Re: [SLUG] RE: VMware 5 Workstation + Linux = kernel panic?
I run vmware 5.0 on breezy/amd64. I have definitely run two vms at once on this system. I was running a team of up to about 6 vms with the same vmware, but can't remember if it was hoary or breezy. I use bridged networking. The team also had internal networks. -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] drivers
On Mon, Nov 07, 2005 at 04:06:16PM +1100, Christopher JS Vance wrote: I've been asked to get a particular touchscreen (eGalax usb) going before Tuesday next week on Breezy. Thanks all. Somebody moved the goalposts, and the target is now FC3, which is covered by the supplied drivers. -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] drivers
I've been asked to get a particular touchscreen (eGalax usb) going before Tuesday next week on Breezy. The manufacturer has actually provided some Linux software in the way of binary bits plus some sources for a number of old rpm-based distributions (RH=9, FC=3, Mdk=10), all for XFree86 and a 2.4 kernel. I see the following issues, with possibly varying degrees of significance: .deb != .rpm 2.6 != 2.4 Xorg != XF86 package names and divisions are all different My questions (to those who might actually know): Does anyone by chance have working bits for this beast? Should this be relatively simple to kludge into Breezy? What gotchas can I expect if I try to do it myself? Should I start from FC3 or one of the other sources (as closest to Breezy)? -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Your top-ten linux desktop apps
On Sun, Oct 02, 2005 at 10:30:36PM +1000, [EMAIL PROTECTED] wrote: Not everyone codes so that time_t is equivalent to int so there's nothing wrong with time_t being a long (and fix the code that can't handle it). Of course not. The synopsis for time(2) on v7 Unix says | long time(0) | | long time(tloc) | long *tloc; precisely because int was 16 bits, and therefore not long enough. Mind you, time_t wasn't even a sparkle in anybody's eyes then... -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Keeping passwords safe
On Wed, Oct 05, 2005 at 08:03:44AM +1000, [EMAIL PROTECTED] wrote: Is anything obvious from the man page? Not on an OS where the people don't care about documentation. -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] package installation initial state
Let's say I have a running Linux kernel with working user space, and I want to construct a minimal chroot environment from a bunch of .deb files (I have the Breezy preview ISO). No optional or recommended, just the bits I say, with their prerequisites. The chroot doesn't need to be bootable, but does need to be self-contained. With RH/Fedora, I can make a few empty directories, run rpm --initdb (or whatever the spelling is), and then use rpm either after chroot or with an option to change the directory it considers as root. What is the equivalent in Ubuntu/Debian? I have come to prefer apt over rpm for upgrading, but it's not obvious how to start the thing out. The man page for dpkg is up(?) to the usual GNU standards for documentation. Or is it significantly easier just to start with a huge installation and try to remove things I don't want or need until there's nothing left to remove? -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Re: package installation initial state
On Thu, Sep 29, 2005 at 12:58:07PM +1000, Matt Palmer wrote: What you want is debootstrap. Does everything you're looking for, but in a single command. I note with some interest that somebody has recently written rpmstrap (name might be slightly different) to provide an equivalent to debootstrap for RPM-based distros. Thankyou. I'll investigate. -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] BSD packages on Linux installation ????
On Mon, Sep 26, 2005 at 06:26:46PM -0700, pesoy misak wrote: I am wondering about FreeBSD tgz binary packages is compatible with linux distros such as debian or fedora. and also can i just unpack the bsd package and run them on any of these two linux distros No. Explanation follows. BSD *.tgz binary packages are gzipped tar files containing a number of control files (each has a name starting with '+') and then the stuff you actually want installed. The +CONTENTS files may contain several @cwd directives which change where different files in the archive get put, so you can't always just untar in /usr/local. (In order to build some BSD packages on Linux, I've just had to hack around what looked like a bug in GNU tar where it couldn't make a BSD package because it didn't obey multiple -T arguments.) But even if you could unpack a BSD package in the right place, it will almost certainly not work for you. One reason for this is that the BSD package will almost certainly expect you to have appropriate libraries installed on your machine. Most of these libraries will be part of the BSD base system, and are not available as or in packages. Your Linux libraries will not be adequate. BSDs also have different system calls from Linux and from each other. Unless you have a Linux kernel which knows how to emulate FreeBSD sufficiently to handle its system calls, you'll be in trouble, even if you could copy the right libraries in. As a side note, I will say that all the BSDs successfully emulate most of Linux, but I've not heard any suggestion that Linux is adjective enough to reciprocate. I build Linux stuff on BSD from time to time. -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] distro footprints
quote who=Christopher JS Vance At the moment we're using Thinstation as a diskless Linux distribution for running Tcl/Tk applications with Firefox, but I'm finding it quite person-hostile for reconfiguration, etc. I wonder what kind of diskspace or memory footprint I'd need if I wanted to replace it with a working subset of a real distribution like Ubuntu or Fedora. I want to use official packages for easy maintenance, but to avoid cruft like Gn*me and K*E. Sound use includes music and VoIP. Hardware is VIA motherboards and CPUs with 512MB memory and a touchscreen. No shell access for regular users, no window manager, etc. Hints about which directions to avoid would also be useful. On Mon, Sep 19, 2005 at 04:57:32PM +1000, Jeff Waugh wrote: You could try out the deeply integrated LTSP functionality in Ubuntu 5.10 (which will be released in mid-October, but available as a Preview release right now, and well worth testing). It'll work with whatever X session you choose to set up. I think you might have misunderstood. We have a file server already, running a different open OS. We want to boot from the server, but run all programs including Firefox and Tk locally on the minimal footprint diskless Linux. No swap. I managed once to get a chrooted Red Hat kernel build environment in only 50 .rpms and that was painful. I guess I could try doing the same sort of thing with Ubuntu, but was wondering how many .debs will it take me to get this kind of stuff running with no recommended packages, just the obligatory ones. And how many MB of ram disk that'll cost. -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] distro footprints
At the moment we're using Thinstation as a diskless Linux distribution for running Tcl/Tk applications with Firefox, but I'm finding it quite person-hostile for reconfiguration, etc. I wonder what kind of diskspace or memory footprint I'd need if I wanted to replace it with a working subset of a real distribution like Ubuntu or Fedora. I want to use official packages for easy maintenance, but to avoid cruft like Gn*me and K*E. Sound use includes music and VoIP. Hardware is VIA motherboards and CPUs with 512MB memory and a touchscreen. No shell access for regular users, no window manager, etc. Hints about which directions to avoid would also be useful. -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] distro footprints
On Mon, Sep 19, 2005 at 04:57:32PM +1000, Jeff Waugh wrote: You could try out the deeply integrated LTSP functionality in Ubuntu 5.10 (which will be released in mid-October, but available as a Preview release right now, and well worth testing). It'll work with whatever X session you choose to set up. I'll give it a look, ta. -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Aptitude command to TOTALLY remove KDE
On Wed, Aug 31, 2005 at 09:47:29AM +1000, Terry Collins wrote: Now all I need to do is remember the pager proggy for twm and I'm set. I'm happy with ctwm. -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] NTPD behind a masquerade
On Wed, Aug 17, 2005 at 01:50:51PM +1000, Terry Collins wrote: I find that a lot of the syncs time out with my main ntp boxen. I suspect that the main servers are extremely overloaded now that every home use has easy to use software to enable them to sysnc with stratum 1 timeservers. Trying to sync with stratum 1 servers is the problem. Check pool.ntp.org. -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DNS Issue
On Wed, Aug 17, 2005 at 05:19:23PM +1000, Kevin Fitzgerald wrote: This is probably an Obvious Question but I'm not sure where to look next. Running Fedora Core 4, set up a DNS server and DNS record for my machine. If I do #host kevnote I receive the correct IP Address. I'm trying to install Scalix Community Server. As part of the install it checks the network settings and comes back with an error Network Check Failed - IP Address associated with host name 'kevnote.tcgtech.dyndns.org is '127.0.0.1' Yet the Host request returns the correct info. Any ideas where I should be looking to rectify the 127.0.0.1 address? Every time I've installed anything from Red Hat or Fedora, it put stupid things in /etc/hosts. Remove your hostname from every line in /etc/hosts which starts '127.0.0.1' or '::1'. Those addresses should have the name 'localhost' with no domainname. -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] recursive tree log grep ?
On Sat, May 21, 2005 at 11:12:22PM +1000, Voytek Eymont wrote: I have tree like: /home/domain.tld/logs/ /home/domain2.tld/logs/ ... how do i run a grep across all logs, as in 'grep a-string /home/*/logs/*' If the logs are all immediately in the relevant logs directory, what you've typed in your question is a good start. If the logs appear at different directory depths, then maybe you need find and xargs. Say something like find /home/*.*/logs -type f | xargs grep a-string -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Re: Weird login behaviour
CONFIG_*=y means 'put this functionality into the kernel at compile time, so that it's always there' If your kernel boots, you've got this stuff. CONFIG_*=m means 'make a separate file which *might* be loaded into the kernel after the kernel has already started' This loading can only be done if the module loading functionality is compiled 'y'. When your kernel first boots you don't have this stuff, but it can be added after boot, as if it were always there. Usually, you can also make it go away again, if you really care. CONFIG_*=n means 'do not put this functionality into the kernel at compile time, and do not make a separate file for possible later loading' If you built with a particular feature being 'n', there's still a possibility somebody (maybe even you) might still compile it separately with 'm'. If kernel module loading is 'y', this stuff could then still be loaded later as a module, even though you said 'n'. Lots of stuff could be marked any of the three ways, while some stuff is either 'y' or 'n', and can't be 'm' (the module loading stuff comes to mind here...). The issue basically resolves to 'did we compile the stuff - and if so, how does it get into the kernel?'. I would normally compile stuff 'y' that is needed to boot, although using an initrd may make such paranoia unnecessary (this depends on how pedantically you read the word 'needed' within this sentence). Some other free OSs take the Linux approach of using lots of modules, while others think it more secure never to do this, and just use bigger kernels with lots of stuff compiled in. At compile time it's not a big deal - unless you make a habit of turning unused things off with 'n', you're still building the stuff either way. At run time, there may be some savings with slightly smaller memory footprints in modular kernels. -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Timezone and Daylight Savings Time
On Tue, Feb 01, 2005 at 01:43:11PM +1100, Jesus M. Salvo Jr. wrote: When I run the 'date' command while we are in daylight savings time, it says: Tue Feb 1 13:35:07 EST 2005 1) Why is it EST ? Shouldn't it be EDT or AEDT ? S for Summer. We don't do Daylight Time here. There are no legal or universal abbreviations for timezones in Australia. I've also seen EASST and EADT. In some places, I think it's also called L (or K in Queensland, right now...). http://www.timeanddate.com/library/abbreviations/timezones/au/edt.html And someone in Norway knows something that Australian governments and people don't? I have checked the following already: 1a) TZ variable: set | grep TZ TZ=Australia/Sydney 1b) /etc/localtime: $) cksum /etc/localtime 1227095042 785 /etc/localtime $) cksum /usr/share/zoneinfo/Australia/Sydney 1227095042 785 /usr/share/zoneinfo/Australia/Sydney Looks good. 1c) dump of the timezone file: $) /usr/sbin/zdump -v /usr/share/zoneinfo/Australia/Sydney snip /usr/share/zoneinfo/Australia/Sydney Sat Mar 27 15:59:59 2004 UTC = Sun Mar 28 02:59:59 2004 EST isdst=1 gmtoff=39600 /usr/share/zoneinfo/Australia/Sydney Sat Mar 27 16:00:00 2004 UTC = Sun Mar 28 02:00:00 2004 EST isdst=0 gmtoff=36000 /usr/share/zoneinfo/Australia/Sydney Sat Oct 30 15:59:59 2004 UTC = Sun Oct 31 01:59:59 2004 EST isdst=0 gmtoff=36000 /usr/share/zoneinfo/Australia/Sydney Sat Oct 30 16:00:00 2004 UTC = Sun Oct 31 03:00:00 2004 EST isdst=1 gmtoff=39600 /usr/share/zoneinfo/Australia/Sydney Sat Mar 26 15:59:59 2005 UTC = Sun Mar 27 02:59:59 2005 EST isdst=1 gmtoff=39600 snip Looks good. Okay, the above file says EST with isdst=1, but again, shouldn't that be EDT instead ? No. EDT is for New York and places nearby. 2) Anyone know how to change the output of date so that it shows the GMT offset instead ? Check the manual pages date(1) or strftime(3) if you have them. Depends on your OS. -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Timezone and Daylight Savings Time
On Tue, Feb 01, 2005 at 02:36:23PM +1100, Jesus M. Salvo Jr. wrote: S for Summer. We don't do Daylight Time here. Huh ? What do you mean we don't do daylight time here in NSW ? http://www.lawlink.nsw.gov.au/crd.nsf/pages/time2 That page doesn't mention daylight time anywhere. It talks everywhere of daylight saving, and doesn't give a name to the timezone. Your proposed abbreviation for the timezone name didn't include an S for the saving part of daylight saving, either. I've heard the timezone most frequently called Summer time around here. I know that you can format the output of the date command. What I was after was having the _default_ ( that is, no options ) output of date to display GMT offset instead of EST. The Unix standard was three characters for a time zone name. POSIX has increased the maximum number allowed to at least six, but it took a while before more than three wouldn't break things. POSIX does say that the timezone name may not include digits, so you won't be changing the default on any systems but your own. Look for zic, zdump, etc. Or just recompile date. Or make a wrapper script. -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Timezone and Daylight Savings Time
On Tue, Feb 01, 2005 at 03:58:29PM +1100, Jesus M. Salvo Jr. wrote: http://www.lawlink.nsw.gov.au/crd.nsf/pages/time2 That page doesn't mention daylight time anywhere. It talks everywhere of daylight saving, and doesn't give a name to the timezone. OK ... I finally see what your original statement meant. The previous page http://www.lawlink.nsw.gov.au/crd.nsf/pages/time1 actually calls it summer time, too, although never in capitals. That's probably as close to official as you'll get in NSW. Hence EST. :-( Understood ... but it just seems that it has been sort of accepted practice that there is a different abbrevation for standard vs. daylight savings time. It's a good idea from a computing viewpoint, but governments don't seem to have made an official Act or Regulation to specify any abbreviations. The words wise and government don't often appear in the same sentence, unless mixed with sarcasm. .. and it says, under Time zone names, that AEDT is the abbreviation I do agree that's probably the most common and sensible 4-letter abbreviation for it, but it's not the only one. I usually just say +1100 (and write dates like 2005-02-01) but AEDT also works for me when the conversion is explicitly about Australia. -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html