulator available (probably even
floppy-bootable) that'll turn any random cast-off PC into a
terminal with all the deisred characteristics.
___
gnhlug-discuss mailing
My wife and I drive right past there on our weekly
jaunts (in season) to hike Pack Monadnock Saturday
or Sunday mornings - might you be open then? I didn't
see hours mentioned in your email or WWW site...
In any case, luck to you!
___
And tune in next week when Ben takes us through the SunOS -
Solaris name mapping! Following that will be a general discussion
tracking the ATT releases and how they were tracked by BSD :)
This UN*X family tree is known to be b0rken in some ways but
it's still interesting:
Once a fedora release goes end-of-life, there are no more updates,
period. For example, Fedora Core 6 went end-of-life a few months
ago, and hasn't had a security update of any sort released since.
So you have to upgrade the system to the next Fedora release
(or the one after) to keep
I believe RHEL3 U9 was the last full active development release
Unfortunately, the bug in question was apparently introduced
with the U8 release which I understand was supposed to be the
final one, but it's bad enough that I'd suspect there'll be a
what bug did you run across? Don't suppose there's a BZ open on it?
Sorry, I don't have the number in front of me at the moment but
here's the patch that fixes it:
https://bugzilla.redhat.com/attachment.cgi?id=294675
Symptom is that the system appears to hang when filesystems
are
That way the list can continue to have discussions [...]
with out having me bother everyone. :)
The signal on this channel is Linux, so if you're talking
Linux you're not bothering anyone because that's why we're all
gathered here. Of course, you get extra credit for taking
newbies and
This is all so totally fascinating that I urge you to create
another mailing list (similar to gnhlug-jobs or gnhlug-announce)
on which this discussion can be given the attention it deserves.
After you've created that list and invited all interested
parties to join it, I'm sure the resultant
Many of the suggestions made so far seem to assume that the
users are cooperative and will not try to defeat any of the
suggested auditing mechanisms. Is that assumption correct?
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
Many of the suggestions made so far seem to assume that the users
are cooperative and will not try to defeat any of the suggested
auditing mechanisms. Is that assumption correct?
That is a safe assumption. The users are the ones that have asked
for better monitoring then what is done
I need to log everything from the time a user logs in to the time
they log out, including all commands, output of commands, etc.
ttysnoop appears to be close to what you want, though straight
out of the box it looks like it's not configured to relay
session automatically.
Using script isn't
If you can SSH in to your machine it's probably not strictly
a pty problem you're having, since SSH uses them. A couple
of other things to try:
- Try running something like screen which uses ptys but not X.
- Try launching one of the failing apps with strace so you can
maybe see the last
right now it is a pain to add the extra NICs
In some situations (and it sounds like yours might be one of them)
you can just use aliases for the interfaces to get the effect
of adding NIC cards. The syntax and the location of the config
files varies with the distribution but you're basically
Alt-Ctrl-F3 to a terminal window and Ctrl-Alt-F7 once didn't
work, but Ctrl-Alt-F4 and Ctrl-Alt-F7 three times did :)
Bizarre, but it's a work-around that I thought might help others.
You have my sympathies - it's humiliating to have to resort to
that sort of voodoo. FYI, you may be able
I don't see it. I have monthly, weekly and daily jobs that I want
to run sequentially, in that order, starting at 8pm. How would
I do that using 'at'? Preferably without making the actual jobs
know about each other, e.g. not teaching the weekly job that it
should be followed by the
Can somebody who's an admin for gnhlug-sysadmin please
check that list's admin queue for an item from me?
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Just thinking out loud here, but one extremely low-tech way to do it
might be to have your daily/weekly/monthly cron jobs create a work queue
by depositing disposable copies of the desired scripts into a special
directory with names guaranteed to be sequential and unique (like,
say,
You won't be able to do this from userland. That spot of memory
is off-limits because the kernel needs to preserve it in the event
that another process wants to enter a vm86 mode.
Maybe I haven't been following this thread closely enough, but I
don't see how either User mode accesses or
(The nominally cost-effective thing to do is throw the computer out
and buy another one, but I trust you are aware of that and have
particular factors which are affecting the math.)
I let principle override practicality all too often. It irks me
to toss a beautiful piece of engineering
It depends on a number of factors like which compiler version
you're using (try cc --version to find out) and how badly
written that code is. You'll probbaly get a lot farther if
you get your build to stop using -pedantic but that's not an
all-purpose solution and removing it may even allow
301 - 400 of 944 matches
Mail list logo