The general *nix chat thread

stipud question. what am I doing wrong with the following?

this gets no results:
grep -f index.txt searchfile.csv

this gets me the results I was after:

#!/usr/bin/perl
#use strict;
#use warnings;

open (IndexFile, 'index.txt');
while (<IndexFile>) {
chomp;
print `grep $_ searchfile.csv`;
}
close (IndexFile);



Is there a limitation to the size of files which grep can use for strings? index.txt is about 3k and searchfile.csv is around 66mb.

 

What does "file index.txt" say? My guess is index.txt was created on Windows, and is full of extra DOS end-of-line markers.

 

yeah, I just opened it up and showed symbols, and each line has a <cr><lf>... so the chomp was un-breaking it.

Thanks!

 

Ah Windows, is there anything you can't fuck up?

Most Linux distros come with a tool called "To From DOS", which is amusingly abbreviated to "tofrodos" (To Frodo's?). It will trim the extra end of line character, and is safe to re-run on files that don't need it. I tend to run it over input text files in scripts just by policy, given that I don't trust Windows users to know what the hell they're doing ever.

 

Thanks guys, I am still learning how to troubleshoot this stuff, looking at the logs.. I should have thought of that.. I am used to pretty useless windows logs. I don't get any error when starting, just [fail] so here is what I got in the logs.

Code:

Aug 30 12:14:05 kashmir named[2958]: starting BIND 9.8.1-P1 -u bind
Aug 30 12:14:05 kashmir named[2958]: built with '--prefix=/usr' '--mandir=/usr/share/man' '--infodir=/usr/share/info' '--sysconfdir=/etc/bind' '--localstatedir=/var' '--enable-threads' '--enable-largefile' '--with-libtool' '--ena$
Aug 30 12:14:05 kashmir named[2958]: adjusted limit on open files from 4096 to 1048576
Aug 30 12:14:05 kashmir named[2958]: found 4 CPUs, using 4 worker threads
Aug 30 12:14:05 kashmir named[2958]: using up to 4096 sockets
Aug 30 12:14:06 kashmir named[2958]: loading configuration from '/etc/bind/named.conf'
Aug 30 12:14:06 kashmir named[2958]: /etc/bind/named.conf.options:26: undefined ACL '.none.'
Aug 30 12:14:06  named[2958]: last message repeated 5 times
Aug 30 12:14:06 kashmir named[2958]: /etc/bind/named.conf.options:27: undefined ACL '.none.'
Aug 30 12:14:06 kashmir named[2958]: loading configuration: failure
Aug 30 12:14:06 kashmir named[2958]: exiting (due to fatal error)

so I look into /etc/bind/named.conf.options and checked lines 26 and 27.. then I remembered a few days before I added something to this (probably never rebooted since).

Code:

26: // allow-transfer {.none.;};
27: // allow-recursion {.none.;};
28: recursion no;
29: version "Not Disclosed";

so I commented out these 2 lines, and now it works again. nothing to do with the hostname I was adding those lines as someone told me it will stop my server being used for DDOS requests, now I just need to fix this and test it again.

Thanks for the help I now know where to look.. I feel very enlightened.

 

Hey elvis, did you get around to playing with the wife's laptop? Just curious as I still haven't managed to get optimus running yet, seeing if you had some magic up your sleeve.

 

Nope, haven't had time yet.

 

All good, thought I'd ask as most answers I find online are a dead end.

 

Anyone had much experience with GNU's coreutils on a Win32 machine? Got some configuration files that would be handled reasonably easy (well, it'd cut down a lot of tedious effort at any rate) using coreutils but not quite sure just how compatible it is.

 

I copied grep out of it, didn't even require any dependencies, just worked. I don't have any experience with the other programs. Also be wary of windows new line.

 

http://nextmicrosoftceo.com/

...

 

Ended up moving ~/.cache to /var/cache/org/user/ instead of RAM, cos can't mount anything under NFS-mounted /home/user. NFS server load dropped from medium to absolutely nothing (~120 clients on a 2.4Ghz quadcore, 16GB RAM, also runs Samba for about 150 PCs). So it was Firefox, Chrome and Google Talk/Hangout smashing it badly.

The other candidate for a similar migration is ~/.local, particularly gvfsd-metadata.

Bonus for both: reduced backup time / size.



I still use ~/.cache on tmpfs for machines that don't have NFS /home, esp the ones with SSD (longevity++).

Yeah I've already got /scratch for the power lusers to write their big temp data sets to. Occasionally there's still a newbie dumping it to /home though.

 

http://www.linux.com/news/featured-...ideos-from-linuxcon-and-cloudopen-new-orleans

11 Keynote Videos from LinuxCon and CloudOpen New Orleans

The Linux kernel developers panel with Tejun Heo, Greg Kroah-Hartman, Sarah Sharp and Linus Torvalds. The session is moderated by Ric Wheeler.



Eben Upton, Raspberry Pi



Gabe Newell, Valve



"Developers, the new super power." Jonathan Bryce, OpenStack



"New frontiers." Dirk Hohndel, Intel



"The Technium." Kevin Kelly, Wired



"What five years of Android has meant." Chris DiBona, Google.



"The state of Linux." Jim Zemlin, The Linux Foundation



"Before I die." Candy Chang, Artist and TED Fellow



"Linux, cloud and next generation workloads." Brad McCredie, IBM



"Open matters." Brian Aker, HP.

 

Cheers

Here's a reason to not copy and paste from websites into terminal:

http://thejh.net/misc/website-terminal-copy-paste

From:

http://www.reddit.com/r/linux/comments/1norlj/fix_ubuntu/

 

Interesting trick

 

I assume this is more of a CSS trick than a vulnerability? Still good advice, if you just blindly copy paste without looking.

Code:

<p class="codeblock">
      git clone
      <span style="position: absolute; left: -100px; top: -100px">[201~/dev/null; clear; echo -n "Hello ";whoami|tr -d '\n';echo -e '!\nThat was a bad idea. Don'"'"'t copy code from websites you don'"'"'t trust!<br>Here'"'"'s the first line of your /etc/passwd: ';head -n1 /etc/passwd<br>git clone </span>
      git://git.kernel.org/pub/scm/utils/kup/kup.git
    </p>

edit: just realised you would have to paste into a text editor first, or it would just start to execute the code before you got a chance to read it and press enter.

 

Tried Archlinux for the first time over the weekend. I think this is going to end up being my favorite distro.

Managed to get a Archlinux VM to boot only using 28Mb of RAM. That is nuts.

 

The AMD Radeon Performance Is Incredible On Linux 3.12

http://www.phoronix.com/scan.php?page=article&item=amd_linux312_major&num=1

 

I've come across a real niggling little issue which is proving to be annoying to resolve. I've setup a CentOS 6 VM running BIND (9.8.2) as a secondary NS and run into a very frustrating file permission error.

As it's CentOS / RHEL, the bind process runs as "named" with user / group both named. Zone files are stored in /var/named, again standard config.

Initially running it the zone files refused to write so of course I double checked permissions. The file ownership is correct and I've confirmed that the process is running as named.

Now for the gotcha, I ran named manually (/usr/sbin/named -u named) and it worked perfectly.

The run line in the init.d script is as follows:

daemon --pidfile "$ROOTDIR/$PIDFILE" /usr/sbin/"$named" -u named ${OPTIONS};

Both appear exactly the same from a process list, so it has me stumped! I've double checked and both are reading the same config so therefore trying to access the same files.

I've run strace to see if I can identify anything quirky but nothing stands out. I can't figure out why they're they're different yet running as the same user.

Any ideas?!

 

SELinux.

try, setenforce 0 (to temporarily disable SELinux).
see if the problem goes away if it does, you probably want to do a relabel.

edit: did you install bind via yum? it should've put the right labels on the folders for everything to work.