Re: OT: Telecom Proxy servers?
Telstra subject their cable customers to a transparent proxy as well. Their proxy behaves when told to get the uncached version of an object thankfully. - neil On Thu, Mar 04, 2010 at 01:00:40PM +1300, Craig Falconer wrote: CABLE CABLE CABLE!
Re: OpenMoko Open Source / Linux / Hackable cellphone is on sale now.
I'm currently using pre-pay with Vodafone. Do you know off-hand if it will just work with my existing SIM? If so then I'm definitely in. we've got to support these guys otherwise open hardware as a concept won't work. Having a hackable phone is going to be so cool. With the GPS, there's no reason your phone shouldn't double as an open source SatNav. - neil On Tue, Jul 08, 2008 at 11:00:33AM +1200, John Carter wrote: The Open Source / Embedded Linux Open / Hackable Cellphone is on sale now! I note their 900/1800/1900 Mhz tri band is sold out, but the 850/1800/1900 is still available for $399 or $3690 for packs of 10. http://www.openmoko.com/product.html I note Vodafone 900 and 1800 services and Telecom has 850mhz planned. http://www.gsmworld.com/roaming/gsminfo/cou_nz.shtml The telecom site suggests that you'll not be able to use the openmoko on their network... http://telecom.custhelp.com/cgi-bin/telecom.cfg/php/enduser/std_adp.php?p_faqid=104 http://telecom.custhelp.com/cgi-bin/telecom.cfg/php/enduser/std_adp.php?p_faqid=135 Question 1 to the Group: Do we need to have the 900Mhz version here or is the 850 preferable because in the longer term it will allow us to switch to Telecom if we wish? Question 2 to the Group: Anyone planning on distributing these things in NZ? Question 3 to the Group: I'm planning on ordering one for myself, and would be prepared to go for a 10 pack if enough others are interested. Anybody else interested? John Carter Phone : (64)(3) 358 6639 Tait ElectronicsFax : (64)(3) 359 4632 PO Box 1645 ChristchurchEmail : [EMAIL PROTECTED] New Zealand
Re: Joys of not quite monolithic kernels...
in the same vein, we plugged a USB hard disc caddy into a laptop here not five minutes ago and it greeted us with: mount: unknown filesystem type 'ntfs' so we: cd /usr/src/linux-2.6.23.14-callisto/ make menuconfig nice -19 make sudo make modules_install mount /media/usbdisk/ ..and the error no longer appears. how cool to be able to add NTFS support to your kernel without rebooting? - neil On Wed, Feb 13, 2008 at 09:14:57AM +1300, John Carter wrote: What ever the merits of the grand debate about Micro vs Monolithic Kernels are ...here is a curious data point. My CD writer ceased to write. Cease to notice blank disks. T'was nailed to the perch. I was about to invoke guarantees and the like when I thought... perhaps I should reboot (you know The Computer Industry joke, push it back up... tell me if you haven't heard it, we computer types really deserve that joke.) When I thought, I don't want to reboot, I have too much running... Hm, perhaps... what module controls the device... # lsmod . . . cdrom 37536 1 sr_mod . . # Aha! There it is! So let's try... # rmmod sr_mod And bang udev decides it needs sr_mod to handle this device and pulls it in again... it sees the blank disk, I burn the disk and off we go without a reboot. John Carter Phone : (64)(3) 358 6639 Tait ElectronicsFax : (64)(3) 359 4632 PO Box 1645 ChristchurchEmail : [EMAIL PROTECTED] New Zealand
Re: usb wireless...
I've got an Edimax EW7318USg with external antenna from Computer Dynamics that uses the rt73 module from: http://rt2x00.serialmonkey.com/wiki/index.php?title=Downloads ..and a Dick Smith XH8344 that uses the rt2570 module from the same URL, and a SafeCom SWMULZ-5400 that uses the zd1211rw module from recent kernels. They all seem to work ok, although some of them need the interface brought up before they can associate, others require firmware and I'm left with the impression they're not 100% reliable but they do work. - neil On Tue, Dec 18, 2007 at 07:29:22PM +1300, Steve Holdoway wrote: Can anyone recommend a supported wireless usb device? 11b is fine, just need connectivity for browsing the internet - nothing flash - and I haven't got any PCI slots available. I was going to use my old dlink dwl-120+ but they seem to have broken the drivers in gutsy. Which is annoying. To say the least. Cheers, Steve -- Steve Holdoway [EMAIL PROTECTED]
Re: Wiki Software
Whatever wiki software is easiest for you would be grand (even if it means using a different wiki markup). How will the existing content be migrated? On Tue, Dec 11, 2007 at 04:23:01PM +1300, Jim Cheetham wrote: From a user perspective, however ... you all might like a slightly different feature set :-)
Re: workstation power consumption
There are gubbins that can be plugged into a power outlet and which measure the power used by anything that is then plugged into their integrated power outlet. You are welcome to borrow mine if you can't get hold of one. It measures VA, W and the power factor between them and if left in overnight will record the kWh used, which can obviously be combined with the unit charge on your electricity bill to find out how much it cost to run an applicance overnight. Using my gubbin, I read 20W for my EPIA server, 15W for a Transmeta laptop, and 200W for a workstation, although I would guess that modern workstations would use more than 200W. - neil On Fri, Sep 21, 2007 at 11:11:04AM +1200, Roger Searle wrote: Marginally on / off topic I guess... I am interested in quantifying how much it costs to leave a computer running over night, and therefore what the power saving is per year to an organisation if computers that are otherwise left running all night are turned off via a schedule. The question I don't have the answer for so am asking the list is, what would an approximate power consumption be for a typical modern workstation that is sitting idle? Assuming that the monitor has powered down by the operating system. Cheers, Roger
Re: Ubuntu apt archives
Make use of an archive proxy: http://clug.net.nz/index.php/Approx (once the wiki is back up) and you'll never lose another .deb again. And better, the same proxy can be shared between multiple machines. You can even use such a proxy in off-line mode: http://mozart.amiton.ltd.uk/mysite/approx.html - neil On Fri, 2007-05-18 at 02:03 +1200, Rik Tindall wrote: Hi, Has anyone kept all the 6.06 update downloads, for sharing? - A pain to retrieve all over again, for each new install.. Have found /etc/apt/apt.conf.d/20archive for where to stop them being cleared out in future i think - disable line thus?: #APT::Archives::MaxAge 30; cheers
Re: Ubuntu Feisty Fawn upgrade.
On Mon, 2007-04-23 at 18:31 +1200, Robert Fisher wrote: On Monday 23 April 2007 6:10 pm, Kerry Mayes wrote: Here at work we had three machines running Edgy and it seemed wasteful for each to download 900Mb I have an IPCop box with a 30Gb cache which means that our 3 Mepis boxes only get files once from the internet when upgrading/updating. Have a look at the approx or apt-proxy packages. There's even an account of approx on the wiki: http://clug.net.nz/index.php/Approx - neil
Re: debian (knoppmyth actually) eth0 not coming up
On Thu, 2007-03-22 at 20:38 +1200, Steve Holdoway wrote: Nick Rout [EMAIL PROTECTED] wrote: Due to a power supply letting out the magic smoke I moved my mythtv main OS drive into another case with a different motherboard etc. Of course all the chipsets are different, what was via-rhine is now sis900 etc. now the sis video card is recognised on boot, and there is an eth0 created, as one would expect on a reasonably modern kernel. The trouble is it is not getting an IP address on bootup. /etc/networking/interfaces looks right (and indeed is unchanged, so should work). It is set to dhcp. Running /etc/init.d/networking restart after boot works, but is a PITA for a machine with (usually) no keyboard. Can anyone tell me where to look to determine why eth0 is not automatically getting an ip address? knoppmyth is based on knoppix which is basically debian. I figure that: 1. I changed nothing in terms of settings (except the /etc/X11/Xf86config-4 file) 2. eth0 is detected and the driver loaded by the kernel, same as before (albeit a different driver, but its still eth0) 3. By the time the startup scripts run /etc/init.d/networking it should ready to obtain an IP address, the same as before. This usually happens when the relevant module isn't loaded when the networking is started. Try stuffing a modprobe whatever at part of the networking start code, just to be sure... do you mean something like: iface eth0 inet dhcp pre-up /sbin/modprobe sis900 (in /etc/network/interfaces), or aliasing eth0 to sis900 in /etc/modprobe.d/aliases? - neil
Re: Music Player Daemon
assuming you issued the apt-get command below then MPD and a command-line client for it are already installed on your system. further, MPD is already running. to configure it, edit /etc/mpd.conf to show MPD where your music directory is and then: sudo /etc/init.d/mpd restart ..to tell MPD to take notice of your changes. to control MPD, use mpc. use: man mpc ..to learn how to use mpc, but in brief: mpc update (to update the database of music files) mpc add PATH (to add the music file at relative path PATH to the active playlist) mpc play mpc pause mpc stop as you can see, it a lot more complicated to use than a typical music player and is will therefore only be of interest if you can see its advantages. - neil On Mon, 2007-03-05 at 22:00 +1300, David Merrick wrote: Having downloaded it to Unbuntu how do I install and use it?? On 3/1/07, Neil Stockbridge [EMAIL PROTECTED] wrote: apologies in advance if you already know of this, but check out MPD. it's a music player (MP3, OGG, etc.) but with a client-server architecture. you could use one of the existing clients or write your own UI so you can have everything exactly as you like it. you link with libmpd and send commands like select playlist ambient, play, pause, set volume to 50%, etc. and it sends back messages like now playing Run Run Run by Goldenhorse, volume set to 50% and so on, not to mention that it builds a database of ID3 tags and presents a friendly API for mining that data. web site: http://www.musicpd.org/ packages: sudo apt-get install mpd mpc libmpd-dev (or similar for your distro) - neil
Re: anyone want any ISOs?
On Fri, 2007-02-09 at 20:01 +1300, Nick Rout wrote: On Friday 09 February 2007 11:33, Neil Stockbridge wrote: i got myself a 25GB data plan with Slingshot and now i can't use it all up. does anyone want any ISOs or other large files downloaded from the Internet? latest Sabayon DVD would be sweet ;-) ok. the torrent client says it should be here by Monday at the current rate.
anyone want any ISOs?
i got myself a 25GB data plan with Slingshot and now i can't use it all up. does anyone want any ISOs or other large files downloaded from the Internet? - neil
Re: anyone want any ISOs?
On Fri, 2007-02-09 at 12:50 +1300, Phill Coxon wrote: On Fri, 2007-02-09 at 12:42 +1300, Christopher Sawtell wrote: On Friday 09 February 2007, Neil Stockbridge wrote: i got myself a 25GB data plan with Slingshot and now i can't use it all up. does anyone want any ISOs or other large files downloaded from the Internet? A current knoppix DVD would probably be of interest to more than one member. I'll second that. ok, if no-one comes forward with 5.1.1 on DVD already then I'll start bittorrent later on today
Re: OT: Still looking.... Java, c++
Is this the same project we were talking about in October? I'm back in Christchurch now. I'm house hunting full-time at the moment, but I'll be looking for a job shortly after we've completed. - neil On Fri, 2007-01-26 at 12:03 +1300, Steve Holdoway wrote: Linux/C++/Java developers/designers. Currently having to look as far away as the UK at the moment. Any more local people looking for a challenging position in a laid-back environment??? If so, get in touch, and join us fighting the spam wars. Cheers, Steve
OpenOffice spreadsheet diff
i seek your aid in my quest to compare two OpenOffice spreadsheets. like a fool, i forked the sheet and now i have two versions of the same spreadsheet both containing modifications that i would like to keep. the .sxc files can be unzipped, to yield these files: content.xml META-INF meta.xml mimetype settings.xml styles.xml the XML is all on one line, so: sed -i 'y//\n/' *.xml ..puts those XML files into a more comparable form yet the actual changes are swamped in a sea of formatting changes that i didn't make. the sheets could be exported as CSV and compared but changes to formulae would be missed. so is there a better way of comparing two spreadsheets? - neil
Re: OT:[Fwd: TP: How NSA access was built into Windows]
On Fri, 2007-01-19 at 19:08 +1300, Volker Kuhlmann wrote: but that was accidental. How many million transistors on a chip now? So, where do you start? here: http://www.homebrewcpu.com/ - neil
SiS integrated graphics with Ubuntu 6.10
with Ubuntu 5.10 on a laptop with SiS integrated graphics, the LCD was fine. after installing Ubuntu 6.10 the LCD flickers in an almost imperceptible way but enough to be irritating as if the LCD didn't have enough power. after fiddling for hours with refresh rates, different X drivers and even xscreensaver at the behest of google it was solved by turning the brightness up. the problem only occurs when the LCD brightness is set to minimum. - neil
Re: Reg Help Grep....
both of these are a bit hacky and might break if the HTML changes. using cut: grep 'span id=lblValuationNumber' cHTML | cut -d'' -f2 \ | cut -d'' -f1 (look up the meaning of the -d and -f options) using sed: grep 'span id=lblValuationNumber' cHTML | sed 's/[^]*//g' \ | sed 's/^ *//' (using regular expressions to cut out the HTML tags and then strip the white space) - neil On Thu, 2006-12-07 at 21:08 +1300, Don Gould wrote: [EMAIL PROTECTED] shared]# grep 'span id=lblValuationNumber' cHTML span id=lblValuationNumber22115 36400 B/span/td [EMAIL PROTECTED] shared]# Except I only wanted 22115 36400 B (without the quotes). I've tried some different options... [EMAIL PROTECTED] shared]# grep 'span id=\lblValuationNumber\(.*?) \/span' cHTML [EMAIL PROTECTED] shared]# grep 'span id=\lblValuationNumber\(.*?)/span' cHTML [EMAIL PROTECTED] shared]# grep 'span id=lblValuationNumber(.*?)/span' cHTML All return nothing. Idea's what I'm doing wrong? Cheers Don
Re: OnT technical - Java install problem
i had some trouble with the Sun JRE. it hadn't finished unpacking properly. there should be a /usr/java/jdk1.5.0_09/bin/unpack200 tool that it uses to unpack .jar files. look in /usr/java/jdk1.5.0_09/jre/lib and other places for .jar files and see if some of them are called javaws.jar.pack or something similar (i don't remember exactly). then use the tool to unpack them manually. this may or may not be your problem because i remember a slightly different error message but it's worth checking. - neil On Thu, 2006-11-30 at 11:51 +1300, Barry wrote: Hi all, I am trying to get java up running because OO2 requires JRE to save docs in win2003 xml format. Java is installed in /usr/java/jdk1.5.0_09/.etc. As usr I have r-x access to all files. PATH is set to /usr/local/bin:/bin:/usr/bin:/usr/X11R6/bin:/usr/games:/usr/java/jdk1.5.0_09/bin:/home/barry/bin by making an alt in /etc/profile as root In a term the cmd 'java -version' produces the following - There was an error trying to initialize the HPI library. Please check your installation, HotSpot does not work correctly when installed in the JDK 1.2 Linux Production Release, or with any JDK 1.1.x release. Could not create the Java virtual machine. Google has produced some install info only. Chris supplied the file from which I installed java, his works ok. Any ideas please Barry
Re: Java install problem
On Thu, 2006-11-30 at 21:57 +1300, Barry wrote: I reinstalled, the msg is now--- Error occurred during initialization of VM java/lang/NoClassDefFoundError: java/lang/Object this is the error i used to have. Does java require a runtime file rt.jar because it has not been installed/exists on my box?!?!? it does require it and it should be at /usr/java/jdk1.5.0_09/jre/lib/rt.jar. see if there is one of those packed files i was talking about in my other post there instead and if so, unpack it. - neil
Re: CLUG - my observations
On Tue, 2006-11-28 at 23:22 +1300, Rik Tindall wrote: 'Push technology'? (I have a backlog of replies, that peace holds back.) Why advocacy work gets nowhere, I've realised, is that *nix/Net is a 'pull technology' of quality - so that everyone capable of using it already knows /or does (use it). this is interesting although the LES students and many people who have posted saying CLUG was their route into unix provide counterexamples. some people use unix happily if they don't have to administer it. my mum falls into this category. Internet cafes can deploy unix without their punters noticing. but this is the dirty secret isn't it? in order to use unix at all unless it is administered for you, one needs to be interested enough to want to learn about it, which few people are.
Re: fluxbox??
KDE and GNOME envoronments require more RAM and CPU than lightweight environments such as fluxbox and IceWM although KDE and GNOME offer more features. i use IceWM on a laptop with 112 MB RAM and 600 MHz CPU and GNOME on my beefier laptop. - neil On Fri, 2006-11-03 at 19:54 +1300, Christopher D MAHER wrote: Hi Everyone, I had a guy come in to work today and talk about Linux. He said he ran Gentoo with something called fluxbox... is this right? Can anyone shed some light on this for me please. He said to the effect that its a light GUI that runs over the top of a command prompt. Apparently GNOME chews ram quickly, at least as a GUI. Please help! Chris Maher. PS Sorry again for the flame:)
Re: Suggestions for automated backup system?
i concur concurrently. here's another example of an rsync script [ #!/bin/sh BACKUP_TO=$YOURSERVER:$PATH_ON_YOURSERVER THIS_HOST=`hostname` OPTS= OPTS=$OPTS --archive OPTS=$OPTS --delete OPTS=$OPTS --delete-excluded OPTS=$OPTS --hard-links OPTS=$OPTS --temp-dir /tmp OPTS=$OPTS --compress OPTS=$OPTS --bwlimit=4 OPTS=$OPTS --exclude=/dev/ OPTS=$OPTS --exclude=/mnt/ OPTS=$OPTS --exclude=/proc/ OPTS=$OPTS --exclude=/sys/ OPTS=$OPTS --exclude=/tmp/ OPTS=$OPTS --exclude=/var/ rsync $OPTS $* / $BACKUP_TO/$THIS_HOST ] - neil On Tue, 2006-10-31 at 22:05 +1300, Hadley Rich wrote: On Tuesday 31 October 2006 21:51, Volker Kuhlmann wrote: It's really about time I installed a decent backup procedure for my linux box. Any other suggestions? rsync, bash. Bonus: does the job *exactly* as you want it done. Negative: no GUI. But then, a GUI is doing you a fat load of good with cron anyway. I concur. I use something based around this main line; rsync -aRv --stats --delete-during --backup \ --backup-dir=/$NAME-$DATE $OPTIONS \ {/etc,/home,/var,/root} \ $SERVER::$MODULE/$NAME 21 /var/log/hackup.$NAME hads
Re: Syntax help for script...
did you sort all of this out in the end? i find that reading the shell scripts in /etc/init.d teaches me how to do things in shell scripts. i didn't see anyone give an example of if in a shell script, so if you've not seen it yet: if [ x1 == x`iptables -vL INPUT | grep banned2 | wc -l` ]; then echo y; fi (the x in x1 above is so that when the backtick command returns the empty string, the shell interprets if [ x1 == x ].. rather than if [ 1 == ].., which is illegal. the banned2 string above if what i'm looking for the presence or absence of in the iptables output. wc -l counts the lines in the output) - neil On Fri, 2006-10-27 at 12:00 +1300, Don Gould wrote: Good news! Last night, with the help of Tusker and Neil I got the iptables stuff logging! :) Bad news... I'm stuck again :) I have two issues - GREP IN A SCRIPT, DO WHILE LOOP GREP IN A SCRIPT The question... how do I script a test? if (iptables -L traffic_in -vn | grep 192.168.3.136) = then iptables -A traffic_in -d $3 end if (Yes, this is how I'd think of it in VB) The problem... Duplicate records. # iptables -A traffic_in -d $3 This creates a new counter every time the ip is refreshed by dhcp. So I need to test to see if it's present. I understand that I use grep, but I don't know how to code the syntax to test it in a script. # iptables -L traffic_in -vn Gives me... Chain traffic_in (13 references) pkts bytes target prot opt in out source destination 9856 6879Kall -- * * 0.0.0.0/0 192.168.3.136 0 0all -- * * 0.0.0.0/0 192.168.3.136 0 0all -- * * 0.0.0.0/0 192.168.2.130 7053 586Kall -- * * 0.0.0.0/0 192.168.2.124 34918 35Mall -- * * 0.0.0.0/0 192.168.3.183 Which as you can see has duplicate records for 192.168.3.136... So at the command line I go ... [EMAIL PROTECTED] shared]# iptables -L traffic_in -vn | grep 192.168.3.136 9856 6879Kall -- * * 0.0.0.0/0 192.168.3.136 0 0all -- * * 0.0.0.0/0 192.168.3.136 But how do I script this so I can use it in the batch file... if (iptables -L traffic_in -vn | grep 192.168.3.136) = then iptables -A traffic_in -d $3 end if DO WHILE LOOP Next I need to collect the traffic information into the database. Sudoo code... for each line in iptables -L traffic_in -vn Ip = $8 Data = $2 mysql -h bowenvale -u oncs -pbutterfly -e INSERT INTO oncs.tblData (IPAddress, DataUsed) VALUES('$Ip', '$Data'); loop From: http://forum.openwrt.org/viewtopic.php?pid=30841 [EMAIL PROTECTED] shared]# iptables -L traffic_in -vn | awk '{if (int($1)!=0) print $8\t$2}' 192.168.2.124 33M 192.168.2.148 5140K 192.168.3.136 6879K 192.168.2.124 721K 192.168.3.183 35M [EMAIL PROTECTED] shared]# This is great, but how do I turn it in to a loop so I can stuff each record into the database? TIA Cheers Don
Re: http://www.nzdsl.co.nz/module-Speedtest.phtml Speed test for Linux...
you could: 1. find a speed test that doesn't require flash or java 2. use wget to run the speed test from a cron job 3. parse the results from the document wget downloaded 4. use grace to make charts in the web server document root your grace script might look like this: [ #!/bin/sh USE=$0 input-data-file output-graph-file [width height] DATAFILE=$1 GRAPHFILE=$2 WIDTH=$3 HEIGHT=$4 if [ ! -f $DATAFILE ]; then echo cannot read input file: $DATAFILE. echo $USE exit 1 fi if [ $GRAPHFILE == ]; then echo $USE exit 1 fi if [ == $HEIGHT ] || [ == $WIDTH ]; then WIDTH=400 HEIGHT=300 fi COMMENT=`head -1 $DATAFILE` ARGS= for i in $COMMENT; do if [ 1 == $ISCOMMENT ]; then if [ != $ARGS ]; then ARGS=$ARGS;; fi ARGS=$ARGS s$j comment \$i\ fi j=$((j+1)) if [ # == $i ]; then ISCOMMENT=1; j=0 fi done gracebat -nxy $DATAFILE -legend load -hdevice PNG -printfile $GRAPHFILE -fixed $WIDTH $HEIGHT -pexec $ARGS -pexec title \$DATAFILE\ -pexec 'xaxis ticklabel format yymmdd' -timestamp chmod 644 $GRAPHFILE ] ..which charts files like this: [ # load smoothed-load smoother-load 2006-10-04T00:00 0.00 0.01 0.00 2006-10-04T00:01 0.06 0.03 0.00 2006-10-04T00:02 0.02 0.02 0.00 ] - neil On Wed, 2006-10-04 at 23:40 +1300, Don Gould wrote: I'd like to set up my server to run a speed test every half hour so I can graph performance. Does anyone know or a linux system that will do this for me? Download Speed: 1091 kbps (136.4 KB/sec transfer rate) Upload Speed: 130 kbps (16.3 KB/sec transfer rate) That's my 3.5mbit link! Cheers Don
Re: regex guru needed...
this rule looks like it must match doesn't it? .*did not issue MAIL/EXPN/VRFY/ETRN during connection to MTA$ the only thing i can think of is trailing space (as Hadley already said), escaping those slashes and maybe some of those spaces are really tabs even though it seems unlikely. does logcheck need the regex to match the whole line? does at least: .*did.*not.*issue.*MAIL.* work? - neil On Wed, 2006-06-28 at 09:17 +1200, Steve Holdoway wrote: 'cos it doesn't work! And I really don't understand why. Grrr. On Wed, 28 Jun 2006 08:30:35 +1200 Carl Cerecke [EMAIL PROTECTED] wrote: Why not just: .*did not issue MAIL/EXPN/VRFY/ETRN during connection to MTA$ Cheers, Carl. On 28/06/06, Steve Holdoway [EMAIL PROTECTED] wrote: To preserve what's left of my sanity, can anyone help me with this logcheck rule? To ignore messages like this, Jun 28 06:02:18 server sm-mta[7813]: k5RI2I6g007813: 23-52-175-62.user.auna.net [62.175.52.23] (may be forged) did not issue MAIL/EXPN/VRFY/ETRN during connection to MTA I added the following line to /etc/logcheck/ignore.d.server/sendmail: ^\w{3} [ :0-9]{11} [._[:alnum:]-]+ (sendmail|sm-(mta|msp|que))\[[0-9]+\]: [[:alnum:]]+:.*did not issue MAIL/EXPN/VRFY/ETRN during connection to MTA$ ...and a couple of hours of variations on the theme. Can anyone point out the glaring error??? Cheers, Steve
Re: Modem problems
I have the same problem. I'm with Xtra at the moment and using a Dynalink Prestige 600 router. I see upstream and downstream noise margins of all zeroes when it craps itself and resetting the modem in the router via the admin interface causes it to come back after a little while, sometimes thirty seconds, sometimes ten minutes. I guess Telecom got a snowflake on a switch somewhere. - neil On Mon, 2006-06-12 at 13:23 +1200, Robert Fisher wrote: Today we are frequently having to cycle the power off then on again on our DSL-302G ADSL modem. Is it likely to be the weather / power or maybe our modem? Anyone else having the same problem? Our power has not gone off but several times it has flickered.
Re: any nfs gurus out there??
i'm shooting from the hip a bit here and i've not tried this out but isn't NFS managed by portmap? portmap has a -i option that binds portmap only to the specified address. is that option any good? - neil On Thu, 2006-05-25 at 10:58 +1200, Steve Holdoway wrote: ... I'm looking for a way to start up the relevant processes so they only listen on one of many ethernet interfaces. Is there a way to do this without firewalling the interfaces I don't want to use? Cheers, Steve
Re: 3 png linux icons.
why does the icon for root have a tie? root ought to know better: http://foldoc.org/foldoc.cgi?quantum+bogodynamics - neil On Fri, 2006-05-19 at 19:48 +1200, jason wrote: Evening all , had somebody overseas send me these today for my site , quite cute icons of root and user based on the penguin. If you can make use of them then please enjoy.
Re: The CLUG wiki
thanks Nick. i'd like to say thanks to everyone who made the wiki happen: hardware, software, bandwidth and content. i think it's a great tool. it's got a bit of an oddity in that it files away minor changes on a different page: http://clug.net.nz/index.php/RecentEdits?days=90 you can also have a backup of the wiki content on request, in case you're worried as i was that the work you put in could go up in smoke. - neil On Wed, 2006-05-17 at 14:25 +1200, Nick Rout wrote: I'd like to present the wiki updating award to Neil Stockbridge who has committed 18 changes to the wiki in the last 90 days - in fact since 28 March. Cheers mate, a job worth doing! Many of them new pages too. http://clug.net.nz/index.php/RecentChanges?days=90
Re: Who donated that dunger? Anyone for a WiFry?
On Thu, 2006-05-18 at 13:17 +1200, Nick Rout wrote: I built one of these about a year and a half ago and got nearly 10km range from Edgeware to Yobbo on the hill. do you use yobbo at all? would you if others were using it? what would you want to transfer around town for free if you had an unlimited account around chch but not to the wider net? i'd like to see all popular distros mirrored in chch. it would be gratifying to have security updates at 54 Mbit/s. - neil
RE: Computer Locking up
it could be: http://www.badcaps.net/ try using a different power supply and then if that fails looking on the motherboard for swollen or discoloured capacitors. - neil On Tue, 2006-05-16 at 10:52 +1200, HappyEvilSlosh wrote: Thanks to Carl, Nick and Volker for the advice. Now an update. I swapped the hard drive with another PC which had a fully functioning install, after which the Hard drive worked fine. I swapped the cables in the PC with ones I knew worked. However I'm unable to run any additional tests as I'm now lucky if the computer boots past whatever the part of booting that allows you into BIOS is called, let alone completing an operating system start up (this goes for LiveCD's as well). --Slosh
Re: How to reset the dns cache?
On Wed, 2006-05-10 at 00:59 +1200, Don Gould wrote: If so, do you know how to reset the dns cache in mdk10? iirc, the data is held in ram, so a shutdown and re-start would be the easiest at this time of day. I shutdown at the end of the meeting... it must be on the disk somewhere i thought restarting the caching name server should work too. if you're convinced it's caching you could try the brute force approach: sudo find /var -type f -print0 | sudo xargs -0 grep -l IP ..where IP is the IP address is the one you think is being cached. this approach is good for find things in /etc too, say if you can't work out what is starting eth1 when you don't want it to: sudo find /etc -type f -print0 | sudo xargs -0 grep -l eth1 - neil
Re: Tuesday 9 May 2006 CLUG Meeting
On Wed, 2006-05-10 at 00:39 +1200, Christopher Sawtell wrote: I also talked very briefly about the Festival/Mbrola/kttsd text to speech system, and demonstrated it. http://accessibility.kde.org/developer/kttsd/ is there a command-line interface to TTS that can be scripted? it would be cool if unhelpful and irritating to have servers speak up when they're running low on disc space. I ran the script program too, the resultant typescript is at: http://shell.clug.org.nz/~chris/typescript-09-Aug-2000.txt why the strange date? - neil
Re: Tuesday 9 May 2006 CLUG Meeting
On Wed, 2006-05-10 at 09:55 +1200, Nick Rout wrote: I also talked very briefly about the Festival/Mbrola/kttsd text to speech system, and demonstrated it. http://accessibility.kde.org/developer/kttsd/ is there a command-line interface to TTS that can be scripted? it would be cool if unhelpful and irritating to have servers speak up when they're running low on disc space. it is in fact a front end to festival, so go back one step and use festival :-) flite was going through my mind during the presentation. that's festival lite, right? is there any scriptable tool for saying things with festival? if so, what's it called? i wrote a Java server that uses JFreeTTS that i can make say things using: echo when thinking about the history of the universe | telnet localhost 31324 ..but the Java takes up a lot of RAM and i'd like a substitute written in C. - neil
Re: Tuesday 9 May 2006 CLUG Meeting
On Wed, 2006-05-10 at 10:09 +1200, Nick Rout wrote: On Wed, 2006-05-10 at 10:05 +1200, Neil Stockbridge wrote: On Wed, 2006-05-10 at 09:55 +1200, Nick Rout wrote: I also talked very briefly about the Festival/Mbrola/kttsd text to speech system, and demonstrated it. http://accessibility.kde.org/developer/kttsd/ is there a command-line interface to TTS that can be scripted? it would be cool if unhelpful and irritating to have servers speak up when they're running low on disc space. it is in fact a front end to festival, so go back one step and use festival :-) flite was going through my mind during the presentation. that's festival lite, right? is there any scriptable tool for saying things with festival? if so, what's it called? i wrote a Java server that uses JFreeTTS that i can make say things using: echo when thinking about the history of the universe | telnet localhost 31324 ..but the Java takes up a lot of RAM and i'd like a substitute written in C. echo when thinking about the history of the universe | festival -tts fantastic! it's: echo when thinking about the history of the universe | festival --tts ..for me (note the longopts double dash), but it works very quickly (less than one second delay before speaking). - neil
server live CD
does anyone know of a server live CD? i want to download a live CD ISO with a system that when booted, configures all network interfaces via DHCP and then starts sshd. if it has parted too that would be great. i have a headless box but no monitor available that i can plug into it. i want to log into the headless box and resize its ext2 partitions. such a live CD should ideally boot without user interaction because i have no monitor to see anything it puts on the screen. cheers, - neil
Re: Tomorrow night's meeting
On Mon, 2006-05-08 at 16:05 +1200, Dale DuRose wrote: Where and when is the meeting tomorrow? it's at the bottom of this page: http://clug.net.nz/index.php/MeetingSchedule ..or from the page: [ St Albans Community Resource Centre, 1047 Colombo Street This is the old Library building at the northern end of Colombo Street. It is quite distinctive and should be difficult to miss. ] the meeting starts at 7:30pm. - neil
Re: Last nights meeting - Thanks Dale and Rik...
On Fri, 2006-05-05 at 13:15 +1200, Don Gould wrote: It was interesting to see the the BSD camp consider a wifi card nothing more than another ip device so it's completely integrated into ipconfig. I am perplexed about why linux requires a whole new subset of tools just to configure a wifi card. Thou I'm sure that in later versions of linux we're going to see the iw range of tools integrated into ifconfig to reduce confusion. google said this first about OSI levels: http://computer.howstuffworks.com/osi1.htm when a cable is plugged into an network ethernet adapter and also into a hub, this is part of the physical layer (OSI layer 1). the air is obviously very different from a cable in that signals radiate in all directions instead of only down the cable. with eleven (or so) different radio frequencies, the air is divided into as many different channels that don't interfere with each other, just as cables side by side but not connected to the same network don't interfere with each other much. this is OSI layer 1 for wireless networking. iwconfig configures the radio inside the wireless network card so it's concerned with OSI layer 1. TCP/IP (the stuff that ifconfig is concerned with) is related to OSI layers 2, 3 and 4, so i guess that's why the tools are separate, not that i'm advocating one approach or another. - neil
Re: multimedia files
On Thu, 2006-05-04 at 15:07 +1200, Roger Searle wrote: Hi, I have several problems with attempting to use various multimedia files - .wmv, .mpg, .mpeg, .mp3, .wma. I'm also unable to get Amarok to play wav or ogg files, perhaps one thing at a time... this may not be what you want but i've always used xmms to play MP3 and OGG although i've never tried playing a WMA. it doesn't do RealAudio either but then what does, apart from RealPlayer? i use mplayer to play video and it plays everything i throw at it (apart from QuickTime maybe). there may be mplayer packages available for your distro. i like the mplayer-nogui package. - neil
Re: Server crashing
i had a similar problem with my EPIA MII-1200. it had been running reliably for just under two years at around 34 C and then suddenly developed this problem: when the power is turned on, it beeps then goes through the BIOS and starts booting Linux normally, then between thirty and ninety seconds after the power has been applied (at a different part of the boot up each time) the screen goes blank and the hard disc light stays on. no seeking sound is heard from the hard disc after this point. pressing any keys on the keyboard has no effect and the machine must be hard reset. the people i bought it from ( http://www.hushtechnologies.com/ ) will replace it because it's -just- inside the two year warranty and they think this may be the problem: http://www.badcaps.net/ they suggested bad capacitors on the mainboard but i suspect bad capacitors in the power supply. either way, Hush must have seen a lot of bad capacitors to share this with a customer. - neil On Mon, 2006-05-01 at 18:25 +1200, Andrew Errington wrote: Hi, Recently I have had trouble with my Linux box crashing (or rather, hanging). Until last week it had an uptime of over a year, but now it has frozen after a couple of hours and a couple of days. Is there any obvious place I can look for 'last thing I was doing before I died'? The symptoms are (this time round) local screen shows last system message (i.e. VGA is still running), local keyboard is unresponsive (CAPS LOCK light won't respond, and CTRL-ALT-DEL won't work). Can't get local console, can't ssh in. Ethernet lamps are on solid. I am inclined to blame the PSU dropping out for some reason (maybe too hot?), but I also heard the UPS beeping plaintively a couple of times recently when in fact there was plenty of power around (i.e. the lights were on...), so maybe that's on its way out. The system is an EPIA-5000 mini-ITX mobo, with 80Gb Seagate Barracuda in a Sereniti 2000 case. Even though this has only happened a few times I marvel at the robustness of the filesystem. Everything comes back with no trouble, and KDE starts up just how I left it. Hints and Tips would be welcomed, but I think it's just entropy. Thanks, Andy
Re: Linux help
i'm not familiar with MEPIS but looking at the comment line: # Dynamic entries below, identified by 'users' option ..suggests the script that is automatically messing about with the /etc/fstab file (and undoing Bernard's original changes from early on in this thread) is using the users option to identify which lines it owns (as automatic mounts) and which ones the user has edited and which the script should therefore leave alone. could someone with access to MEPIS confirm or deny this please? - neil On Mon, 2006-05-01 at 21:18 -0400, Robert Fisher wrote: On Mon, 01 May 2006 04:48, Bernard wrote: I have just installed Mepis and need some help having my Fat32 partitions mounted automatically. I can use KwikDisk and click /mnt/hda2 to mount it but I want to have it mounted automatically every time I boot. My /etc/fstab line looks like this /dev/hda2 /mnt/hda2 vfat,ext3,ext2,reiserfs noauto,users,exec 0 0 What should I change it to? I have been playing with my Mepis laptop on Bernard's behalf and can add that even if I change fstab manually it seems to revert back to the line above after a reboot with the following line above it.. # Dynamic entries below, identified by 'users' option Back to Google for me.
Re: Ubuntu sound problems
On Fri, 2006-04-28 at 11:59 +1200, Stephen Irons wrote: It is not clear to me how various things work together (or not) in linux to make sound work. There seem to be too many layers and alternatives: alsa, oss, esd, jack? OSS is a known interface between application software and physical sound cards and includes OSS kernel drivers for each sound card. ALSA is similar and has mostly made OSS obsolete. ESD is another interface but one that sits on top of either OSS or ALSA and mixes audio from multiple applications (OSS and ALSA support only one application using the sound card at once i think). what's jack? I did notice from /proc/interrupts that interrupt 11 is shared between the soundcard and the PCI-based USB card, and the number of interrupt events for interrupt 11 is zero. This suggests some sort of conflict between the sound card and the USB card at the interrupt level. 4. I thought that PCI was meant to have solved interrupt sharing? I have the BIOS set to non-PnP operating system, with resources allocated automatically. 5. Do I need to change this? did you change this and did it make any difference? - neil
Re: Introducing ... me
On Tue, 2006-04-25 at 22:52 +0100, Malcolm Locke wrote: Actually, that may not be the case. I live 'out in the sticks' in the UK, and its only been in the last 2 years I have been able to get 512K ADSL. Speed is not really that much of an issue for me, but I do have the luxury of an uncapped service which doesn't look like it will be an option. Having said that, I've never monitored how much I transfer so I may not need it anyway! i have uncapped in the UK and thought i would need uncapped in NZ and desperately tried to find it here but i only use about 2 Mbyte each month and i thought i was a reasonably heavy user, although i don't do media files. - neil
Re: Introducing ... me
On Tue, 2006-04-25 at 01:01 +1200, Christopher Sawtell wrote: Telecom provide an ADSL service in most of the main towns and centres, but not in rural areas. I'm told it works. http://www.xtra.co.nz Xtra works. their home page has MSN branding though so you'll be subjected to that when you're looking up your infrastructure details. i only chose Xtra because they're one of only two that offer no minimum contract. most other providers want a twelve month minimum. i'm going to try iHug soon. they at least have an NTP server, which Xtra refuse to even after being asked more than once. the fastest DSL in NZ is 2Mbit down, 128kbit up. reliability is pretty poor compared to the UK (i am with Zen and have had only one outage in six years, which was BT's fault). i've had a few outages over the course of six months with Xtra including one time when they bollocksed the routing to the UK. - neil
Re: Introducing ... me
On Tue, 2006-04-25 at 10:08 +1200, Hadley Rich wrote: You can now get faster, I'm with iHug 3.5Mbit/512Kbit it is a bit expensive though. I don't really need the 3.5 down but it's the only way to get a decent upload rate on ADSL. As for reliability, mine has been reasonable. sorry, Hadley's right. they upgraded to 3.5Mbit/s recently: WAN Port Statistics: Link Status: Up Upstream Speed: 160 kbps Downstream Speed: 4288 kbps - neil
farnell
thanks for the link to Farnell whoever that was: http://nz.farnell.com/ i'll be placing an order on Tuesday next week. does anyone want to get anything from Farnell and get free delivery*? it's usually $10 for delivery. * as long as you keep the weight under 10 kg - neil
Re: LDAP
i'm not recommending LDAP, but i like to use it. i had trouble migrating contacts from one mail program to another so i put them in LDAP instead. i probably use only a tiny proportion of the LDAP feature set but i hope that having my contacts accessible via a standard will make switching e-mail clients easier in the future. i put up some notes: http://www.zen9658.zen.co.uk/sysadmin/ldap/ - neil On Mon, 2006-04-10 at 21:45 +1200, Don Gould wrote: I noticed support for ldap directory services while rebuilding a number of systems and installing thunder bird on them all. I now have my mail set up so that it's all on the server but contact lists are getting lost on each machine and we don't have anyway of sharing contact. In the past I just used exchange server. Is LDAP what I'm after? Does anyone know much about this that they can advise me? Webmin on my cc box has a place holder but nothing seems to be installed. I'm doing a bit of reading on the cc web site but don't want to spend hours reading up on something that's not the right thing. Cheers Don
Re: LDAP
On Mon, 2006-04-10 at 23:22 +1200, Don Gould wrote: then the base dn in thunderbird becomes 'bowenvale' What is a 'DN'? DN is short for distinguished name. it seems to indicate a full path to a name such as: cn=admin,dc=dist,dc=ro (here there is a CN (common name) and a base) ..so i suspect the base name in Thunderbird is the same as the Search base in Evolution in which case it should be dc=dist,dc=ro or in your case dc=bowenvale,dc=pointclark,dc=net. - neil
Re: IMAP Subfolders
Is there a way to make subfolders within folders in a imap store? I use Evolution, which uses right-click then New Folder too, but if you want to it the hardcore dangerous way: $ telnet your-imap-host 143 Trying your-imap-hosts-ip... Connected to your-imap-host. Escape character is '^]'. * OK Courier-IMAP ready. Copyright 1998-2002 Double Precision, Inc. See COPYING for distribution information. a login your-account-name your-password a OK LOGIN Ok. a create INBOX.your-new-folder-name a OK INBOX.your-new-folder-name created. a logout * BYE Courier-IMAP server shutting down a OK LOGOUT completed Connection closed by foreign host. This document has more details: http://www.ietf.org/rfc/rfc2060.txt ..although beware that you're actually talking IMAP using telnet and therefore arbitrarily bad things could happen to your data if you get it wrong. - neil
Re: 64 bit ubuntu support...
there is probably a fix for this since i've heard of the problem before, but the reason i'm going to revert to 32 bit is because WINE doesn't work on 64-bit and because Skype is not yet available as a 64-bit build and after three hours of following other people's ramblings on how to run 32-bit Skype on Ubuntu amd64, i've decided that a 32-bit re-install would be easier. - neil On Sat, 2006-04-08 at 18:22 +1200, Steve Holdoway wrote: I'm having a bit of a problem with my new server. I've built it up with 64 bit ubuntu, and I have 3 problems that I can't seme o find any reason for at all: 1. The clock keeps gaining big time - hours a day. No way ntp can control that. 2. The keyboard keeps repeating keys. I've read that a poor quality KVM can cause that, but why only on this box, and none of the others? 3. It keeps crashing for no reason. The log gets confusing because of the clock, but, for example Apr 8 19:36:34 localhost dhcpd: DHCPREQUEST for 10.0.0.202 (10.0.0.201) from 00:00:24:c6:1e:74 via eth0 Apr 8 19:36:34 localhost dhcpd: DHCPACK on 10.0.0.202 to 00:00:24:c6:1e:74 via eth0 Apr 8 18:13:26 localhost syslogd 1.4.1#17ubuntu3: restart. Apr 8 18:13:26 localhost kernel: Inspecting /boot/System.map-2.6.12-10-amd64-k8-smp Apr 8 18:13:26 localhost kernel: Loaded 28582 symbols from /boot/System.map-2.6.12-10-amd64-k8-smp. Apr 8 18:13:26 localhost kernel: Symbols match kernel version 2.6.12. absolutely no error messages at all! I tried building my own kernel, only to find that a) the compile delivered by default (gcc 4.0.1) is not the one used to compile the kernel (suspicious), and b) building a kernel with gcc 4.0.1 creates a solution that will not boot. I suspect that all the problems are tied in to this clock issue. Has anyone suffered a similar problem and got any pointers on how to fix it? My current view is to scrap using 64 bit until it grows up, and reert to 32 bit. Any comments? Cheers, Steve
Re: X-Server question
Rather than buy anything I grabbed one of the junk PC's I've acquired and set it up with enough for it to connect to X. I was wondering if does this mean the junk PC has the X libraries installed along with some X clients and that you run applications by shelling into the junk PC from the everyday PC then exporting $DISPLAY to your everyday PC before running X applications, which then run on the junk PC even though their windows appear on the everyday PC? there is a way to use the local CDROM rather than the one on my everyday PC? Also is there a way to run synaptic locally as well? assuming you're doing the above, any X applications (including synaptic) that run on the junk PC will -only- see the CDROM installed in the junk PC. when synaptic is run on the junk PC, it will -only- manage packages on the junk PC even though it's user interface appears on the screen of the everyday PC. sorry if i've missed the point. - neil
Re: Hint for the Day: User limits and Hog Wild Processes.
that's well useful. can i put it in the wiki or do you want to? - neil On Tue, 2006-04-04 at 11:19 +1200, John Carter wrote: Every now and again a process or program goes hog wild and chews up all the resources on a Linux machine. Its the usual thing, a memory leak or something. Shouldn't happen but it does. Unfortunately modern Linux distributions have an attitude that the user is all glorious and knows what he is doing and should be allowed all resources he possesses. Thus a memory leak can chew up all available ram, all available virtual ram and never let go. Then Linux's dread, and somewhat arbitary, OOM Killer (Out Of Memory killer) wakes up and starts clobbering random processes. Usually the wrong ones. A better administrated box sets ulimit's on resources so that processes that are go wild get clobbered. Here is a set of ulimits you can put in your own ~/.bash_profile to stop the rampage. Comments, suggestions, tweaks welcome. # This is the critical one. Tweak this to be a bit less than your physical RAM. #virtual memory(kbytes, -v) 50 ulimit -v 50 # This tends to drop 'core' files when ever _anything_ seg faults. # How ever, you can then fire up gdb and see why it died. #core file size(blocks, -c) ulimit -c 4 #data seg size (kbytes, -d) 50 ulimit -d 50 # file size (blocks, -f) unlimited #ulimit -f #max locked memory (kbytes, -l) unlimited ulimit -l 10 #max memory size (kbytes, -m) 40 ulimit -m 40 #open files(-n) 200 #ulimit -n 200 #pipe size (512 bytes, -p) 8 #ulimit - #stack size(kbytes, -s) 8192 #ulimit - #cpu time (seconds, -t) unlimited #ulimit - #max user processes(-u) ulimit -u 160
DNS servers
does anyone run their own DNS server? my ISP sent me an e-mail slapping me for leaving recursion on. this is all that's required for bind9: options { allow-recursion { none; }; }; - neil
Re: DNS servers
On Fri, 2006-03-31 at 00:08 +0100, Jim Cheetham wrote: On Fri, Mar 31, 2006 at 09:38:51AM +1200, Neil Stockbridge wrote: http://www.dnsreport.com provides a nice overview of the provision for a domain. thanks, this is a brilliant site. i didn't know about SPF before. - neil
Re: DHCPDISCOVER
On Tue, 2006-03-28 at 23:05 +, Robert Fisher wrote: Thanks, your page led me to edit /etc/dhclient.conf and uncomment and change the timeout from 60 to a lower number. nice one. your dhclient.conf is in a different place to mine. which distro is this on? if it's Debian or a derivative, is it because you have a DHCP client package other than dhcp3-client? Kanotix has a great tool for changing the addresses (or selecting DHCP) for each interface. ..but when we edit configuration files by hand, we can pretend we're one of those elite hackers in the movies. - neil
Re: DHCPDISCOVER
On Thu, 2006-03-30 at 16:35 +1200, Robert Fisher wrote: Yes it is Kanotix I do not know what DHCP client it uses - and the machine is not here at the moment. it looks like you have the dhcp-client package then. the instructions assume the dhcp3-client package but it makes no difference apart from where the dhclient.conf file is found. ..but when we edit configuration files by hand, we can pretend we're one of those elite hackers in the movies. Yes but my original question was - how do I change/shorten the time it tries to find an address from DHCP. i'm confused. is your DHCP working the way you want it to now? - neil
Re: DHCPDISCOVER
thanks for the ifplugd tip. before, i had to manually reconfigure the laptop as i moved around. i've written up some notes that may give inspiration to people on Debian or derivatives: http://www.zen9658.zen.co.uk/sysadmin/net/index.html - neil On Mon, 2006-03-27 at 21:42 +1200, Christopher Sawtell wrote: ifplugd solved the problem for me. emerge ifplugd ( Yes I know you are using something else, you'll just have to translate to its lingo, something like apt-get blah, blah blah, I believe )