March 2012 - luv-main

comparing bulk files
by James Harper 23 Mar '12

23 Mar '12

I have just finished recovering the data from a failing NT4 installation. The dd took _ages_ but reported no i/o errors, but booting failed with a checksum failure and investigation revealed there was a single bit error in a critical file which I subsequently replaced from a download of NT4 sp6a. A single bit error in one file obviously raises the possibility of more such errors... So I have a directory tree of NT4 files in the image I took of the original disk, mounted under Linux, and a flat directory containing the files from service pack 6a, and I want to compare them. There are heaps more files on the disk but these are the only ones I have the originals of. So the disk image fs looks like (for illustrative purposes only - actual filenames may not be correct): /WINDOWS/welcome.exe /WINDOWS/system32/mem.exe /WINDOWS/system32/WINSRV.dll /WINDOWS/file_that_does_not_exist_in_sp6a.dll and the sp6a directory looks like: /sp6a/WELCOME.EXE /sp6a/mem.exe /sp6a/winsrv.exe /sp6a/file_that_does_not_exist_on_installation.dll The files are possibly in different directories, and are possibly different cases. I could write some horribly inefficient shell script to just go through each file in the sp6a directory and find a file with the same name in the /WINDOWS tree and diff it, but maybe there is a better way? This would at least give me a very rough idea of how many errors I might be looking at. Thanks James

7 11

Any practical way to improve glacial page downloads?
by Erik Christiansen 23 Mar '12

23 Mar '12

About every second day now, my ADSL broadband connection turns to sludge. Just loading the ABC news homepage, or BBC, or anything else, takes up to a minute, or sometimes stalls indefinitely. Power cycling the ADSL modem sometimes fixes this problem, and also the times when DNS stops working, i.e. the stall is during "Looking up ...". I just did that now, without improvement, but hitting the "Reload Current Page" icon in firefox sometimes helps. In firefox, I have network.dns.disableIPv6 = "user set boolean true", but that's more for slow "Looking up ..." AIUI, whereas it's glacial page transfers which make the service almost useless at the moment. The link looks more than OK: Connection Status Connected Us Rate (Kbps) 128 Ds Rate (Kbps) 1536 US Margin 27 DS Margin 21 Trained Modulation T1413 LOS Errors 0 DS Line Attenuation 54 US Line Attenuation 31 Peak Cell Rate 301 cells per sec CRC Rx Fast 3 CRC Tx Fast 1 CRC Rx Interleaved 0 CRC Tx Interleaved 0 Path Mode Fast Path /proc/sys/net/core/rmem_default is 112640 /proc/sys/net/core/rmem_max is 131071 Repeated long latencies within Internode's network don't look like they'd help much, but I don't have anything to compare that with. The fact that it seems to have failed to get to the target in the end looks ghastly, and beyond my fixit toolkit. $ traceroute www.linuxcnc.org traceroute to www.linuxcnc.org (69.163.248.64), 30 hops max, 60 byte packets 1 router (192.168.1.254) 1.110 ms 1.313 ms 1.532 ms 2 lns20.mel4.internode.on.net (150.101.212.44) 16.206 ms 22.500 ms 29.324 ms 3 te2-2.cor3.mel4.internode.on.net (150.101.208.65) 36.055 ms 42.548 ms 49.196 ms 4 gi6-0-0.bdr1.mel4.internode.on.net (150.101.210.41) 225.556 ms 231.354 ms 238.425 ms 5 te5-0-0.bdr1.syd6.internode.on.net (150.101.208.106) 250.423 ms 253.429 ms 257.861 ms 6 pos1-0-0.bdr1.lax1.internode.on.net (203.16.213.190) 263.005 ms 269.143 ms 274.384 ms 7 144.223.30.1 (144.223.30.1) 282.270 ms 183.912 ms 183.287 ms 8 sl-st31-la-0-0-3-0.sprintlink.net (144.232.5.100) 187.641 ms 194.321 ms sl-st31-la-0-4-2-0.sprintlink.net (144.232.5.102) 200.827 ms 9 144.232.24.222 (144.232.24.222) 207.956 ms 144.232.9.46 (144.232.9.46) 184.232 ms 144.232.24.222 (144.232.24.222) 190.154 ms 10 te0-2-0-2.ccr22.lax01.atlas.cogentco.com (154.54.29.201) 197.100 ms te0-3-0-5.ccr22.lax01.atlas.cogentco.com (154.54.44.133) 183.526 ms te0 (154.54.29.201) 181.851 ms 11 te8-1.mag01.lax01.atlas.cogentco.com (154.54.28.146) 236.065 ms 238.283 ms 239.757 ms 12 38.122.20.218 (38.122.20.218) 207.794 ms 214.622 ms 183.125 ms 13 ip-66-33-201-114.dreamhost.com (66.33.201.114) 204.995 ms * * 14 * * * ... 30 * * * Unless they're just hosted on dreamhost, and local firewalls are intervening? Just guessing, since we also have this: $ traceroute abc.com.au traceroute to abc.com.au (203.2.218.208), 30 hops max, 60 byte packets 1 router (192.168.1.254) 0.615 ms 0.786 ms 0.987 ms 2 lns20.mel4.internode.on.net (150.101.212.44) 15.727 ms 22.029 ms 28.338 ms 3 te2-2.cor3.mel4.internode.on.net (150.101.208.65) 37.551 ms 43.832 ms 50.405 ms 4 gi6-0-0.bdr1.mel4.internode.on.net (150.101.210.41) 57.123 ms 63.398 ms 70.258 ms 5 te6-0-0.bdr1.mel6.internode.on.net (150.101.210.50) 75.448 ms 81.458 ms 88.341 ms 6 abc.melbourne.pipenetworks.com (218.100.13.6) 107.321 ms 113.388 ms 115.609 ms 7 * * * ... 30 * * * It'd be nice to be able to do something to improve the page transfer speeds on the bad days. Erik -- "Political Correctness is a doctrine fostered by a delusional minority and by the mainstream media which holds forth the proposition that it is entirely possible to pick up a turd by the clean end." - Author unknown.

21 57

xterm mouse pointer colour won't change
by Erik Christiansen 23 Mar '12

23 Mar '12

Environment: Ubuntu 10.04, gnome. After a couple of decades of running xterms with "-fg yellow -bg darkslategrey", I'm finding it harder to locate the mouse pointer when it's over text, even if I thrash the rat, to try to detect the motion. So I added "-ms red" to the xterm invocation. But that little I-like pointer remains white with a black border. I then tried putting this: *XTerm.pointerColor: red into ~/.Xdefaults , but it also had zero visible effect. Maybe it doesn't like red? Let's try: $ /usr/bin/xterm -bg darkslategrey -fg yellow -ms blue Nah. Humph. I'll try again tomorrow, I think. Erik -- Less is more or less more - Y_Plentyn on #LinuxGER

6 17

auto-complete bug in byobu
by Andrew Miller 22 Mar '12

22 Mar '12

Hey guys, I installed Byobu version 5.16 recently and I'm experiencing an annoying bug when I press tab to auto-complete a command. If you're not familiar with it, Byobu is a collection of scripts for screen and tmux. See http://en.wikipedia.org/wiki/Byobu_%28software%29. I'm using it with screen version 4.0.3-r4. I noticed the bug the first time I ran Byobu, so I'm pretty sure Byobu is what's causing it. When I press tab the first few characters of the command appear as normal. However the last character is a circle. The auto-complete options appear below the command as normal. I have no problems entering the command. However, after entering the command backspace and delete no longer work. Each time I press one of these keys it produces a circle character. Here's a photo I took after pressing the backspace key a dozen times: http://i.imgur.com/PsZeL.jpg At first I thought the bug goes away if you detach from Byobu (go to the tty console) and resume. This does happen but later the bug reappears. Any ideas how to fix it? If it's a setting that I have to change, there are a lot of configuration files to read and I don't know what I'm supposed to be looking for. Do you think I should submit a bug report? Thanks --Andrew

1 0

importing dvd's into mythtv
by James Harper 20 Mar '12

20 Mar '12

I'm having a bit more success with importing some dvd's into mythtv, but the 'bad sector' copy protection is painful. dd_rescue allows me to make an image but the deliberate bad sectors turns a 30 minute process into a 1 hour 30 minute process as reading bad sectors is terribly slow. The bad sectors are almost always going to be in a chunk (512 sectors for the Toy Story 3 DVD I'm importing now) so I was wondering if there exists an application to do it a bit smarter, eg if you get 4 bad sectors in a row, jump forward some small number of sectors (eg 64) and keep doing that until a good sector is found, then work backwards to the last bad sector, and then assume that the sectors inbetween are bad. That would dramatically speed up the process... Does such an app exists? Or failing that, a way to make the DVD drive not spend excessive time trying to read corrupt sectors and let the application re-read if it thinks it needs to? I could simply rip the movie out but that requires more effort than I'm prepared to commit to the process (finding the movie file is increasingly difficult as modern dvd's contain phantom files of similar sizes) and I value my time more than I do the $$$ required to store the raw iso's. And the kids enjoy the little dvd games too. Mainly I want to avoid the piles of DVD's that never seem to get put away :) James

2 1

Cobbler, Koan, Juju and Orchestra
by Toby Corkindale 16 Mar '12

16 Mar '12

Hi, So: Cobbler, Koan, Juju and Orchestra I wondered if anyone here is using any of these, and if so, what's your opinion? Do they work as well as advertised? Are they any good when running with local virtual machines rather than on an EC2-compatible cloud? The ideas look good, but it's surprisingly hard to come up with documentation showing how to go from an Ubuntu install to actually having the whole lot working. (The things I did find gloss over important steps, or only get one part going, such as PXE boot, but not juju.. or juju or koan to boot a machine, but not configure it..) Cheers, Toby

3 3

Audio cleanup of an MPEG4 video
by Peter Ross 16 Mar '12

16 Mar '12

Hi all, we have an instruction video with a bad sound. Most of it caused by breathing into the micro. I thought about using Audacity to clean it up. But first I have to "extract" the audio only, and later I have to re-attach the audio afterwards. I do not have any experience with it so any recommendation is welcome. Thank you Peter

3 3

Re: Delay emails
by Mike Fabre 15 Mar '12

15 Mar '12

On Wed, Mar 14, 2012 at 02:27:42PM +1100, Peter Ross wrote: > Quoting "Mike Fabre" <mike+luv(a)fabre.id.au>: > >> Hi, I have a postfix & dovecot server set up, and I have been given the >> task of making it put a arbitrary 5-10 minute delay for every email that >> comes through, whether it be for sent mail or received mail doesn't >> matter. > > Does a configuration as described here work? > > Postfix on a dialup machine > http://www.postfix.org/STANDARD_CONFIGURATION_README.html#dialup > > You could put one more relay in your chain and run the delivery > > randomized, instead of in the dialup script. > > Apologies if this was discussed before, I probably missed a handful of > mails in the thread. Brilliant, this works perfectly. Thanks, I'll use a cronjob to push all messages through every 5-10 minutes that should work well enough. -- Mike Fabre

3 2

Connecting linux laptop to eduStar networks
by Daniel Dalton 15 Mar '12

15 Mar '12

Hi, I'm attempting to connect my linux laptop to my government school's wifi network via wpa_supplicant. Firstly, I've been given some instructions for network-manager, which are as follows: open network manager (probably top right) and manually add a new wireless network use eduSTAR as the SSID Security: WPA & WPA2 enterprise - (WPA2) Authentication: TLS Identity: host/8808-DDLINUX.services.education.vic.gov.au user certificate: point to extracted folder: computers.crt CA Certificate: point to the extracted folder: cacert.pem Private key: point to the extracted folder: key.pem Password: password Next I've looked through some example wpasupplicant config files and also consulted man wpa_supplicant.conf And developed the following: network={ ssid="example wpa2-eap network" key_mgmt=WPA-EAP proto=WPA2 pairwise=CCMP group=CCMP eap=TLS identity="user(a)example.com" ca_cert="/etc/cert/ca.pem" client_cert="/etc/cert/user.pem" private_key="/etc/cert/user.prv" private_key_passwd="PKCS#12 passhrase" } Next I modify this file for my school's network according to the instructions I got at the top. network={ ssid="eduSTAR" key_mgmt=WPA-EAP proto=WPA2 # pairwise=CCMP # group=CCMP eap=TLS identity="host/8808-DDLINUX.services.education.vic.gov.au" ca_cert="/path/to/cacert.pem" client_cert="/path/to/computers.crt" private_key="/path/to/cacert.pem" private_key_passwd="password" } I'm suspecting this probably won't work though since it's usually the case:) So some questions: 1) HOw do I make this configuration file accurately reflect the configuration for this network, according to the instructions for network-manager? 2. Do I need those lines which I commented out in my final configuration file which has been modified to reflect the information my school provided me (my last paste) 3. The original "private_key" certificate in the raw instructions I got from the man page was a "prv" file. However, my school says to use the .pem file. Will this work? 4. The same is true for client_cert, the original config specifies a "pem" file while my school says to use the "crt" file. Which makes me a bit confused about what to do for 3 and 4. So does anyone have any ideas how to make this network configuration likely work? I haven't tested it yet, but I thought I'd try and get it as perfect as I could first before testing since I don't really know what I'm doing :) Thanks very much for any help. Dan

2 5

Delay emails
by Mike Fabre 14 Mar '12

14 Mar '12

Hi, I have a postfix & dovecot server set up, and I have been given the task of making it put a arbitrary 5-10 minute delay for every email that comes through, whether it be for sent mail or received mail doesn't matter. The first thing we thought of was to make all users use procmail, and set up a global procmail rule with a sleep in it, while this will almost certainly work, the users must not be able to run commands on the server, which they can do with a ~/.procmailrc file. I have tried to set up a postfix filter which would delay the mail (similar to this: http://pintant.cat/2010/04/14/how-to-delay-the-mails-postfix/) but that has an issue where it will only allow 5 of these filters to be running at a time, so if I where to send 10 messages with a 5 minute delay, the first 5 will take 5 minutes and the next 5 will take 10 minutes. I can change this limit within postfix but I would prefer to use a method that is more scalable. Anyone got any ideas? -- Mike Fabre

11 28