Re: [Ltsp-discuss] Re: NeoMagic driver
Try reading the documentation: http://www.ltsp.org/instructions-3.0.html http://www.ltsp.org/documentation/ltsp-3.0-4-en.html#AEN1022 You can use either version of XFree86. It sounds to me like you'd specify XSERVER = auto (for X4) or XSERVER = XF86_SVGA (for X336). NeoMagic chipsets are used for laptops. It should be a PCI connection, so 'auto' should work. Are you sure you already have everything else working with the laptop with LTSP? Jason Begin Original Message From: Joseph [EMAIL PROTECTED] Date: 25 Nov 2002 08:27:29 -0500 Subject: [Ltsp-discuss] Re: NeoMagic driver Does this mean this would be automaticly detected or do I need some setting in the ltsp conf file? On Fri, 2002-11-22 at 23:56, Jason Bechtel wrote: Go to the source: http://xfree86.org/current/Status21.html#21 Jason -- respectfully, Joseph - (606) 477-2355 x140 == End Original Message Race to Save the Primates - every click provides food! http://www.care2.com/go/z/primates --- This SF.net email is sponsored by: Get the new Palm Tungsten T handheld. Power Color in a compact size! http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0002en _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] black screen after X starts
Sorry, I haven't been saving my LTSP messages and I can't seem to find this thread in the archives even though it was just within the last couple days... Someone was having trouble with X going to a black screen. They were concerned about the video card or the font server or some other server-related config. It sounds like they weren't exactly sure where to start. I've been skimming the digests and didn't see a response, so... In my experience, X going to a black screen means that the driver is fine and the video card is fine and there is nothing wrong with the font server. The problem is likely just a matter of finding the right modeline for your video card and monitor combination. Most likely is that the video card is exceeding one of the ranges on the monitor. The monitor recognizes this and just doesn't display anything. To debug, use runlevel 3 and start X manually: # /tmp/start_ws /tmp/x.log # cat /tmp/x.log | less This should let you see what settings X is using. I'm not sure how LTSP is setting up its default XF86Config files these days (or what you may already have tried), but this will let you see which modelines it is eliminating and why and which modeline it finally settles on. Then, find out the physical specs of the monitor. If at all possible, put a multisync monitor on the unit and see if it works with that. You probably just need to reign in the ranges. If you have an old monitor, you may need to generate a custom modeline to hit a particular dotclock right on the nose. You're almost there! Jason Race to Save the Primates - every click provides food! http://www.care2.com/go/z/primates --- This sf.net email is sponsored by: To learn the basics of securing your web site with SSL, click here to get a FREE TRIAL of a Thawte Server Certificate: http://www.gothawte.com/rd524.html _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: neighbor table overflow with 3c905cx-tx-m
Dirk, IIRC, the neighbor table overflow error is due to not having an entry in /etc/hosts (or DNS) for the workstation. Or maybe it was needing the use-host-decl-names on; option in /etc/dhcpd.conf. Perhaps the latter, because it is due to the workstation not knowing its own name. Jason Date: Thu, 19 Sep 2002 12:20:33 +0200 From: Dirk Wirsbitzki [EMAIL PROTECTED] Subject: [Ltsp-discuss] neighbor table overflow with 3c905cx-tx-m Hi, I successfully flashed a 3c905cx-tx-m with Etherboot 5.0.7. The 3c905cx-tx-m ist the successor of the 3c905c-tx-m and Vendor and device IDs are the same. The rom starts and the machine boots up fine (means dhcp and tftp works) until it tries to mount the nfs resources. Then it gets stuck with a neighbor table overflow error. I already tried to change the PCI slot but no go. I never hab problems like that with the older 3c905 cards. Is this a problem with the 3c905 driver? I=B4m still using 2.x and I can=B4t change to 3.x right now. Best reagrds, Dirk Wirsbitzki -- --==- Jason Bechtel, Software Engineer | (419) 861-3331 Unique Systems, Inc. | Mailto:[EMAIL PROTECTED] 6920 Spring Valley Drive, #106 | http://www.uniqsys.com Holland, OH 43528| UNIX/Linux Solutions for Business --==- Think-Linux The Solutions Show October 30-31, 2002, Toledo OH www.think-linux.net --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: Abiword Font Error with LTSP client and server using SuSe 8.0
Jim, Now that's a helpful set of information... :-) From: Jim Rich [EMAIL PROTECTED] Date: Thu, 19 Sep 2002 11:09:12 -0700 /etc/X11/XF86Config-4 does not exist but /etc/X11/XF86Config does exist and changes could be made to this file if one ignores the line in the file that says VaX generated file As long as you don't rerun the X configuration stuff in SuSE, you will be fine ignoring that line. If you are worried about it being overwritten, make a backup copy after you've modified it. /etc/rc.d/init.d/xfs does not exist but /etc/init.d/xfs does exist but the abiwordfont.txt file said to change two lines that do not exist in this file the only other xfs file I found was a binary file (executable) /etc/init.d/xfs is the file you want. If those lines don't exist, it's probably because SuSE likes to put everything in /etc/rc.config. Look in there for xfs options: cat /etc/rc.config | grep -i xfs Look at the change that the abiwordfont.txt document is telling you to make. It wants you to replace a '-1' argument with a '7100' argument for the '-port' option to xfs. If you can find out where SuSE sticks its xfs command-line options, you can make this modification. Or, show us the /etc/init.d/xfs file and we'll tell you what to do with it. If you choose this option, please also provide the relevant lines from /etc/rc.config. Restarting the X font server as stated in the abiwordfont.txt file did not work i.e. service xfs stop service xfs start error msg was : service: command not found The 'service' command is a Red Hat Thing(TM). You can just do the following on SuSE: /etc/init.d/xfs stop /etc/init.d/xfs start I have even tried to copy the abifonts files to the a /opt/ltsp/i386/usr/share/AbiSuite/fonts directory This does not match up with the strategy that the abiwordfont.txt document takes. From the document: This problem results when the X font server can not serve up the fonts that AbiWord requires. The solution is to get the Xserver to communicate with the LTSP terminals through a TCP port so that LTSP terminal can connect to the X font server (xfs), and to set the X font server's path to find the AbiWord fonts. This is only one solution. The other solution is to move the fonts like you just did, and then leave xfs disabled and instead modify /opt/ltsp/i386/etc/rc.setupx (and rc.setupx3) to build the local (to the workstation) XF86Config file with the extra FontPath information. Look for the section that lists the paths in which to search for fonts: Section Files RgbPath/usr/X11R6/lib/X11/rgb FontPath /usr/X11R6/lib/X11/fonts/misc/ FontPath /usr/X11R6/lib/X11/fonts/Type1/ FontPath /usr/X11R6/lib/X11/fonts/75dpi/ FontPath /usr/X11R6/lib/X11/fonts/100dpi/ EndSection and add your path so it looks like this: Section Files RgbPath/usr/X11R6/lib/X11/rgb FontPath /usr/X11R6/lib/X11/fonts/misc/ FontPath /usr/X11R6/lib/X11/fonts/Type1/ FontPath /usr/X11R6/lib/X11/fonts/75dpi/ FontPath /usr/X11R6/lib/X11/fonts/100dpi/ FontPath /usr/share/AbiSuite/fonts/ EndSection Now you've only had to modify two files and should get the same result. This is unclean, though, because you've now created two copies of the Abiword fonts and created non-standard LTSP modifications (i.e. outside of lts.conf). So, while it modifies fewer files, this solution must be maintained from one LTSP release to the next. If you ignore all of this about moving fonts around and instead just take the advice about mapping the doc onto SuSE's way of doing things, you'll be just fine. Hope this helped. Jason -- --==- Jason Bechtel, Software Engineer | (419) 861-3331 Unique Systems, Inc. | Mailto:[EMAIL PROTECTED] 6920 Spring Valley Drive, #106 | http://www.uniqsys.com Holland, OH 43528| UNIX/Linux Solutions for Business --==- Think-Linux The Solutions Show October 30-31, 2002, Toledo OH www.think-linux.net --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: Fonts: how to tell if everything is correct ?
My point is: - I have made a link from the /opt/ltsp/i386/usr/X11R6/lib/X11/fonts/ to = /usr/X11R6/lib/X11/fonts. That doesn't work (I get a text screen on the = client that local X was unable to load DEFAULT font) - Why do I have two font directories ? Do I need to install all fonts in= duplicate directories ? - How can I tell what fonts are used under a session on the server / on t= he client ? (in order to reproduce on the client what happens on a direct= login on the server) - Are fonts some window manager config ? If so, how come I don't get the= same fonts, since I mount the same /home/USER/ with its .gnome and other= .* files ? - Why did the link (above) not work ? The client should be fooled to hav= e the same fonts as what happens on the server, no ? Marcel, The link above did not work because NFS mounts do not traverse links. Imagine if a user who could remotely mount their home directory decided to create a link to /etc in their home directory. Then they could just cd directly into that system directory! While file permissions would hopefully prevent them from doing anything bad, NFS is made to work like this *on purpose*. It's a feature, not a bug. :-) What you can do is just turn the link around. Actually move all of the fonts to the LTSP tree under /opt/ltsp/... and then point the main system directory at that: /usr/X11R6/lib/X11/fonts - /opt/ltsp/i386/usr/X11R6/lib/X11/fonts/ This allows you to keep one copy of your fonts and everything keeps working smoothly. To tell what fonts are being used, refer to the commands 'xlsfonts' (CLI) and 'xfontsel' (X11). Jason -- --==- Jason Bechtel, Software Engineer | (419) 861-3331 Unique Systems, Inc. | Mailto:[EMAIL PROTECTED] 6920 Spring Valley Drive, #106 | http://www.uniqsys.com Holland, OH 43528| UNIX/Linux Solutions for Business --==- Think-Linux The Solutions Show October 30-31, 2002, Toledo OH www.think-linux.net --- Sponsored by: AMD - Your access to the experts on Hammer Technology! Open Source Linux Developers, register now for the AMD Developer Symposium. Code: EX8664 http://www.developwithamd.com/developerlab _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: SIS 6215
It sounds like the XSERVER setting is fine. I believe these cards work with XF86_SVGA. I'm not sure about which XFree86 v4.x driver they use... Ah! According to xfree.org docs (http://xfree.org/3.3.6/SiS.html and http://xfree.org/4.1.0/SiS.html) they are indeed under SVGA. Now, it sounds to me like you just need a working modeline! See the URLs I provided for tips on building a suitable modeline. Jason Date: Tue, 10 Sep 2002 12:27:53 +0700 From: diwakoe [EMAIL PROTECTED] Subject: [Ltsp-discuss] SIS 6215 my lts is working fine, but i have one WS with an sis 6215 chipset VGA Card can not display XScreen, is there any setting for this SIS VGA Card at lts.conf, which XSERVER i can use, i already XF86_SVGA still can not display with error screen not define. -- --==- Jason Bechtel, Software Engineer | (419) 861-3331 Unique Systems, Inc. | Mailto:[EMAIL PROTECTED] 6920 Spring Valley Drive, #106 | http://www.uniqsys.com Holland, OH 43528| UNIX/Linux Solutions for Business --==- Think-Linux The Solutions Show October 30-31, 2002, Toledo OH www.think-linux.net --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: Mouse Scroll wheel
imps/2 From: Joseph [EMAIL PROTECTED] Date: 10 Sep 2002 12:21:46 -0400 What protocol would it take to make a scroll wheel work on a client with Microsoft Mouse? -- --==- Jason Bechtel, Software Engineer | (419) 861-3331 Unique Systems, Inc. | Mailto:[EMAIL PROTECTED] 6920 Spring Valley Drive, #106 | http://www.uniqsys.com Holland, OH 43528| UNIX/Linux Solutions for Business --==- Think-Linux The Solutions Show October 30-31, 2002, Toledo OH www.think-linux.net --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: PCI-Problem
Ernst, I believe I once had to set up 4 Compaq Deskpro systems (model 2100, I think) as LTSP workstations. I also installed a PCI video card in each one because the onboard one was either not supported or had insufficient video RAM. Make sure you have gone into the BIOS and disabled the onboard video! XFree86 will not know which one to use unless you do this. Jason From: evonklein [EMAIL PROTECTED] Date: Thu, 5 Sep 2002 14:49:07 +0200 Subject: [Ltsp-discuss] PCI-Problem E. Klein I have a Compaq Deskpro as DC. It contains one PCI-slot. I put different cards into this slot (e.g Mach64, SVGA, S3,..). None of them worked with xf86config-4, only with xf86config. During the boot sequence I got these messages: . . . PCI: BIOS32 entry (0xc00fa000) in high memory, cannot use PCD: Using configuration type 1 PCI: Probing PCI hardware PCI: cannot allocate resource region 0 of device 00:00.0 PCI: cannot allocate resource region 1 of device 00:00.0 PCI: cannot allocate resource region 2 of device 00:00.0 PCI: cannot allocate resource region 3 of device 00:00.0 PCI: cannot allocate resource region 4 of device 00:00.0 PCI: cannot allocate resource region 5 of device 00:00.0 got res[1000:101f] for resource 0 of PCI device 1022:2000 . ... some more messages . (II) Module mouse:vendor: The XFree86 Project compiled for 4.1.0, module version 1.0.0 (EE) No drivers available Fatal server error no screens found I'm very confused about it. The first time I thought, a damaged grafic card is the reason. --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Re: Do what with /etc/hosts?
Kris Adcock wrote: Sent: 06 September 2002 18:27 Subject: [Ltsp-discuss] Re: Do what with /etc/hosts? I think you've got your columns reversed... It goes IP_Address, then white space, then Hostname1 Hostname2 etc. But it should say But that is what I should be adding to /etc/hosts, then? An entry for each workstation? Right. My lines usually have IP hostname hostname.domainname like this: 10.1.1.1 ws001 ws001.domain.net Jason --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Using extra VRAM
True. I don't think AGP and EDO coexisted much, if at all. It's been a while (couple years) since I've looked at high-end video cards, but there was a time that one could find PCI versions of some of the latest cards. Perhaps no longer... A 16MB video card might actually just do the trick for a system with 16MB in SIMMS already. LTSP isn't reliable with 16MB, but is rock solid with 32MB (X isn't *that* bad). So, leaving 4MB to actual video and giving the remaining 12MB to the system as swap brings you to 28MB effective RAM. Not bad... :-) I agree, though, that this would be a possibility for only a fraction of a percent of the LTSP community. Jason On Mon, 2002-09-02 at 21:30, [EMAIL PROTECTED] wrote: Jason, I saw that article, and found it interesting as well. But, aren't the EDO simms rather old technology ? And, if you've got an older motherboard that requires EDO memory, then you probably don't have an AGP slot for a fancy video card. You're not likely to find an nvidia PCI card. You probably could find an ATI PCI card with 8 or 16mb of ram, but that doesn't make a great swap device, as it's rather small. So, maybe there is some application for this vram usage in LTSP, but I think it will be really small number of people who can actually take advantage of it. Jim. On Mon, 2 Sep 2002, Jason Bechtel wrote: You're all probably avid Slashdot readers like me and have or will soon see this article: http://slashdot.org/articles/02/09/02/2321211.shtml?tid=137 which links to this article: http://hedera.linuxnews.pl/_news/2002/09/03/_long/1445.html which details how to map your unused video RAM as a usable memory space (MTD = Memory Technology Device) under Linux w/ XFree86. Of course, I immediately thought of the LTSP! I've had several occasions where I had an older machine that only took expensive EDO SIMMS which one had to special order and had an onboard video card that either didn't play well with Linux or didn't have enough VRAM. With this technique, I could avoid having to buy SIMMS for the thing (16MB just isn't enough for LTSP with X) and just buy a higher end graphics card with a bunch of memory and use it as swap! The VRAM swap would be a hell of a lot faster than local hard disk or NFS swapping... It would be almost like adding that RAM directly to the system! In this situation I could also have used the otherwise ignored VRAM associated with the onboard video card as additional swap... The process described could be automated to some degree and reduced to a simple LTSP variable, VRAM_SWAP. :-) Jason --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Using extra VRAM
You're all probably avid Slashdot readers like me and have or will soon see this article: http://slashdot.org/articles/02/09/02/2321211.shtml?tid=137 which links to this article: http://hedera.linuxnews.pl/_news/2002/09/03/_long/1445.html which details how to map your unused video RAM as a usable memory space (MTD = Memory Technology Device) under Linux w/ XFree86. Of course, I immediately thought of the LTSP! I've had several occasions where I had an older machine that only took expensive EDO SIMMS which one had to special order and had an onboard video card that either didn't play well with Linux or didn't have enough VRAM. With this technique, I could avoid having to buy SIMMS for the thing (16MB just isn't enough for LTSP with X) and just buy a higher end graphics card with a bunch of memory and use it as swap! The VRAM swap would be a hell of a lot faster than local hard disk or NFS swapping... It would be almost like adding that RAM directly to the system! In this situation I could also have used the otherwise ignored VRAM associated with the onboard video card as additional swap... The process described could be automated to some degree and reduced to a simple LTSP variable, VRAM_SWAP. :-) Jason --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] mouse
Tom, You have the info you need. Just put imps/2 in for the X_MOUSE_PROTOCOL in lts.conf and the ground work is complete. ZAxisMapping is hard-coded into the recent (v3) LTSP rc.setupx files, so you don't need to worry about it. The only additional steps are telling your apps to listen for the Z-Axis buttons (4 and 5). This is an app-by-app thing. Some support it natively (StarOffice, Mozilla, ...) but others require some settings in each user's home directory in a config file. I'm not sure of the name off the top of my head. I think it's '.Xdefaults'... Jason Date: Tue, 20 Aug 2002 08:37:34 -0400 From: Tom Allison [EMAIL PROTECTED] I have a three button MSFT mouse with wheel for both my PC and my ltsp terminal. One ltsp doesn't have a 3-button mouse. logged directly into the PC, the mouse + wheel work fine. under ltsp it's a 2 button mouse. Device: = /dev/psaux (correct in ltsp.conf) Type ImPS/2 (ltsp.conf = PS/2) Is this it? X_MOUSE_BUTTONS=3 I don't know what to add for Z-Axis mapping either.. --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: Minimal Firewall Rules?
This is the list of holes I had to open up on SuSE to get LTSP working in VMWare: tcp,69 udp,69 tcp,2049 udp,2049 tcp,369 udp,369 tcp,111 udp,111 tcp,1028 udp,1028 tcp,177 udp,177 udp,514 Jason --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] CrossOver Office Server Edition
People using WinTSE or Citrix, Just pulled this from one of the many reports on LinuxWorld. It sounds like you can run Windows apps to terminals from a Linux server running this product. It costs $4,895 for a 100 concurrent user license. An additional 100 users cost $2,995. This sounds like a lot, but I think WinTSE is like $20,000 or something absurd like that. So, you may be even less tied to Windows than you think! You can use Linux or Windows on user desktops and they can have access to their necessary Windows apps from a *Linux* server instead of a Windows server! Jason CrossOver Office® Server Edition Version 1.2: http://www.codeweavers.com/about/press_releases/?id=20020813 --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] hard drives mice
Yes, mouse settings are controlled by the XF86Config file, but on an LTSP workstation this file is (normally) built at boot-time and optionally takes various X-related settings from the lts.conf file. So, before resorting to a static XF86Config file, explore the existing options... First, you have X_MOUSE_BUTTONS (the docs call it X_BUTTONS; I'm not sure which is right off-hand), which takes an integer argument (you probably want 3). You will probably want the more common X_MOUSE_PROTOCOL option (you will probably specify IMPS/2) and the X_MOUSE_DEVICE option (/dev/psaux will probably be right for you). It's all in the docs: http://www.ltsp.org/documentation/ltsp-3.0.0/ltsp-3.0.html#AEN1022 I also recall that the rc.setupx (or rc.setupx3 for XFree86 v3.3.6) automatically includes the ZAxisMapping 4 5 required for wheel mice. The rest is up to you just as it would be on a normal system. You need to tell the apps to obey the wheel. Do a Google search for more on this. Also, I might mention a little thing I released a while back with little fanfare: LTSP Mouse Autodetection. This LTSP add-on allows you to put auto in for some of the lts.conf X_MOUSE_* options. It only adds a couple hundred Kb to the lts tree and ~2 seconds to the boot time (only for stations who receive these auto commands, of course). I aim to reduce the footprint in the next release. For now, it's available at the following URL: http://www.uniqsys.com/~jbechtel/automouse Jason Date: Sat, 27 Jul 2002 22:02:10 -0500 (COT) From: John Karns [EMAIL PROTECTED] To: Tom Allison [EMAIL PROTECTED] Subject: Re: [Ltsp-discuss] hard drives mice On Thu, 25 Jul 2002, Tom Allison said: 2) How do you tell the server about a 3button + wheel mouse? Or, tell the client about the different mouse options? This is handled in the /etc/XF86Config file. Without a line in lts.conf to explicitly point to an existing XF86Config, LTSP defaults to creating one on the fly. One way then to accomplish your goal would be to find the XF86Config file for the particular client and copy it to some location on the servers hd; then manually edit to suit the mouse cfg, and use the parm in lts.conf (XF86CONFIG =nameOfFile if I remember correctly) to point to the modified XF86Config. --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] best solution
Date: Wed, 24 Jul 2002 16:56:11 +0530 From: Adrian D'Costa [EMAIL PROTECTED] I am being assigned a project for a school. They have a tight budget. Right now the are on a windows 2000 server and 28 nodes. The server is P III 800 Mhz, 64 mb ram, 40 GB hdd and a DSL connection using a usb modem. Yeah, the server RAM will need to be *much* higher, especially if you are planning on users running big apps, like Netscape and StarOffice. For 28 clients, figure at the bare minimum 256MB and try to get 512MB. See how much the mobo can take. If RAM is cheap, just max it out. Thanks. I was thinking something like 256MB. I was also thinking of Clustering. Have you tried it out? I haven't tried clustering yet. I'm not sure it would relieve the memory crunch. If your server is swapping, however, your users will notice it in the form of slow performance immediately. 256MB will only be enough to get your first 10 or so clients online without swapping. To get the rest online without performance degradation you will need more RAM in the server. At least 512MB, but perhaps 768MB or 1GB. Also make sure the USB DSL modem is supported under Linux before you make the leap. Of course, *always* make sure your hardware is supported before switching to a new system... Well I have some doubts about the USB modem. Hardware is all Intel sytems. Mobo is 830 I think. Red Hat posts their supported hardware lists on their website. If it's not supported explicitly, then just get another one and save yourself time and headaches. The nodes P III, 450 Mhz, 64 mb ram, 40 GB HDD. Some have cdroms some scanners. All the NICs are Rl8193 Comfortable amount of RAM, but not enough for much local apps work. If you can double it to 128MB, that's a start. Again, is it possible to use the local harddisk or is there a way to mount them on the server and utilse them? If you intend to use LTSP, then the terminals are not reliable storage units. The power could be cut at any time. You do not want to be using them for anything other than temporary storage and only for operations that influence the operations of the user sitting at that terminal. Yes, you could mount them with ENBD, but to what end? What data would you trust to be stored there? System data? User data? They're only good for swap under LTSP. As someone else suggested, you might *not* want to use LTSP if you feel you need to use these drives fully. Install Linux on the harddrive of each system and then figure out a way to manage them all centrally. No, I can't recommend this. This is the entire reason that LTSP exists. Maintaining separate complete OS installs on each machine is the management nightmare that has driven most of the people to LTSP. LTSP embodies the idea of centralized management. As little as possible goes on the workstations. You need to choose between ease of management or using all of your existing hardware. The two don't mix. Their only request is that they need to utlise each node's harddisk. only real use for hard disks in LTSP clients is for local swap, which you will definitely want if you want to do local apps. It this easy to setup up? Is it in the documentation? Easy? That depends on your skill level. If you're comfortable with bash scripting and command-line stuff, then it's not bad. Some people have posted things to the list. Not sure who or when. If you can't find the posts in the archive, you could try sending an explicit request to the list for such details and hope that someone feels like posting them again. I don't believe such things are in the documentation as it is not standard LTSP procedure yet. They would be using Netscape, StarOffice, Java and the usual stuff for schools. I have some knowledge in Linux but not successfully in installing ltsp on my home systems. I think I can use local applications and would use each harddisk etherboot to the server using etherboot's lilo option. So I need you people's assistance in what would be the best way. I guess you could use the HDs for that and for swap. I don't understand what you're asking with regard to the best way though. Do you want details of how to set it all up? As far as I recall, local swap is not a simple option The best way to use the local harddisks as the client does not want it to go for waste. Yes I need some way to set up local applications. Local applications in the LTSP sense means applications running in the RAM and CPU of the workstation. By definition, LTSP workstations are diskless and have no hard drive and therefore nothing is actually installed locally (they just run locally). Local apps is a standard part of LTSP and is well supported. This does not use the hard disks in the clients, however. I've heard of new mobos that allow up to 8 IDE connectors. You could build a couple of LTSP servers with IDE RAID and do load balancing between
Re: [Ltsp-discuss] XDM Problem
Stephen, It sounds like you've got the service configured properly and the restrictions released. Just make sure there's no firewalling blocking the packets at the kernel level. Make sure ipchains/ipfilter/iptables is turned off. Jason From: Stephen Berry [EMAIL PROTECTED] Date: Tue, 23 Jul 2002 10:27:17 +0100 Subject: [Ltsp-discuss] XDM Problem I have set up LTSP 3.0.4 /RH 7.2. My workstation boots but ends up with the grey screen with the big X. I have read the docs, /etc/inittab contains x:5:respawn:/etc/X11/prefdm -nodaemon netstat -ap | grep xdmcp gives this. udp17280 *:xdmcp*:*1491/gdm SERVER= is set to the correct IP in lts.conf, along with the line XDM_SERVER= /etc/X11/xdm/xdm-config has this line DisplayManager.requestPort: 0 commented out. Xacess has this line * #any host can get a login window. and [xdmcp] is set to Enable=true in gdm.conf. What am I doing wrong? HELP! --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] best solution
Date: Tue, 23 Jul 2002 15:04:21 +0530 From: Adrian D'Costa [EMAIL PROTECTED] I am being assigned a project for a school. They have a tight budget. Right now the are on a windows 2000 server and 28 nodes. The server is P III 800 Mhz, 64 mb ram, 40 GB hdd and a DSL connection using a usb modem. Yeah, the server RAM will need to be *much* higher, especially if you are planning on users running big apps, like Netscape and StarOffice. For 28 clients, figure at the bare minimum 256MB and try to get 512MB. See how much the mobo can take. If RAM is cheap, just max it out. Also make sure the USB DSL modem is supported under Linux before you make the leap. Of course, *always* make sure your hardware is supported before switching to a new system... The nodes P III, 450 Mhz, 64 mb ram, 40 GB HDD. Some have cdroms some scanners. All the NICs are Rl8193 Comfortable amount of RAM, but not enough for much local apps work. If you can double it to 128MB, that's a start. Their only request is that they need to utlise each node's harddisk. Why? What do they care? If they really want to use the hard disks, they could put them all in the server in a huge IDE RAID array, but those generally aren't reliable. The only real use for hard disks in LTSP clients is for local swap, which you will definitely want if you want to do local apps. They would be using Netscape, StarOffice, Java and the usual stuff for schools. I have some knowledge in Linux but not successfully in installing ltsp on my home systems. I think I can use local applications and would use each harddisk etherboot to the server using etherboot's lilo option. So I need you people's assistance in what would be the best way. I guess you could use the HDs for that and for swap. I don't understand what you're asking with regard to the best way though. Do you want details of how to set it all up? As far as I recall, local swap is not a simple option in LTSP 3.0. You'll have to modify rc.local to probe for the IDE disk, setup the swap partition (or probe for an existing one), and then mount it and enable it. This has all been done before, but it's not standard yet. Jason --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] KDE 3.1a's Desktop Sharing
After everything with x0rfbserver and desktop shadowing, this seemed relevant to LTSP. This is by far not nearly the hole-in-1 solution that x0rfbserver is sure to become, but I thought y'all might be interested. http://static.kdenews.org/mirrors/qwertz/kde31alpha/s6.png More joy, Jason --- This sf.net email is sponsored by:ThinkGeek Gadgets, caffeine, t-shirts, fun stuff. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: how do i install a new x-win manager?
Matt, See my comments below... From: [EMAIL PROTECTED] Date: Thu, 11 Jul 2002 23:38:58 +0100 Subject: [Ltsp-discuss] how do i install a new x-win manager? Ive had a look through the mailing list and couldn't find anything about = this, Im sure it must have been asked before so even a pointer to the = corect posting would be great. It has been asked before, but this is a difficult topic to search on because the keywords are used so heavily on this list. Im new to LTSP (tried it a couple of times in the past and got nowere) = but having just reformated my server and installed RH7.3 I tried again = and it works!! Thanks for not giving up on us! And congratulations on the successful install. :-) After installing and seting up LTSP (kernel 3.0.1) and then found a = window manager that looked like it would do for the more stuborn members = of my family (the other half!) called qvwm-1.1.12-1 that claims to look = a lot more like windoz (yuk!). After installing the RPM I cant find a = way of getting qvwm started on my workstations (it doesn't apear in the = selection list at login). Now I know there is a config file I need to = change somewhere to alow this to work but I cant find anything that = looks right :( The config you are looking for depends on your Display Manager. You are probably using one of these three: xdm, gdm, kdm. If you tell us which one you're using, then someone will be able to give you decent instructions. I havn't run qvwm on the server itself yet, as the normal command I use = to change desktops (switchdesk) dosn't show qvwm ether :( I've never used switchdesk, so I can't help you with that. Have you tried 'man switchdesk' yet? If you can find out where it learns about the available desktops... Jason PS: Please, no HTML mail to the list. Some of us are subscribed in digest mode (like me) or use text-only mail readers. Thanks. --- This sf.net email is sponsored by:ThinkGeek Gadgets, caffeine, t-shirts, fun stuff. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: Client hangs when Loading192.168.0.112/lts/vmlinuz-2.4.9-ltsp-6
Looks like you're missing the colon (:) between the IP address and the path to the file. This is configured in /etc/dhcpd.conf. Of course, this could just be what you typed into the email. Without actual error messages we can't tell you why it's not working. Have you checked the logs on your server? Are you running any firewalling software or do you have a restrictive hosts.allow? Jason From: jgmcbride [EMAIL PROTECTED] Date: Fri, 12 Jul 2002 06:28:02 -0400 Subject: [Ltsp-discuss] Client hangs when Loading 192.168.0.112/lts/vmlinuz-2.4.9-ltsp-6 Client hangs when loading kernel Loading 192.168.0.112/lts/vmlinuz-2.4.9-ltsp-6 This file is in the tftpboot directory. --- This sf.net email is sponsored by:ThinkGeek Gadgets, caffeine, t-shirts, fun stuff. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: Problem with start_ws or syslogd or whatever?
Sebastian, I've never seen this error before, I mean the one from syslogd. How is syslogd being run on your server? It is launched by a script called rcsyslogd (SuSE thing). Do a 'vi `which rcsyslogd`' and look for where the daemon is being started. What command-line options are there? Try restarting the daemon: 'rcsyslogd stop; rcsyslogd start'. Oh, and is 192.168.0.2 your syslog server? As for the XF86_S3 error, it looks like you put XSERVER = XF86_S3 in lts.conf, but you don't have the LTSP XFree86 3.3.6 S3 Xserver installed. Either install it or change your lts.conf entry to read just s3 instead of XF86_S3. This will use the LTSP XFree86 version 4.x Xserver and tell it to use the s3 driver module. Jason From: Sebastian Schulze [EMAIL PROTECTED] Date: Fri, 12 Jul 2002 18:21:01 +0200 Subject: [Ltsp-discuss] Problem with start_ws or syslogd or whatever? I have installed the ltsp-packages on an old PII/333/SuSE 7.3. After solving some minor problems with the nfs and dhcp I have still an=20 important problem on the first (and at this time the only) client (PI/100= ,=20 10Mbit network): The client is booting the kernel, but in runlevel 5 the X-Server is not=20 starting. The last messages on the screen are: Building the XF86config file Building the start_ws script Starting syslogd syslogd: syslogd: cannot write to remote file handle on192.168.0.2:514 /temp/start_ws: /usr/X11R6/bin/XF86_S3: no such file or directory [10 tim= es] --- This sf.net email is sponsored by:ThinkGeek Gadgets, caffeine, t-shirts, fun stuff. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] 2 DHCP Servers on same segment
Follow-up to my own post...
Now that I've had a chance to look at the new kernel from work
(bandwidth!), I see that the root/linuxrc file in the initrd has not yet
been modified to include either of Ken Yap's patches.
Don't despair, though! Without having to recompile dhclient, you can
just apply his second patch yourself and build a new kernel and
reconfigure DHCPD. Here's the outline of what you need to do:
Get the ltsp_initrd_kit package and open it up. Apply the following
patch (copied directly from Ken Yap's email):
--- root/linuxrc Tue Mar 5 14:31:09 2002
+++ linuxrc.alt Fri May 3 20:56:56 2002
@@ -112,7 +112,15 @@
fi
[ ${INITRD_DBG} = 5 ] exec /bin/sh
-cat /etc/dhclient.conf
+ /etc/dhclient.conf
+if [ $DHCP_REJECT ]
+then
+ cat /etc/dhclient.conf
+select-timeout 2;
+reject $DHCP_REJECT;
+EOF
+fi
+cat /etc/dhclient.conf
interface eth0 {
request subnet-mask,
broadcast-address,
Now do whatever is required to build the kernel:
http://www.ltsp.org/documentation/ltsp-3.0.0/ltsp-3.0.html#AEN779
Now, edit your /etc/dhcpd.conf on the LTSP server to add
DHCP_REJECT=ip_of_competing_dhcp_server to your option-129 (strangely,
I can't find details on this, but I've seen it on the list dozens of
times). Make sure you also have option-128 and that it contains the
Etherboot code (NOT a MAC address!).
That should prevent your clients from listening to the NT server on
boot. And, once again, as long as you don't put a 'range' statement in
your LTSP dhcpd.conf, then your Windows machines shouldn't be listening
the wrong DHCPD either.
Jason
Jason Bechtel wrote:
Mandar, Stephen, and others battling with or considering
dual DHCP servers,
While configuring the LTSP DHCP server to only hand out
addresses to particular MAC addresses will save the Windows
machines from trouble, this solution requires that his
existing Windows DHCP server also hand out fixed IP
addresses (to prevent the diskless clients from
occasionally getting a response from the Windows DHCP
server first). If the administrator is resistant to change
and is handing out dynamic IP addresses, then this does not
solve the entire problem.
It's not just a problem of having the diskless clients
getting a response from the Windows DHCP server during the
initial Etherboot DHCP query, either. There's also the
dhclient query when the kernel boots!
FYI, depending on how the DHCP client is configured, it can
wait a pre-determined amount of time and then choose from
multiple DHCP responses. The default behavior of
immediately responding to the first offer is what leads to
seemingly random behavior (sometimes one server answers
first, sometimes the other).
I believe there were some relevant posts a short while back
on the ltsp-developer mailing list. I think it was Ken Yap
talking about how you could place a special string in your
dhcpd.conf file to allow it to identify itself to Etherboot
clients as their preferred server. There was also a way to
instruct the Etherboot clients to look for that string.
Here's the link to the message and the relevant quote:
http://sourceforge.net/mailarchive/forum.php?thread_id=683572forum_id=2542
For people who have to coexist on a network already
running a DHCP server, Etherboot has the
REQUIRE_VCI_ETHERBOOT option which works well for selecting
only the Linux DHCP server for the reply. Unfortunately
dhclient in LTSP will get its info from any DHCP server.
dhclient.conf is apparently flexible enough to implement
the same filtering scheme as Etherboot. I'll work on some
changes to root/linuxrc to implement this scheme. It will
be triggered by the existence of a kernel parameter of
DHCP_VCI=Etherboot-5.0 (in option-129) so it will still
work in the normal case.
A follow-up message includes that patch to the root/linuxrc
and a config example.
http://sourceforge.net/mailarchive/forum.php?thread_id=690001forum_id=2542
According to Ken, the patch would've required a more recent
version of dhclient, however. Just an hour later, Ken
posted an alternative solution, which would work with
existing versions of dhclient:
http://sourceforge.net/mailarchive/forum.php?thread_id=690149forum_id=2542
Here's a lower tech approach which can be deployed right
away. If DHCP_REJECT is defined in option 129 as the IP
address of a server to ignore, then dhclient will not
accept replies from it. So you would set it to the IP
address of the rival DHCP server. This works with the
dhclient in current LTSP initrds.
I don't know if either linuxrc patch made it into the LTSP
3.0.4 kernel to support this, though. Perhaps Jim can
bring us up to date on this.
In any case, it seems that it is absolutely possible to
have LTSP w/ DHCPD coexist with another DHCPD on the same
network. Whether the ability to do that is pre-built into
the new LTSP kernel is another question.
Jason
[Ltsp-discuss] Re: [Ltsp-developer] Graphical booting
I haven't seen 8.0 yet, but my SuSE 7.2 has a nice graphical boot. They must have something similar in their most recent offering which we could use. There are probably similar approaches used in all the major distros nowadays. We could take our pick... Jason Jason A. Pattie wrote: Mandrake 8.2 is using 2.4.18, IIRC. [EMAIL PROTECTED] wrote: Mandar, Yes, we have used the LPP patch (Linux Progress Patch). I didn't include that kernel in the latest ltsp_kernel package, because that kernel is based on the 2.4.18 kernel sources, and there isn't yet an LPP patch available for that kernel. But, if you can use the ltsp_kernel-3.0.1-i386.tgz kernel package, there is a kernel called vmlinuz-2.4.9-ltsp-lpp-5 that will give you that graphical boot screen. Jim McQuillan [EMAIL PROTECTED] On Wed, 10 Jul 2002, Mandar Deodhar wrote: Has anybody devloped a graphical booting for ltsp clients ?? By graphical booting i mean, when i put the boot floppy and clients boots, insted of seeing all the mesages about searching for dhcp server, and the kernel boot up , I see some nice logo with a progress bar below it and when all initilastion is complete, i am directly onto the X windows login screen. It's similar to way i see Windows 98 logo when my windows machine boots up. I have alreday implemented LTSP in my office. But the booting procedure and all messages deters new users from trying out the thin client as they become aware that they are using linux and are skeptical about using it coz of lack of knowledge and awarness about linux Mandar Deodhar --- This sf.net email is sponsored by:ThinkGeek Two, two, TWO treats in one. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] 2 DHCP Servers on same segment
Mandar, Stephen, and others battling with or considering dual DHCP servers, While configuring the LTSP DHCP server to only hand out addresses to particular MAC addresses will save the Windows machines from trouble, this solution requires that his existing Windows DHCP server also hand out fixed IP addresses (to prevent the diskless clients from occasionally getting a response from the Windows DHCP server first). If the administrator is resistant to change and is handing out dynamic IP addresses, then this does not solve the entire problem. It's not just a problem of having the diskless clients getting a response from the Windows DHCP server during the initial Etherboot DHCP query, either. There's also the dhclient query when the kernel boots! FYI, depending on how the DHCP client is configured, it can wait a pre-determined amount of time and then choose from multiple DHCP responses. The default behavior of immediately responding to the first offer is what leads to seemingly random behavior (sometimes one server answers first, sometimes the other). I believe there were some relevant posts a short while back on the ltsp-developer mailing list. I think it was Ken Yap talking about how you could place a special string in your dhcpd.conf file to allow it to identify itself to Etherboot clients as their preferred server. There was also a way to instruct the Etherboot clients to look for that string. Here's the link to the message and the relevant quote: http://sourceforge.net/mailarchive/forum.php?thread_id=683572forum_id=2542 For people who have to coexist on a network already running a DHCP server, Etherboot has the REQUIRE_VCI_ETHERBOOT option which works well for selecting only the Linux DHCP server for the reply. Unfortunately dhclient in LTSP will get its info from any DHCP server. dhclient.conf is apparently flexible enough to implement the same filtering scheme as Etherboot. I'll work on some changes to root/linuxrc to implement this scheme. It will be triggered by the existence of a kernel parameter of DHCP_VCI=Etherboot-5.0 (in option-129) so it will still work in the normal case. A follow-up message includes that patch to the root/linuxrc and a config example. http://sourceforge.net/mailarchive/forum.php?thread_id=690001forum_id=2542 According to Ken, the patch would've required a more recent version of dhclient, however. Just an hour later, Ken posted an alternative solution, which would work with existing versions of dhclient: http://sourceforge.net/mailarchive/forum.php?thread_id=690149forum_id=2542 Here's a lower tech approach which can be deployed right away. If DHCP_REJECT is defined in option 129 as the IP address of a server to ignore, then dhclient will not accept replies from it. So you would set it to the IP address of the rival DHCP server. This works with the dhclient in current LTSP initrds. I don't know if either linuxrc patch made it into the LTSP 3.0.4 kernel to support this, though. Perhaps Jim can bring us up to date on this. In any case, it seems that it is absolutely possible to have LTSP w/ DHCPD coexist with another DHCPD on the same network. Whether the ability to do that is pre-built into the new LTSP kernel is another question. Jason From: Stephen Berry [EMAIL PROTECTED] Subject: Re: [Ltsp-discuss] 2 DHCP Servers on same segment Date: Wed, 10 Jul 2002 13:31:17 +0100 It is possible to have 2 DHCP servers on one segment. That is what we are doing here. Our Win2k server hands out fixed DHCP addresses which means that each client will always have the same IP address. Any unused IPs are locked out so that they cannot be used. The LTSP server also dishes out IPs to our LTSP clients with a fixed IP for a specific MAC address. --- This sf.net email is sponsored by:ThinkGeek Two, two, TWO treats in one. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] segmentation fault
On Tue, 9 Jul 2002, Jason Bechtel said: As I said, with otherwise proven software. I take it Sophos is a precompiled binary. In that case, sure. Your underlying system libraries could have an effect, especially the way Red Hat likes to play with versions of glibc... But in this case it was about the LTSP distributed binaries running on an LTSP client. The system That's the part I wasn't sure about. Actually, I was thinking that the DC perhaps was getting the dhclient binary which resides on the server and executing it from within the LTSP environment, which uses a different set of libraries than those which the binary was compiled under. I wasn't questioning the internal consistencies of LTSP. The client should be getting its dhclient binary from the appropriate NFS-root share from the LTSP server (/opt/ltsp/arch/, where arch is [i386|PPC|Sparc|...]). If it were getting it from the LTSP server's host OS, then you wouldn't be able to network book other architectures using DHCP. Now, I'm not a DHCP hacker, but I would *think* that DHCP isn't limited to the i386 architecture and that the network-layer services provided are ignorant of the underlying hardware. I guess DHCPD could supply binaries for a bunch of known architectures and ask the client which one it is and then send it the right one, but I can't imagine that this is how DHCP works. I thought that the Etherboot code did the first call to DHCP, obtaining the network info along with the path for the TFTP download of the kernel. The kernel itself then makes the next call to DHCP to determine its network info (which isn't passed along when the kernel is network-booted by Etherboot). To do this, the kernel has to be compiled with the necessary options. Therefore, I would conclude that the dhclient code is part of the kernel (probably in the initrd) and was compiled with the kernel for a particular architecture. It does indeed seem to be hardware related, but as far as I can tell, the mobo (ASUS P5A) is to blame. I've swapped the entire client machine (with the same mobo) as well as the hub and patch cables with the same problem. Due to the intermittent nature it's difficult to say with real certainty, but the problem appears to occur only when the P5A has an AGP card installed. I tried two different AGP cards - a generic card employing a Trident Blade3D chip, and an ATI Rage Pro. The problem disappears when I replace the AGP card with an ISA video card. Interesting. Thanks for the update on that. I was a computer tech for a while in college at a local shop. I recall with frustration the hard-to-pin-down hardware conflicts. This is the only downside to the PC architecture. Even with standards out the wazoo, these companies manage to make parts that don't work with certain other parts... But then I ask myself if I could spend $2,500 on an Apple system (to run Linux/PPC, of course) and I have to say that the commodity pricing keeps me coming back. Sigh... I am declaring the test a success though, because I won't be using the ASUS P5A machines as clients for the main part of the project. Although it's still disappointing because I do have a few of the P5A which I was thinking I could use as DC's later on. If it really is just a conflict with the AGP slot, you could still try a decent PCI graphics board. They sometimes cost more, but ISA is in its twilight years... Jason --- This sf.net email is sponsored by:ThinkGeek Stuff, things, and much much more. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: Video Problem with VGA Charger and rdesktop
Hersh, The calculation you need to perform is simple. The number of colors is a power of two. 256 colors is 2^8, which means each pixel requires 8 bits to store its color. (This is what you get with X_COLOR_DEPTH = 8 in lts.conf.) 8 bits is 1 Byte. So, now you take your various resolutions and multiply: 800 x 600 = 480,000 pixels 640 x 480 = 307,200 pixels At 1 Byte per pixel (8-bit color), both of these resolutions can be handled by your 524,288 Byte video card. If you want a higher resolution, you will have to lose colors. If you want more colors, you will have to sacrifice resolution. You say you would be happy with 640x480 and 8-bit color. I think I heard that the Windows rdesktop protocol is limited to 8-bit color anyway, right? According to my calculations, you should be able to get 800x600 at that color level (as testified to by your Win95 configuration). If you're not getting it, then you might need a custom modeline. I don't have the links handy, but check the archives (and Google) for modeline, XFree86, custom or something like that. Your best bet is to put a hard drive in the thing (you said you had Win95 on it) and install Linux. Use all of the tools at your disposal to generate a working /etc/X11/XF86Config file (xf86config, Xconfigurator, xvidtune, etc.). Then copy the file to the LTSP server and follow the LTSP manual instructions for using a custom XF86Config file. You can do it without installing Linux, but it might be more awkward and wind up taking longer. In your original post, you mention having poured lots of hours into this project. Indeed, crafting custom modelines demands time and patience. If your time is valuable, you will quickly exceed the cost of a better video card or even an entire diskless client! Do what's right for you. Good luck! Jason From: Hersh Cristino [EMAIL PROTECTED] To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: Re: [Ltsp-discuss] Video Problem with VGA Charger and rdesktop I would be happy with 640 x 480 with 256 colors. When I run the machine with just Windows 95, I have it set at 800 x 600 with 256 colors. Windows 95 identifies the video card as a ATI VGA Wonder. Jeff Roberts [EMAIL PROTECTED] 07/09/02 10:04AM I'm sure somebody on the list can give you exact numbers, but I can tell you that you're not going to get very many colors or very high resolution with only 512k of video memory. I have some isa cards with 1 meg of memory on them, and I believe the best I can do is 800x600 with 256 colors. So I would guess that you could get 640x480 with 256 colors, or maybe 800x600 with 16 colors. I am in no way an expert, so maybe somebody can answer you with more accuracy/detail. --- This sf.net email is sponsored by:ThinkGeek Stuff, things, and much much more. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] segmentation fault
Julius, As I said, with otherwise proven software. I take it Sophos is a precompiled binary. In that case, sure. Your underlying system libraries could have an effect, especially the way Red Hat likes to play with versions of glibc... But in this case it was about the LTSP distributed binaries running on an LTSP client. The system environment is proven. The libs on the client have nothing to do with the libs of the host OS on the LTSP server. The same LTSP binaries are running all over the planet without segfaults. Therefore, hardware... Jason Date: Sun, 7 Jul 2002 21:22:21 -0400 (EDT) From: Julius Szelagiewicz [EMAIL PROTECTED] Subject: Re: [Ltsp-discuss] segmentation fault Jason, try this for size: same physical machine, sophos mailmonitor and antivirys package dies with segmentation fault under rh7.3 and runs fine under rh7.2. i dare say it is the software, not the hardware. julius On Mon, 8 Jul 2002, Jason Bechtel wrote: For me, seg. fault immediately implies hardware problem (at least with otherwise proven software). See below for more comments... --- This sf.net email is sponsored by:ThinkGeek Oh, it's good to be a geek. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] DNS and NFS server
Date: Mon, 8 Jul 2002 00:17:52 -0700 (PDT) From: NS Jambak [EMAIL PROTECTED] Subject: [Ltsp-discuss] DNS and NFS server I'm wondering whether I would need a DNS server setup on the LTSP server, if there's only a server running LTSP server and several workstations connected through a hub. And what would be the purpose of this DNS server. Should I put the names of the workstation in the /etc/hosts file too? You only need DNS and/or /etc/hosts if you want to refer to your workstations by name in lts.conf. You could also just refer to them by MAC address. You might not even need to refer to them at all if they all take the same LTSP settings in lts.conf, in which case they all use the [Default] section. I don't think there's any other reason, is there? And if there is, /etc/hosts should be enough. I setup a LTSP server with RH 7.2 but I didn't use the ./ltsp_initialize. I couldn't get the same login screen as the one in the documentation but the same login screen as the RH 7.2 login screen. And I could access everything on the server as if I was on the server and it seemedt that the exports file did'nt serve any purpose. It sounds like it's working fine. You're saying that on the client you get the graphical login window, just like on the server? You should notice at least that the options to shutdown and reboot aren't there, but other than that, this is great! The exports file only provides access to the files necessary for the boot process and starting X-windows and for any local-apps. Once you get the XDMCP login to the LTSP server, you have full access to the server because you *are* logged into the server, just as if you were sitting at the console. Jason --- This sf.net email is sponsored by:ThinkGeek Oh, it's good to be a geek. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] segmentation fault
John, For me, seg. fault immediately implies hardware problem (at least with otherwise proven software). See below for more comments... From: John Karns [EMAIL PROTECTED] Subject: Re: [Ltsp-discuss] segmentation fault On Sun, 7 Jul 2002, [EMAIL PROTECTED] said: Hi John, from my experience I had seg faults for different reasons. Hardware related: One was bad ram, the other a bad mobo (on our ltsp-machines). I also had seg faults on my personal machine for different (software related). Check if it's defietely not a hardware prob. (This is from the view of a non-programmer :) - perhaps there's two different kinds of seg-faults) regards, Yes, I hadn't really thought of it from that angle, as the machine is new; but you're right - could be bad RAM or whatever. In any case a somewhat strange problem. I had given thought to hardware in a different sense though, as I swapped the hub and NIC. What doesn't make sense though is that after a successful boot, all runs fine; the client inits to xdm then I can run KDE with no problems at all, so that would seem to rule out hardware. No, it doesn't. The thing about hardware errors is that they are often random and unpredictable. That the boot itself was successful tells you that whatever normally is flaky in the machine is running fine, at least for the moment. My first presumption would be hardware problem, especially with seg faults. One other thing I should have mentioned is that a similar problem happens earlier in the boot process the first time the DHCP server is contacted, *before* the kernel is DL'd via tftp. About 4 instances of the message ALERT: got a fragmented packet - reconfigure your server. The msg comes from etherboot (5.0.6) main.c, just after line 1033 with the comment: Till Straumann [EMAIL PROTECTED] added udp checksum (safer on a wireless link) added fragmentation check: I had a corrupted image in memory due to fragmented TFTP packets - took me 3 days to find the cause for this :-( So I was pretty sure that the the dhcp server or client was at fault, so I updated the dhcp components. Yes, you should've mentioned this. This changes everything. This points at the problem being with the networking, which is still hardware, but not necessarily client hardware. A good test to firm up this assumption would be whether other clients also experience problems like this. If not, then the problem is with your client and could be your NIC, the network cable, the hub port, or even still something to do with RAM/mobo. If other clients are having this problem, then it is either network in general (hub, switch, cabling) or the networking to your DHCP server (cable, NIC, port on the hub/switch). As for the Etherboot code comment, it says to me that Herr Straumann is responsible for adding the code to check for fragmentation, which kindly reported to you the reason for which you may be experiencing problems. It indeed sounds like you have a networking problem somewhere. My only confusion lies in that you say these messages occur before the TFTP download begins, yet the comment in the code seems to discuss TFTP packets... I haven't looked at the code, but it could just be that Herr Straumann had a problem with his TFTP packets, but wound up adding checksum and fragmentation checking for other stages of the boot process as well. Jason PS: In reference to your original post, there is no grounds to suspect any binary incompatibility between your DHCP server's system and the LTSP clients. The DHCPD server is compiled and run on your server. The DHCP client code is run on the client hardware. LTSP ships with an entire set of its own libs now for the client and takes nothing from the server. This precludes anything like what you were considering. As long as the two binaries speak the same language over the network they should behave well regardless of what libraries they are compiled against and even what architecture they run on! Your server could be an Alpha or PPC or Sparc architecture and the client i386-compatible. --- This sf.net email is sponsored by:ThinkGeek We have stuff for geeks like you. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: Ltsp-discuss digest, Vol 1 #668 - 17 msgs
From: Hal Vaughan [EMAIL PROTECTED] Subject: Re: [Ltsp-discuss] Access to local devices Date: Sat, 6 Jul 2002 17:31:21 -0400 On Saturday 06 July 2002 05:02 pm, Jason Bechtel wrote: The other way of handling it would be mount the partition(s) and then copy the filesystem contents onto it using 'tar' or just plain 'cp': mount /dev/hda1 /mnt/temp cd /mnt/temp tar xzf /images/linux_partition_1.tar.gz Actually, that's what I'm going to do. I'm using a perl script just to unite all the steps, but I'm copying it over to an nfs mounted directory on the server. I've been so focused on other issues, I hadn't even thought about this, but I'm glad you mentioned it, because I hadn't even thought of making a tarball. I'd have probably remembered to to that AFTER I had everything running perfectly... On the other hand, it's a lot easier to make simple changes (like adding a perl module) if I don't compact it and just use cp. I'm not sure which would give me the most advantage. You're right. For flexibility sake, don't bother with the tarball. It was really only a step toward compression and saving bandwidth. Since your clients are strong, you could do it, but if your main concern is *your* time (not how long it take for the load to run), then forget it. Do whatever is convenient. One thing, though. If you could share your Perl script and LTSP mods with the list when you're done, I'm sure someone here (or someone searching the archives in the future) would be grateful for it. :-) Happy hacking, Jason PS: Since loading all that data over the network will be the main bottleneck of the loading process, you might want to think about making sure you're network is running smoothly. In particular, tuning the MTU and the NFS packet size might help. --- This sf.net email is sponsored by:ThinkGeek We have stuff for geeks like you. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Access to local devices
Hal, Great! Glad you hear you're making some progress. As for mounting, it depends how you will be imaging the freshly formatted disks. For partitioning and formatting, there is no mounting involved. It's what you want to do with your formatted partitions that matters... If you can put together an image of a partition on the server and place it in an NFS exported directory, then you could just dump it to the raw block device with a command like: cat /images/linux_partition_1.img /dev/hda1 For example, if you put the image in /opt/ltsp/i386/images/, then you could reference the file this way. I'm not absolutely sure how you'd assemble the image off-hand, but I imagine the 'dd' command could do it for you. Try something like this on the server: dd if=/dev/hdb3 of=/opt/ltsp/i386/linux_partition_1.img This should create a bit-for-bit image of the third partition on the primary slave IDE drive of the server. The other way of handling it would be mount the partition(s) and then copy the filesystem contents onto it using 'tar' or just plain 'cp': mount /dev/hda1 /mnt/temp cd /mnt/temp tar xzf /images/linux_partition_1.tar.gz This might actually be better if you have well-endowed client systems because you could create a compressed tarball image on the NFS mount and then save yourself some bandwidth. If you have 486's or low-end Pentiums as clients, however, the decompressing would slow down the whole operation, so you might as well just download the whole binary image over NFS. Good luck! Jason From: Hal Vaughan [EMAIL PROTECTED] Subject: Re: [Ltsp-discuss] Access to local devices Date: Sat, 6 Jul 2002 11:20:42 -0400 Got it working! I realize this will differ for a CD or CDRW, but this is how I got the ha= rd=20 drive recognized and working. I added two modules to lts.config: MODULE_01 =3D ide-probe-mod MODULE_02 =3D ide-disk I was also told to go to runlevel 3, so change the runlevel line in lts.c= onf=20 to: RUNLEVEL =3D 3 I was originally told to at ext2 as well, but there was no ext2 mod and i= t=20 worked without it. I'm not sure what you'd need to do to use ext3 or oth= er=20 file systems. When I asked, on the IRC channel, there was one person tha= t=20 seemed to know this well, and he answered quickly. He didn't stay on the= =20 channel long. Someone else was staying on and chatting with other people= =2E =20 He reluctantly answered my questions about this (I wanted to make sure a=20 mount was not required for this -- I think I did have to add a simple mou= nt=20 line to /etc/fstab, something like /dev/hda1 /mnt/localhd defaults 0 0,= =20 although he told me that would not be required). --- This sf.net email is sponsored by:ThinkGeek Got root? We do. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Access to local devices
Hal, You do not need ENBD unless the controlling scripts are running on the LTSP *server* in which case they need access over the network to the client drive. From what it sounds like, this is not what you want. Many people have modified and played around with the rc.local script in LTSP to get it to do all sorts of things with local partitions, usually to setup local swapping. Search the list archive for some of these keywords: local, swap, ide, partition. There've been lots of posts about this and I think at least one either included or linked to an example modified rc.local script. Basically all you need is to make sure your LTSP kernel includes all the necessary flags to do IDE access. I may be wrong, but I think the default LTSP kernels now include this. Then you just boot your clients into runlevel 3 (via lts.conf) and put your script in /opt/ltsp/i386/etc/rc.local. In this case, the script doing the action is on the client, so you don't need ENBD to access the hard drive. Then to access any additional data from your LTSP server (or some other server), all you need to do is use read-only NFS, which is already part of standard LTSP. Another thing that occurs to me is that you will either have to devote LTSP entirely to this bootstrapping project or find a way to distinguish between normal LTSP clients and new systems that need to be formatted and loaded. The first thing I would try is to configure DHCPD such that known hosts (MAC address hard coded into dhcpd.conf) receive a normal LTSP kernel and boot params, while unknown hosts receive a (possibly modified) kernel and an argument to put it into a special runlevel (say, 4) which could correspond to the bootstrapping script in inittab. Another possibility is to have the default section of lts.conf include a variable which tells rc.local to attempt to bootstrap the client, while known clients all receive a negative value for this parameter. Jason From: Hal Vaughan [EMAIL PROTECTED] Date: Thu, 4 Jul 2002 16:35:38 -0400 Subject: [Ltsp-discuss] Access to local devices I am looking for a good way to access my local devices on my ltsp client.= I=20 am using ltsp to install software on client systems. The idea is that I = can=20 take the client system, plug it in, hook it up to my lan, and turn it on = and=20 ltsp and a few scripts will do everything I need. Part of the setup is formatting a blank hd in the client system. I'm loo= king=20 for ways to access the hard drive on the client system to first format it= ,=20 then copy over the system files I will need, then install GRUB or LILO on= it. I'm currently reading through the documentation on ENBD, but I'm confused= =2E =20 When I set up a box using LTSP, the box booting is the client and the one= =20 supplying all the LTSP info and kernel is the server. I'm not sure, but = it=20 seems to me that once I get this running, since the drive I want to acces= s is=20 on the client, for the purposes of ENBD ONLY, the client is the server. Have others used ENBD? Is this the only way or best way to access local=20 devices? Can I use ENBD to be able to format the drives in the client=20 system? --- This sf.net email is sponsored by:ThinkGeek Bringing you mounds of caffeinated joy. http://thinkgeek.com/sf _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Re: x0rfbserver authentication problem
Well, no word yet from Jens Wagner, so I went ahead and hacked in a -consent option. :-) It's available here: http://www.uniqsys.com/~jbechtel/shadowing/. It was a much more extensive hack than the -stealth one, but it seems to work. As I just wanted to get something working and I'm leaving tomorrow for almost two weeks of travel, this leaves out options like Disable Remote Control and setting a password for the session. You just get a Yes/No dialog. Also, it was easier to just have -consent imply -stealth. A bit more hacking and they could be independent options. Let's also be sure that this is what we want before someone takes the time to do it! :-) So, I'll be on the road and likely incommunicado for a while. Please, take what I've done and keep going. This is so cool! Happy hacking! Jason PS: I totally apologize for not really commenting what I've done. I just whipped it out as quickly as I could. As Jim pointed out, the original code has almost no comments to begin with and I mostly patterned my work after existing code. Again, sorry. I'll clean it up when I get back if things are still moving along. John_Cuzzola wrote: This is the kind of hacking that i'd really like to do, but if you guys wait for me to do it, you'll all grow really old :) I'm sure that someone else (John Cuzzola maybe ?) could hack that into it very easily. *** It wouldn't be all that difficult (maybe next weekend), but it would be nice to get the current changes accepted by the author so this wouldn't need to be re-done with rfb-0.7.0 ___ Sponsored by: ThinkGeek at http://www.ThinkGeek.com/ _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: Dual CPU
Romain, The number of CPUs in the *server* has nothing to do with your LTSP configuration. LTSP is oblivious to the server's hardware configuration and is only concerned with booting the diskless clients and getting them to a log-in prompt (text or graphical). Now, once people at the diskless stations log in to your dual-CPU server and start doing things, the two CPUs start to matter. And, no, you don't have to do anything special. Just make sure your server is running an SMP kernel. Jason From: Romain Surleau [EMAIL PROTECTED] Date: Wed, 12 Jun 2002 12:11:12 +0200 Subject: [Ltsp-discuss] Dual CPU My question is : is there something special to do if my server has two CPUs ? Is it a good idea to do this on that machine (dual PIII 550 Mhz) ? ___ Sponsored by: ThinkGeek at http://www.ThinkGeek.com/ _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: X0rfbserver Citrix Server semi-sol'n
From: Charles Marcus [EMAIL PROTECTED] Subject: Re: X0rfbserver Citrix Server semi-sol'n Date: Tue, 11 Jun 2002 10:01:46 -0400 From: Jason Bechtel [EMAIL PROTECTED] Date: Tue, 11 Jun 2002 01:15:52 +0200 1: Pushing a session -- I believe the original messages regarding the existence of x0rfbserver were closely followed by someone mentioning a presentation in western Canada (Vancouver?) in which people were already doing this. So, yes, x0rfbserver can be used to push a session out for training. This shouldn't be hard to do, either. Once you setup the source session, all you have to do is fire up a bunch of viewers. You could devise any of a number of ways to do this... A new menu option on desktops or an auto-run command on login, if everyone is supposed to see the demo Actually, I think a more secure method would be to simply send out an email with a link that can be clicked to 'join' the published session. The password could be encoded (more secure) and auto-entered, or simply provided as text (less secure) in the body of the email, and which would then have to be entered manually when prompted for. I see... You're concerned about them accessing the session with a password. Look at the file 'rfm_fbs.1.0.html' that comes with the rfb source. It describes the (Remote Framebuffer Macro) file format version 1.0, a human readable scripting language used to send messages to an RFB Server. It includes the ability to script in the password. Also, why do you care about the security of the password if you are *trying* to publish/broadcast the session? The ability to shadow sessions on different LTS networks over a WAN link. For example, say I have 5 LTS networks located in different cities... I would like to be able to shadow sessions in our local office *and/or* in remote offices... Is this gonna require additional coding, or is it capable now? I have no idea about whether this would just work, but if it doesn't, you can always ssh to the remote LTS server (setting up an X-forwarding tunnel with -X) and then run the xrfbviewer/vncviewer there and have its display directed back across the WAN to your desktop. Doing it via an email with a link (app?), this should be possible, as long as the firewalls are set up properly, no? I don't see how you expect the email app to just execute a link as a command, but assuming you could do that... Sure. You'd just have to have the command start the viewer pointed at the appropriate session and send the Remote Framebuffer Macro file containing the password info. Jason ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas - http://devcon.sprintpcs.com/adp/index.cfm?source=osdntextlink _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: X0rfbserver Citrix Server semi-sol'n
Charles, I'm really excited about shadowing, too! I'm also really excited about the comments regarding LEAF+LTSP. This is a great time for LTSP. I sense a breaking story about Linux in the Enterprise using LTSP and LRP/LEAF. The combination of these two advances could make LTSP as powerful as Citrix/WinTSE! With feature parity, the $0 price tag is a jaw-dropper! :-) Anyway... 1: Pushing a session -- I believe the original messages regarding the existence of x0rfbserver were closely followed by someone mentioning a presentation in western Canada (Vancouver?) in which people were already doing this. So, yes, x0rfbserver can be used to push a session out for training. This shouldn't be hard to do, either. Once you setup the source session, all you have to do is fire up a bunch of viewers. You could devise any of a number of ways to do this... A new menu option on desktops or an auto-run command on login, if everyone is supposed to see the demo 2: Passive Shadowing (a.k.a. spying) -- The patches that add the -stealth option to x0rfbserver effectively make exactly what your boss is asking for possible (and it only took John Cuzzola one rainy Saturday to add those). With no intervention on the part of the user, their session can be remotely observed without interfering with the keyboard and mouse. Even if you'd setup encrypted email so that no one in between could decypher it, the boss could just sit back and watch you type it! x0rfbserver is also distributed with a program called rfbcat, which allows you to save x0rfbserver output to a file, thus preserving the spyed upon session! Jason From: Charles Marcus [EMAIL PROTECTED] Subject: Re: [Ltsp-discuss] RE: X0rfbserver Citrix Server semi-sol'n Date: Mon, 10 Jun 2002 08:40:54 -0400 I am very excited by the discussions for implementing shadowing capability using X0rfbserver in LTSP, but had one question and one comment... Question: Will it be very simple to 'reverse' the concept, so that someone (the Admin, or someone in the 'Trainers' Group) could *push* their desktop out to multiple workstations (or all of them) for Training purposes? A little 'Chat' window would come in quite handy too (for both regular Shadowing, and Pushing). Comment: As distasteful as it sounds, my boss has asked me about the possibility of what I'll call 'Passive Shadowing', ie, shadowing someones workstation without the ability to take it over, and *without them knowing it*, so he can see what they are doing. The reason for this is, we have a big problem with people spending lots of time 'playing' on the internet when they should be working, and the boss would like to be able to look in on them from time to time. ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas - http://devcon.sprintpcs.com/adp/index.cfm?source=osdntextlink _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: DHCP: no root-path for new ws, but ok for others
Jean-Louis, I think you should first put your group section inside the corresponding subnet block. Then, put the root-path option inside the group section. It only applies to your LTSP clients anyway. I know the other machines ignore it, but humor me. Then restart dhcpd and see the problem still exists. Jason From: Jean-Louis Leroy [EMAIL PROTECTED] Date: 10 Jun 2002 09:29:40 +0200 Subject: [Ltsp-discuss] DHCP: no root-path for new ws, but ok for others I had two workstations set up properly. I removed one, then later on wanted to add one again. I reused the old station's IP etc, only changed its MAC address (in dhcpd.conf) and video card (in lts.conf). I restarted DHCP. This is all fine, of course. Well the newly installed workstation manages to get its IP and download the kernel, but it stops on No root-path error after Running dhclient. Now this is especially weird: while I was setting things up the kernel managed to boot and X started. It didn't work because I had forgotten to fix the video driver line in lts.conf. I did that, but the kernel froze as I described above. I went back to previous situation - I'm nearly certain! - but now I got that error, persistently. I'm not sure I follow this last paragraph entirely. It sounds like you were letting the workstation boot while you were altering configurations and things somehow worked at one point, but you're not sure why and they don't work anymore... Very confusing. I know this takes forever, but try rebooting the workstation after each change... every time you restart dhcpd and every time you save a new version of lts.conf or rc.local or rc.setupx... If it helps you figure out what is going on, it will wind up saving you time. I know that a common source of error is named misconfiguration. Well, from jimi - the server - both nslookup ws002 and nslookup 192.168.0.102 work and agree: snip Here's my dhcpd.conf. snip ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas - http://devcon.sprintpcs.com/adp/index.cfm?source=osdntextlink _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: X0rfbserver Citrix Server semi-sol'n
Venkat, You may choose to see it that way, but the owner of a company who is paying for the equipment, the IT staff, the office space, the office furniture, and for the employees to sit there and to do actual work, may choose to see it differently. What matters is whose point of view is legally supported. IANAL, and maybe I'm being cynical, but it seems to me that your employer should have every right to observe anything you are doing on the company systems. It can't be an invasion of personal privacy because you shouldn't be doing anything personal to begin with! I agree that it rubs me the wrong way, and it goes against notions of privacy and freedom and democratic society, but the workplace is not a democratic forum. You sign away a lot of rights when you take a job, especially at a big firm with a large, play-it-safe legal staff. Of course, everyone does something personal while at work. When you make a free local call home (in the U.S. at least) on your lunch break, you don't expect the lines to be tapped and your call to be recorded! Someone might argue that there is an expectation of privacy with certain computer activities as well... It's a slippery slope. I wouldn't want to work at that company either, but the truth is that such capability doesn't have to be built explicity into open source software. I'll repeat part 2 of my reply to Charles: The patches that add the -stealth option to x0rfbserver effectively make exactly what Charles' boss is asking for possible (and it only took John Cuzzola one rainy Saturday to add those). With no intervention on the part of the user, their session can be remotely observed without interfering with the keyboard and mouse. Even if you'd setup encrypted email so that no one in between could decypher it, the boss could just sit back and watch you type it! x0rfbserver is also distributed with a program called rfbcat, which allows you to save x0rfbserver output to a file, thus preserving the data! The intention of the -stealth option is to relieve the user of having to do anything when he/she needs help. The administrator can just jump in. This is not a requirement, but a feature that many people will find useful. Many well-intended features can be used against their beneficiaries. Jason Date: Mon, 10 Jun 2002 09:05:14 -0400 From: Venkat Manakkal [EMAIL PROTECTED] Subject: Re: [Ltsp-discuss] RE: X0rfbserver Citrix Server semi-sol'n Charles Marcus ([EMAIL PROTECTED]) wrote: As distasteful as it sounds, my boss has asked me about the possibility of what I'll call 'Passive Shadowing', ie, shadowing someones workstation without the ability to take it over, and *without them knowing it*, so he can see what they are doing. The reason for this is, we have a big problem with people spending lots of time 'playing' on the internet when they should be working, and the boss would like to be able to look in on them from time to time. I see this as unethical and a violation of an individuals privacy and dignity. I would never work for a company that had such equipment. I do hope that such covert features never make it to mainstream open source code. ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas - http://devcon.sprintpcs.com/adp/index.cfm?source=osdntextlink _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: x0rfbserver
Frank! Man, this is great! Thanks for pitching in. Your hints are fantastic! I don't know how long it would've taken me to figure out the SYSV IPC thing... The X0.hosts file recommendation might also come in handy with LTSP v3.1. Thanks a ton! Jason Date: Sat, 8 Jun 2002 13:45:27 -0300 From: Francisco Castro [EMAIL PROTECTED] Subject: [Ltsp-discuss] x0rfbserver I included preliminary support for desktop shadowing in Netstation 0.8.2, here some hints that can help in building a LTSP package. * I have to deal with the Bad Access error in netstation too, the cause was the lack of support for SYSV IPC on the kernel. * The x0rfbserver is launched from another tty like a normal session, to use it the user must switch to that tty, press enter and then switch back to X tty, the password must be entered by the user after doing that (i'm working on safe methods to transfer this password to the terminals). * The X server that x0rfbserver must be attached to is selected by setting the DISPLAY variable before launch it. * To autorize the connection from the x0rfbserver i included a X0.hosts in /etc with the entry localhost inside. I think there is no security problem doing this because i have only one user on the terminal. I hope this can be usefull for people working on LTSP x0rfbserver support :) ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas - http://devcon.sprintpcs.com/adp/index.cfm?source=osdntextlink _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: X0rfbserver stealth sol'n
John, My most heartfelt thanks... While I was strolling around LinuxTag 2002 in Karlsruhe you were working on this vital project. Thank you! Well after spending most of my Saturday playing with 0xrfbserver source code I believe I have added the -stealth option. Here's how: step 1: Compile xclass version 0.63. Get it at: http://sourceforge.net/projects/xclass (this step is probably optional as older xclasses will most likely work but just-in-case) Or, if you are using SuSE, xclass might not be installed (original 7.2 base distro). I'd keep this step in... :-) step 2: Grab untar x0rfbserver version 0.6.1 get it at: http://download.hexonet.com/software/rfb step 3: Grab my patched files for: x0rfbserver.cc OXServiceApplet.cc which replaces those in the x0rfbserver sub-directory you untarred in step 2. (yes I know I should have made them differential patches!). Get those files at: www.sd73.bc.ca/tux/rfb By not making them patches, you've essentially forked x0rfbserver. Now, I know the word fork carries a lot of negative connotation in the Free/Open Source world, so don't get excited, people! All he's done is alter two files in a project. We're still at a point where it would be utterly simple to make diff patches against the original files. The question is then the classic one. Do we go on maintaining a set of patches for this project as the main development progresses? Or do we submit our patches to the author with an explanatory email and get them accepted into the main source? Obviously, the latter option is preferable as it requires (almost?) no ongoing effort to maintain the desired functionality. Hopefully the author will accept these patches. Has anyone contacted him yet? step 4: Compile as per normal as stated in the INSTALL file: make depend make That'll do it. An x0rfbserver -help should reveal the stealth option. This patch also includes the feature of a GLOBAL .x0rfbserver file. If .x0rfbserver exists in / that will be used over ~/.x0rfbserver. This is useful for those of us who run local apps and/or want the client to activate the daemon themselves but not have them enter their own password or options (like disable keyboard mouse). This global file needs to be placed in the root of the client. So in my case that would be /tftpboot/lts/ltsroot; for other's it would be /opt/... If you issue a: This didn't exist in the original version? Maybe I'm confusing two projects, but I thought it looked for a global config file in prefix/etc/ (default /usr/local/etc/). But whatever program I'm thinking of also allowed user-specific .x0rfbserver files to override the global config file, the opposite of what your patch accomplishes. Now, since it is running as root on the workstation and root's home directory is / (I think), the original policy would seem to suit your current choice of /.x0rfbserver just fine... Anyhow, my take on this is that the file does not belong in LTSP's root directory. It either belongs in ltsproot/etc/ or in ltsproot/usr/X11R6/etc/ (if this dir exists). I'm not familiar with the LSB, but I know that Jim is trying to comply with it as much as possible. Also, the main configuration file should not be a hidden file (i.e. should not start with a 'dot'). x0rfbserver -stealth and do not have either a /.x0rfbserver file or a ~/.x0rfbserver file the user will be prompted with the options box so I would recommend running the daemon first without stealth, entering the wanted password, then copying the .x0rfbserver file from that users home directory and use it for your global config. Excellent. Yes, if x0rfbserver cannot run usefully due to a bad or missing configuration, then it should absolutely appear and ask for the necessary info. I would launch the program as described in my previous email with xinit like: xinit /usr/bin/x0rfbserver -stealth /dev/null -- /ltsbin/XF86_SVGA -ac -query ip of server Or perhaps Jim's new system under 3.1 will provide a better way. I believe he specifically implied that it would. There has been a very valid concern of running this daemon when not needed and the overhead. Looking at the source it appears the client goes to sleep and does nothing until there is a connection. Tests with 'top' confirm this. Even with a connection the daemon's cpu usage is typically less than 8%. Again in my tests running with xinit showed no speed difference while the daemon was waiting and even with a connection it was barely noticeable on the workstation. My test environment is still limited to a VMWare virtual workstation and I cannot switch to a console once in XWindows in the workstation because my *real* console intercepts the keystrokes! Doh! :-) So, can you please look into memory usage by x0rfbserver? This is great news about the light CPU-usage aspect, but that isn't my primary concern. If it likewise has a small memory footprint, however, then this thing is a
Re: [Ltsp-discuss] RE: X0rfbserver Citrix Server semi-sol'n
Matt, I don't think local apps should have to be enabled in order to get this to work. They should be totally separate options with no mutual dependency or exclusivity between them. And I think it could be easily set up to work like this... The only thing required is a way to start and stop the x0rfbserver remotely. This could be done by running an rsh server on all workstations. It could also be done by creating a special-purpose daemon just for this task. Then we could implement our own style of security measures in the controlling daemon. Jason On Thu, 6 Jun 2002 16:32:41 -0500 Egan, Matt B. (Artco) [EMAIL PROTECTED] wrote: Would local apps have to be enabled? I'd prefer it if not. I realize that it needs to run on the client machine but is there a way to work around this so that local apps don't have to be setup just for the use of Xorfb? That's not a big deal I just don't have local apps setup yet so it would be more work (not that I'm lazy) ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] RE: X0rfbserver Citrix Server semi-sol'n
Jim, 1) I think that rsh should be avoided on general principle. 2) I don't think local apps and NIS or even LDAP should be necessary in order to get shadowing capabilities. Perhaps these ideals are unrealistic and I appreciate the old maxim about reinventing the wheel, but couldn't we just find some way totally around all the normal hurdles to make it simple and usable? I know that trading security for usability is not right and I don't intend to sacrifice security. With rsh, it was designed to allow one user on one machine to execute commands as another user on another machine, with all the password prompts required. But in our situation, we only have LTSP administrators (people who know the root password on the LTSP server) wanting to run commands on a machine whose only reliable account is root (not assuming LDAP/NIS authentication) and that isn't even password protected. So, do we really need to institute the whole password database system and authentication prompts to accomplish our goal? I don't think so. Ooh! What about this... What about starting a program with xinit as was recently suggested, but instead of starting x0rfbserver it starts a monitoring app? The app listens for a special key combination, say Shift-Alt-PrintScreen (customizable of course) that would start the x0rfbserver session. It would continue listening and stop the process when it receives the signal again, toggling back and forth. This ensures (pretty much) that the user agrees to the remote access. It admittedly goes against my hope that the user wouldn't have to do anything, but it's better than having to launch an app or switch to the text console to launch the x0rfbserver. The x0rfbserver could be rigged to only allow one connection. Assuming the administrator and the user are talking on the phone or are chatting online (somehow in real-time communication), the admin would connect as soon as the user issues the keystroke combination and no one would be able to hijack the session. If someone does manage to connect before the admin, the admin will immediately know it and tell the user to issue the keystroke combination again to shutdown the server. Damn, this isn't the clean solution I was hoping for. Maybe I'm just being stubborn about not using rsh... I honestly think it's not right for this situation, though. What about creating a special account on the workstations (in LTSP's passwd, group, and shadow files) for running remote maintenance commands? Make it userid 1 and call it 'admin' or something. Provide scripts that make it easy for the administrator of the main LTSP server to set the password on the account (and to change it at any time). Then you could rsh in without needing NIS or LDAP or anything complex. Is this stupid? It leaves the workstations open to repeat password attacks. I think this is going to be hard to avoid, though. And it's equally true for rsh with NIS/LDAP authentication. Thoughts? Am I confused about how rsh functions? Jason The problem with running an rsh daemon on the workstations is that it needs to know WHO is sending the request to run something, and whether they are allowed to or not. By turning on local apps, you enable NIS. That is how the rsh daemon will validate that the WHO is a correct user. An alternative to NIS might be LDAP, but we haven't gotten that to work for local apps yet. In the past, some people have suggested just copying the servers /etc/passwd file to /opt/ltsp/i386/etc, but I can tell you very clearly that will NEVER be part of a standard LTSP package. Creating a special-purpose daemon to handle this, in my opinion, is just reinventing the wheel. Use what is already there, and spend your time solving other more important problems. But, I also invite you to challenge me on this. Maybe you've got a better way. ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Citrix Server
Charles, Me, too... :-) I've taken the college-level courses and gotten my degree. I can bang out some C++, Java, and whatnot. I can pick up new languages with a little time and a reference manual. Still, I need a lot more practice at real coding and until I get that experience I prefer simple scripting. I've been reflecting a lot lately on how nice it is that in the UNIX/Linux world most interfaces and functionality are exported to facilitate command-line interaction. This allows almost everything to be scripted. Thus I don't have to learn some new programming API, just some command-line options... Anyhow, the program is not what I expected. I expected a command-line app, but it seems that the more recent versions have made it more like WinVNC. A little window just pops up and the user has to tell it to start exporting the session. This is not what I had envisioned. I think that to get to Citrix/WinTSE parity, we will have to rip out the vitals of x0rfbserver and streamline it. It should take orders only from the admin on the remote machine and not require any steps by the user at the workstation. And there has to be some measure of security involved, such as only allowing connections from the root user on the LTSP server. Unfortunately, I'm just getting started on a little project here and I have limited time remaining in which to accomplish something. So, I don't think I have the time for something that involved right now. This is something that I desperately want to see in LTSP, though. If someone else is thinking along the same lines I am and wants to take a whack at it, please let us know that you're working on it. We don't want any duplicated effort. Jason From: Charles Marcus [EMAIL PROTECTED] Date: Wed, 5 Jun 2002 14:53:18 -0400 From: Jason Bechtel [EMAIL PROTECTED] Date: Wed, 05 Jun 2002 11:58:20 +0200 I don't see any need to throw money at the problem... yet. I will take a look at it this afternoon (GMT+1) and see how it looks. It could be as easy as dropping a few binaries in the the NFS root tree and writing a script on the admin end. Jason You're a good man, Jason... However, if it starts looking like more work than you initially thought, my offer still stands... Man, I wish I had time to dive into learning how to code... Charles ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] RE: X0rfbserver Citrix Server semi-sol'n
John,
Sounds like lots of good news. Nice work!
Date: Thu, 6 Jun 2002 09:13:18 -0700 (PDT)
From: John_Cuzzola [EMAIL PROTECTED]
*** Hi All,
I have been using x0rfbserver and it CAN work with
remote
(non-local) sessions (with caveats).
snip
First set your lts.conf to runlevel 3 (or whichever ltsp
3 uses) so that
it drops you into a shell prompt as root.
Okay, this is obviously not going to be a part of the final
product. :-) As you mention later, this can all be put
into an rc file and run automatically each time.
ok now launch your X session like this (this is one long
line):
/usr/X11R6/bin/xinit /usr/bin/X0rfbserver /dev/null --
/ltsbin/XF86_S3
-ac -query your ltsp server ip
So, *all* workstations would *always* be running a
x0rfbserver? This is an undesirable requirement. This
puts a load on many terminals that may never need shadowing
help. Even for those who might occasionally need help,
they shouldn't have to always be running the x0rfbserver,
even if it is hidden. Some workstations are lowly 486's
with minimal RAM...
This will launch x0rfbserver and give you a login screen.
You'll notice
that x0rfbserver will ask you for the password. Just
leave it blank for
now. To get rid of it simply run x0rfbserver as some
user. Enter the
password then COPY the .x0rfbserver file from that users
home directory to
the ltsp root. For me that would be:
/tftpboot/lts/ltsroot
Cool trick! Good work! :-)
The x0rfbserver program can be shutdown by the user.
However, if they
do their X session will terminate and send you back to
the login
screen. (after changing your runlevel back to 5). I guess
they would learn
very quickly not to shut the thing down. I tried some
simple stuff in
order to hide it like issuing a -geometry directive and
moving the
x0rfbserver box outside the viewing area but apparently
it doesn't honor
it. I suppose it wouldn't be to big of a deal to edit the
source code and
attempt to hide it (which I may look at when I get a free
moment - I'm
sure the very capable programmers on this list can hack
something up
quicker than I)
This is what I was talking about in my last message. We
don't want any visible face on this program. And in order
to remove the X interface from it, we'll need to modify the
source code.
I think our next step should be to contact the author(s) of
x0rfbserver and explain our requirements. Probably the
easiest way to get what we want is to ask for a feature and
get approval from the maintainer. We don't want to change
the code and then tell the maintainer that he should
incorporate our modifications or face a fork! If the
author has no time or interest in our mod, he or she could
at least recommend the best way to go about doing it, both
from a purely technical standpoint (which source files to
modify) and from a logistical standpoint (perhaps adding it
as a plug-in type feature instead of as a large
modification to the base code. I don't know, but let's
just contact the author first...
I envision a simple command-line arg that disables the
interactive components (X interface). The invocation of
the app would then just be a script run by the
administrator like this:
# the following line returns, leaving
# x0rfbserver running on the workstation (hopefully)
rsh ${WKSTN} /usr/X11R6/bin/x0rfbserver -stealth
# the following line does not return until the admin quits
xrfbviewer ${WKSTN}:0
# thus, the x0rfbserver gets terminated only after the
viewer exits
rsh ${WKSTN} killall -9 x0rfbserver
Not sure about the details yet. Might not need the .
Maybe better to just get the PID of the first rsh call and
kill it off from the server instead of starting another rsh
connection...
Sound good? Is this what other people are also imagining?
Jason
___
Don't miss the 2002 Sprint PCS Application Developer's Conference
August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm
_
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Citrix Server
Venkat- I think the idea is to avoid having to have the user do anything. We also don't want to force the admin into having to use VNC in order to get this feature. x0rfbserver is compatible with VNC, but it doesn't seem to require it. If we can get away with just adding this capability to a simple X session without altogether changing the way LTSP currently works, that would be the ideal solution. Jason Date: Tue, 4 Jun 2002 18:11:30 -0400 From: Venkat Manakkal [EMAIL PROTECTED] Subject: Re: [Ltsp-discuss] Citrix Server Hi, I think the problem could be solved by having the user who has a problem run vncserver with shared mode. Two vncviewer clients (one by the user with the problem and the other by the SA) can connect to the same desktop and shadow each other. ---Venkat. ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Printing
Daniel,
Absolutely. You can do it by user or workstation. I've
done this as follows...
First, you have to consider from which apps you will want
the default printing to work. Most apps acknowledge the
$PRINTER environment variable. Some, like StarOffice, have
their own printer system on top of the OS's printer system.
So, just realize that this might not be your whole
solution.
Now, all you need to do it get the PRINTER variable set
when the user logs in. Many different scripts are run when
a user logs in. Take your pick. I prefer to control
things at the level of the shell, so (assuming you are
using Bash) that would imply using .bashrc, .bash_profile,
or .profile in each user's home directory (man bash if you
don't know the difference between these files). Or, if you
don't want to have to deal with all those files, you could
just do it in /etc/profile or as an added file in
/etc/profile.d/ (if this directory exists in your
distribution). Actually, the file in /etc/profile.d/ is
probably the best bet. It won't get overwritten if you
upgrade your distribution and it's just a single file. You
can call it whatever you want, but I'd probably call it
default_printer.sh.
Now, all you have to do it put a cute little script in
there to do what you want. It might look something like
this (off the top of my head, so please test it out in a
safe place first):
#!/bin/bash
# default_printer.sh
#
# Sets PRINTER to an appropriate value based on the
workstation name.
#
# file containing assignment of workstations to printers
DEFPRNS=/etc/default_printers
# root is exempt from this policy
if [ $USER = root ]; then
exit 0
fi
if [ ! -r $DEFPRNS ]; then
echo $DEFPRNS is unreadable or does not exist!
exit 1
fi
echo -n setting default printer...
WHERE=`echo $DISPLAY | awk -F: '{print $1}' | awk -F.
'{print $1}'`
PRINTER=`cat $DEFPRNS | grep -v ^# | grep ${WHERE} | head
-1 | awk -F: '{print $1}'`
echo $PRINTER
# end of default_printer.sh
Now you just need to make the file /etc/default_printers.
The script makes use of the following file format:
Each important line looks like this:
printername:delimworkstation1delimworkstation2delimetc...
Lines that begin with a '#' are ignored, so you can safely
put comments in the file. Blank lines are also ignored.
The call to 'head -1' makes it safe against accidentally
listing a workstation on more than one line. You can make
the delimeter whatever you want. The only weakness here is
if you had a workstation whose name is a subset of another
workstation's name (like ws1 and ws10). You could
institute a more rigorous format on your printer
assignments file and avoid this, however. For instance, if
you choose your delimeter to be a comma and mandate that
the delimeter must appear before and after each entry, then
you can modify the grep command to look like this:
grep ,${WHERE},
and you would then be safe.
Once these steps are done, you can build on $PRINTER for
special apps like StarOffice... They would require more
scripting to personlize their printer setups, which should
be stored in some sort of text file somewhere. I've done
it for StarOffice, but I don't have the scripts handy right
now.
If you want to do it on username, just replace $WHERE with
$USER and put usernames in the /etc/default_printers file.
Jason
From: Daniel =?iso-8859-2?Q?=A3a=B6?=
[EMAIL PROTECTED]
Date: 05 Jun 2002 14:51:06 +0200
Subject: [Ltsp-discuss] Printing
Hi
I would like to setup my printing system with LTSP, but I
want to select
default printer for workstation, not for user because
many users will
login to server with the same user account from different
thin clients.
Is it possible ? Where should I start to search any
HOWTOs and
documentations?
Daniel Las
___
Don't miss the 2002 Sprint PCS Application Developer's Conference
August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm
_
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Network Neighbourhood
Mandar- I also recommend LinNeighborhood. I've been using with SuSE and there is a big difference between version 0.6.2-48 (SuSE 7.2) and version 0.6.3-54 (SuSE 7.3). I don't know which version comes with 8.0, but version 0.6.4 is available at their main site: http://www.bnro.de/~schmidjo/download/index.html I recommend at least version 0.6.3. Haven't tried 0.6.4 yet... Jason Date: Tue, 04 Jun 2002 19:23:09 +0530 From: Mandar Deodhar [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: [Ltsp-discuss] Network Neighbourhood I am running LTSP thin clients in conjuction with my windows network. Now my linux users want to access some shares on windows network. Basically they want an interface like the Network Neighbourhood in windows. Can anybody suggest an app which works fine in the LTSP environment, so that when the user mounts the shared directory on the windows machine it gets mounted in his own home directory. Also they want to share their directories in linux with the windows machine. Pl suggest any utility Mandar Deodhar ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: Problems with XF86_Mach64
Felix- I've seen this warning message before on an LTSP server which was running the XFree86 3.3.6 Mach64 server. Whenever switching back and forth between the console and the X server via (Ctrl-)Alt-F? this warning would be spit out at the console. Ignore it. The X server was running fine when it was doing this. I think this is a red herring (a distraction from the true solution). Sounds like you just need to tweak your X configuration. Make sure your monitor supports the default refresh/sync rates. When you say you've tried the troubleshooting steps, does that mean you've tried booting into runlevel 3 and running /tmp/start_ws from the workstation console? If so, what were the error messages reported by the X server? This is the direction you need to go to solve the problem. Jason Date: Wed, 29 May 2002 15:55:56 +0100 From: Felix Laate [EMAIL PROTECTED] I have problems with the XF86_Mach64 server (I think). As the client boots it gets past the initial steps, the kernel comes over, the Mach64 module is activated, but then the screen goes black! I tried to do all the stuff mentioned in the LTSP-troubleshooting guide, but to no awail- I tried Ctrl-ALt-F1 and this is what the thing says: mach64ProgramClkMach64CT: Warning: Q 10.6667 Could anybody help me (please) Felix Laate ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Citrix Server
Frank- If you can add what is being referred to as shadowing to LTSP, then PLEASE DO IT! Do you realize that this is probably the #1 feature currently lacking in LTSP?? This is *the* Citrix/WinTSE killer feature!!! I've heard of products/projects (e.g. x2x) that attempt this sort of thing and I've done some initial research into the problem, but it didn't lead anywhere. I've never read anything on x0rfbserver, however. Can you please throw some of the links and info that you might have gathered our way? Thanks, Jason Date: Thu, 30 May 2002 21:22:31 -0300 From: Francisco Castro [EMAIL PROTECTED] On Thu, 30 May 2002 13:04:03 -0700 John Holbrook [EMAIL PROTECTED] wrote: Why? Here's a few reasons: 1) The ability to remotely 'shadow' a users session? This can be done with x0rfbserver, a VNC derivative that works like the windows vnc server shadowing a current session. There are two linux distrubutions for thin clients that are already using it for desktop shadowing (diet-pc and etcterm), and i'm considering to add it to Netstation Linux too. It will be easy to add it to LTSP... ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: Ghost in the Machine
Marcus- As for your vague memory about a limited number of logons, could you be referring to the default configuration of xfs to only spawn 10 daemons? Jim says you can just up the limit in the config file. As for the file-max, there is another parameter you might want up bump up, inode-max! I also toyed with KDE 3.0 when it was first released (haven't tried 3.0.1 yet) and had problems with dcopserver. DCOP is the IPC daemon for KDE apps. Sorry, but I don't have a quick fix. I just switched back to 2.2.2. I think I'm waiting for 3.3.3. :-) Jason Date: Fri, 31 May 2002 09:02:03 -0700 From: Marcus Smith [EMAIL PROTECTED] I've been re-working an LTSP server from the ground up for a friend that I initially set up long ago. I'm running into two problems that I once was able to solve with help from this list, yet I can't find the messages in the list archive. Specifically, I have RedHat 7.3 w/KDE 3.0 running on a dual-Athlon system with 3.5 gigs of RAM and LTSP 3.0.4. My first problem is a vague memory about RedHat systems only being able to export a small number of X-sessions by default. I remember that after adding ten users or so the system would begin to crash all the time. Does this ring a bell with anyone? I already set the file-max option to 65535 in the /etc/rc.d/rc.local file. I thought this had more to do with some xconfig stuff. My second problem is that late last night after getting the system re-installed, I successfully had one single client log in to the system. When I walked back to the server I couldn't launch any programs from the server and I would get a rush of errors saying I had problems with the dcopserver. The last time I received the dcopserver errors I had over a dozen people using the system at any given time. This time it happened after just one login on a remote client. When I logged into the client, I didn't even run any applications. I just logged in to see if I could and then logged out. Thanks for any help. I dig the expertise on this list. ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: Ghost in the Machine
Marcus- Well, it's not xfs, then. IIRC, the inode setting is hard-coded into the kernel (at least it was in the 2.2 kernels). You may need to edit some limits.h file or something like that and recompile. It seems that at some point this requirement was changed or made accessible through /proc, but I'm foggy on the details... It's been a while and I only had to do one of those recompiles once. Agreed, it is confusing. Can you post more info? Perhaps you will need to implement some sort of SNMP monitoring on your stations to get a handle on what is happening there. Jason On Sun, 2 Jun 2002 10:14:57 -0700 Marcus Smith [EMAIL PROTECTED] wrote: I forgot about the XFS bit. But the funny thing is that I'm not even using XFS as yet, or at least I have the lts.conf file parameter USE XFS = N. But I'm glad you reminded me because I'll probably use it again. As for the inode-max, I've been told that RedHat doesn't use the inode-max parameter. In the /proc/dev/fs directory there is a file-max option that I bump up to 65535 using an echo statement in the startup script /etc/rc.d/rc.local. Still all very confusing. ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Announcing Mouse Autodetection
This message is to announce an initial release of a simple add-on module to LTSP which enables mouse autodetection, ala X_MOUSE_PROTOCOL = auto X_MOUSE_DEVICE = auto X_MOUSE_BUTTONS = auto This is accomplished by installed SuSE's hardware detection library (libhd) into the LTSP tree and providing some modifications to the rc.setupx scripts. The website for the module is http://www.uniqsys.com/~jbechtel/automouse/. The module is very rough at the moment and has not been tested outside of my virtual (VMWare) LTSP environment. If you are interested in testing this module or in building upon it (libhd can detect all kinds of hardware!), then please try it out and send me feedback, either personally, or (if appropriate) through one of the ltsp mailing lists. Thanks, Jason ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] DHCP remote root vulnerability
LTSP admins, Apparently, this vulnerability has been out for a couple weeks already. I just received an email from SuSE about updated packages. Anyhow, all recent LTSP installations involve DHCP services, so make sure you're all patched! http://www.cert.org/advisories/CA-2002-12.html Jason ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] NFS mount problems after changing networks
Date: Mon, 20 May 2002 16:53:38 -0700 From: Jesse Scott [EMAIL PROTECTED] Subject: NFS mount problems after changing networks The workstation machine boots, gets it's DHCP address (10.66.6.24), downloads the kernel with TFTP, does the second DHCP request successfully, and then the NFS mount fails. I already went through and changed the IP address and masks everywhere I could find: /etc/dhcpd.conf, /etc/exports, /etc/hosts.allow, /opt/ltsp/i386/etc/lts.conf, but it still refuses to mount: Running dhclient eth0: Setting Rx mode to 1 addresses. Mounting root filesystem: /opt/ltsp/i386 from: 10.66.3.12 (the correct IP of the server) (null)mount: nfsmount failed: Bad file number NFS: mount program didn't pass remote address! mount: Mounting 10.66.3.12:/opt/ltsp/i386 on /mnt failed: Invalid argument ERROR! Failed to mount the root directory via NFS! Possible reasons include: Jesse, At first, I thought maybe you forgot to re-export your NFS shares after altering /etc/exports, but the output from 'showmount -e' proves that the shares are correctly exported. Here are all the appropriate files I can think of, let me know if you need to see another one: snip This brings up my DHCP issue, the above config warns me about the obvious conflict between the 10.0.0.0/255.0.0.0 and 10.66.6.24/255.255.255.248 subnets, but it seems to work and do what I want DHCP wise (only respond to DHCPREQUEST's for the IP's in the range I've been granted control over). I am a bit confused by your dhcpd.conf file. I'm not sure why you declare subnets the way you do. I'm not a DHCP ueberhacker, though, so maybe it's just over my head. It looks like there is a campus-wide DHCP service for your A-class network (10.*.*.*) and you're trying to avoid stepping on its toes. This is the classical problem with an LTSP server joining a network that already has a DHCP server... There are really two problems: 1) Prevent LTSP clients from listening to the other DHCP server. 2) Prevent normal clients from listening to your LTSP DHCP server. There've been some ideas on the ltsp-developer from Ken Yap which would solve problem #1 by adding an option to the etherboot code that would require the DHCP server to contain a string like DHCP_VCI=Etherboot-5.0 and he also suggests a temporary fix: Here's a lower tech approach which can be deployed right away. If DHCP_REJECT is defined in option 129 as the IP address of a server to ignore, then dhclient will not accept replies from it. So you would set it to the IP address of the rival DHCP server. This works with the dhclient in current LTSP initrds. You seem to have solved problem #2 by using the 'not authoritative' option. Am I understanding your situation correctly? /etc/hosts.allow: bootpd:0.0.0.0 in.tftpd: 10.66.6. portmap: 10.66.6. This got me for a while... My copy of mountd was compiled with tcp_wrappers support! You might need to put a line like this in hosts.allow: mountd: 10.66.6. You might also want to check to make sure that ipchains/iptables (a.k.a. firewalling) isn't blocking the packets. Also, any messages in /var/log/messages when the client fails to mount? Jason ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Tarantella
Tarantella is a platform (whatever meaning that word still retains) for delivering application interfaces to a remote client via Java and a proprietary protocol. I use the word platform because it is pretty much OS agnostic at both ends. The server part runs on multiple OSes, (Linux, for our purposes). The server will start a local process, capture the display, and forward it through its proprietary (and very bandwidth-light protocol). The server also will create connections to and pass on application interfaces coming from other boxes (Windows Terminal Server and/or X11 redirected). The client part is usually a Java program running inside of a web browser (IE or Netscape). It is not limited to the confines of the original browser window, however. The server can specify to the client whether the application's display should be opened inside of the Tarantella Desktop (in the current window) or in an external window. The server also determines the size of the window. Of course, the user can run multiple apps simultaneously as with other desktops. The coolest thing is that, like VNC, it is *stateful*. So, if a user disconnects, the applications that are running can remain running (also configurable per application) and will still be there when the user reconnects later. The control is kept at the server running Tarantella. It's all run through a web interface, even the admin. There's no command-line interface. If you want to secure the login screen, you have to use Apache with some sort of SSL enabled. If you want to secure the communications stream for the session (starting after login), then you can buy their SSL module. Here's an example scenario to give you an idea how it looks to the user: 1) User connects to the internet (or is already on the LAN) 2) Fires up browser and goes to https://tta.whereiwork.com/ 3) All kinds of crap gets downloaded to the browser and eventually a login appears. If it doesn't, then some sort of Java support is lacking in the browser. They may need to update the latest IE/Netscape. 4) The user logs in and their Tarantella Desktop appears. 5) The desktop has a column of buttons on the left that launch applications either in the pane on the right or in a new window. The user has no control over anything here, really. It's all up the Tarantella admin how things work. 6) User clicks on a button and the application is started on the appropriate server, grabbed by Tarantella, and then passed through to the browser. At this point, that's basically it. The user uses the program as he/she normally would. You can setup Tarantella's passthru printing to allow the user to print from whatever app to a queue on the Tarantella server. The job is then picked up by Tarantella and is forwarded to the browser of the appropriate user, which prints the job wherever the browser is running. This might take some tinkering, but it does work. For the admin, it's fully customizable and hierarchical. There are users and groups of users and groups of groups of users, etc. There are applications and groups of applications and groups of groups of applications, etc. Each level can impose some settings and it all resolves itself for the context of the current user and application. You can therefore specify which users get which apps and how each app runs. Oh, I almost forgot that it has its own tcp_wrappers-like ACL capabilities so it can be setup to allow connections only from trusted sources. It's a sweet setup with a decent built-in help system. And because it's from SCO (now Caldera), it has perfect Wyse terminal emulation! Enjoy! Jason ___ Have big pipes? SourceForge.net is looking for download mirrors. We supply the hardware. You get the recognition. Email Us: [EMAIL PROTECTED] _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Stunnel and encrypted traffic
Security-minded LTSPers, I've been following the discussion on encrypting traffic with interest. I like the ideas that are flowing on the list. Keep it up! One comment I've had for the person who suggested putting SSH keys on the EPROM is that there are NICs which support flashing the EPROM. You could prepare the images of Etherboot + SSH keys on the server and the flash them to the NICs using a DOS image on a floppy for easy (or easier) updating/management. Still, this method is suboptimal. I also like the discussion about just encrypting the login by passing the SSH public key to the station in response to the user ID and then encrypting the password string. This is a great tactic, but it does require modification to the login procedure. It is probably beyond the scope of the LTSP project to start meddling with XDMCP implementations and the various display managers. For my contribution, I just want to mention stunnel. From the main page: Stunnel is a program that allows you to encrypt arbitrary TCP connections inside SSL available on both Unix and Windows. Stunnel can allow you to secure non-SSL aware daemons and protocols (like POP, IMAP, LDAP, etc) by having Stunnel provide the encryption, requiring no changes to the daemon's code. It's GPL and I've used it to do SSL-POP3 with Netscape (which doesn't natively support it). It works! I didn't have to generate keys or anything! If I'm thinking properly, then we should be able to make an Stunnel+OpenSSL add-on package for LTSP which provides a wrapper on the workstation for the X server's XDMCP request and a new service on the server which functions merely as an ecrypted layer over the existing display manager (port 177). This would provide for an encypted login. For the paranoid or those who require higher levels of security, one could then invoke stunnel again for the permanent X connection (ports 6000-6063), right? It seems too easy... too good to be true... http://www.stunnel.org/ http://www.stunnel.org/examples/ Please, someone deeper into this issue look into it and let us know what comes of it... Jason ___ Have big pipes? SourceForge.net is looking for download mirrors. We supply the hardware. You get the recognition. Email Us: [EMAIL PROTECTED] _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] CD Burning
Dennis, I'm not sure if this will help, but when I was using 'verynice' to control runaway processes, I noticed in the sample config file that it also can give certain processes elevated priority and the example was xcdroast. Give it a try and configure it to make xcdroast have a very high priority. Your other processes may suffer a bit of performance, but the likelihood of turning out more coasters should drop. http://www.tam.cornell.edu/~sdh4/verynice/ Jason From: Katsonis, Dennis [EMAIL PROTECTED] Date: Tue, 14 May 2002 09:08:53 +1000 Just one simple question. I'm using LTSP in a home environment with 2 other PC's, just purely for interest and to play around with it. However I've noticed that if someone logs on while I'm burning a CD, the CD burning process will fail (I'm using X-CD-Roast) Is there any way to prevent this, or does anyone know what the cause may be? Oddly, the FIFO buffer did not appear to be empty, and people logging in from a workstation never used to interrupt the burn process when I first started using LTSP. This isn't hugely important as I can stop DHCPD while burning a CD, but if possible, I'd rather know how to prevent it without doing this if possible. ___ Have big pipes? SourceForge.net is looking for download mirrors. We supply the hardware. You get the recognition. Email Us: [EMAIL PROTECTED] _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] stability
win, The problem is that you only have 16MB of memory in your client. I'm not sure how much the 8 telnet sessions would require, but the X session alone is enough to exhaust your 16MB once you open a few windows. The NFS swapping should prevent it from actually *crashing*, but the swapping is probably what you're experiencing when the client pauses for a few seconds. If the client is actually crashing, however, then either the client is far into swap and thrashing, or NFS swap really isn't working even though you have the setting turned on in lts.conf. Can you determine for certain that NFS swapping is actually working properly? Jason Date: Tue, 14 May 2002 09:49:37 +0800 (CST) From: [EMAIL PROTECTED] hi all, i have stability problems with some clients using ltsp3 RedHat 6.2. they just freeze. sometimes after several seconds they work again, then freeze again. even the mouse freezes too. at other times the screen gets garbled and either X restarts or the client needs rebooting. server is PIII 500Mhz 768MBRAM RH6.2., rtl8139 lancards clients P100, 16MB RAM, rtl8139 lancard, Cirrus logic gd4xxx 1mb video. i have 8 txtmode telnet sessions and also X running. NFS_SWAP enabled. looks like if you open many programs the client crashes or hangs. ___ Have big pipes? SourceForge.net is looking for download mirrors. We supply the hardware. You get the recognition. Email Us: [EMAIL PROTECTED] _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] True Type Fonts not installed (?)
From: Philip A. Roa [EMAIL PROTECTED] Date: Tue, 14 May 2002 17:19:31 +0800 I'm currently experimenting True Type Fonts but haven't really made it run. Since TT fonts are already supported on the new XFree86 i tried to do the ff: 1. Copied some .ttf fonts to /opt/ltsp/i386/usr/X11r6/lib/X11/fonts/Truetype 2. did a ttmkfdir fonts.dir and 3. mkfontdir The files 'fonts.dir' and 'fonts.scale' have been created in the same directory 4. Made a hard reset on the remote workstation (and restarted XFS in the server - not sure if this is needed) 5. Tested if the fonts were present using KDE Control Center-LookNFeel-Fonts - Fonts cannot be browsed (am i doing it right?) Actually, i'm not sure how to test if the TT fonts were being served by XFS. How do you do this? 'man -k font' 'xlsfonts' is one tool you're using xfs, so 'fslsfonts' is another tool you could use. (btw, XFS is a filesystem and xfs is the X font server.) But the real key is to look at xfs's configuration file. The default location is /usr/X11R6/lib/X11/fs/config. It should be there, but check how xfs is being called on the command line using 'ps' to make sure it's not being redirected by the -config option. In this file, you should see a section like this: catalogue = /usr/X11R6/lib/X11/fonts/speedo, /usr/X11R6/lib/X11/fonts/misc, /usr/X11R6/lib/X11/fonts/100dpi/ The directory with your TT fonts must be in this list. Make sure there is a comma at the end of every line except the last one in this section! My final objective really is to have TT fonts in OpenOffce 1.0. Not sure where OO gets its fonts. It might not be simply taking them from the X server or XFS... Jason ___ Have big pipes? SourceForge.net is looking for download mirrors. We supply the hardware. You get the recognition. Email Us: [EMAIL PROTECTED] _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] utmp/wtmp/sessreg
LTSP users, admins, and developers,
What about a more expansive, global, and adaptive approach to the
utmp/wtmp/sessreg problem?
See the following messages discussing prior solutions to this problem:
Ken Godee:
http://sourceforge.net/mailarchive/message.php?msg_id=198655
Derek Dresser:
http://sourceforge.net/mailarchive/message.php?msg_id=1166304
Philip A. Roa:
http://sourceforge.net/mailarchive/message.php?msg_id=1299193
And this message that I have here, but that somehow doesn't come up in
the Sourceforge mailing list search:
Date: Fri, 01 Mar 2002 21:01:42 -0500
From: David Johnston [EMAIL PROTECTED]
Subject: Re: [Ltsp-discuss] Knowing exactly who is currently logged-in
I changed two files to fix this. With these changes, if you are using
GDM, `who` will look like this:
# who
davidws05:0 Mar 1 20:18 (ws05)
beth ws02:0 Mar 1 12:08 (ws02)
Here are the changes:
Put this in /etc/X11/gdm/PreSession/Default:
#!/bin/sh
if [ -x /usr/bin/X11/sessreg ]; then
HOST=`echo $DISPLAY | cut -f1 -d: | cut -d\. -f1`
Display=:`echo $DISPLAY | cut -f2 -d:`
/usr/bin/X11/sessreg -a -w /var/log/wtmp -u /var/run/utmp -l
$HOST$Display -h $HOST $USER
fi
Put this in /etc/X11/gdm/PostSession/Default:
#!/bin/sh
if [ -x /usr/bin/X11/sessreg ]; then
HOST=`echo $DISPLAY | cut -f1 -d: | cut -d\. -f1`
Display=:`echo $DISPLAY | cut -f2 -d:`
/usr/bin/X11/sessreg -d -w /var/log/wtmp -u /var/run/utmp -l
$HOST$Display -h $HOST $USER
fi
Finally, there's Ken Godee's WSMON Perl/Tk program with GiveConsole and
TakeConsole modifications http://www.perfect-image.com/wsmon/:
if [ ${DISPLAY} = :0.0 ] ; then
ltsp=$DISPLAY
else
ltsp=${DISPLAY%%.*.*}
fi
All of these solutions depend on various assumptions about the
environment and the LTSP client hostnames. They were all developed on a
specific platform (Red Hat Linux 7.x) and, as seen recently, do not work
with other distros or in situations in which the assumptions made do not
apply (strange FQDN).
From the WSMON readme file:
assume:
8 max characters for usernames
8 max characters for groupnames
full name equals two words ie. John Doe not ie. John Thomas Doe.
Assumes workstations are on local domain only.
Workstations naming schem must start with ws ie ws001
This allows program to report on LTSP workstations only.
If people name their clients strangely, then simply trimming $DISPLAY
down to just the hostname may not be enough. With a 4 character ut_id
field, one must assure that the appropriate set of four characters is
being used, otherwise every client is assigned the same ut_id string
(original problem).
I'm wondering if under the new (still under development) v3.0.1
installation/configuration system, it would be possible to include an
optional procedure to detect the status of utmp/wtmp logging on the
system and to modify it if necessary? How difficult would this be?
I'm thinking it could scan the {Give|Take}Console scripts for the
'sessreg' command and parse for the argument given to the '-l' option
($DISPLAY, probably) and then scan again for how this argument is being
modified (if at all). Then it could scan the /etc/hosts and/or the
lts.conf file (or ask the administrator) for the naming scheme of the
workstations. It could then determine the domain of the network and
from these determine the appropriate set of 4 characters to use as the
argument to the '-l' option of sessreg.
It could then modify the {Give|Take}Console scripts to perform the
appropriate modifications on $DISPLAY, using the values discovered above.
This approach should be general enough to avoid distro-specific problems
(e.g.: Debian) and should allow admins who don't want to muck around
with their login logging system to install LTSP and have all of their
existing tools (w, who, last, Gkrellm, etc.) *just work*. Admittedly,
it's probably not a *huge* selling point for LTSP, but a rough edge that
it would be nice to round out.
I could probably write this in Bash and awk or sed, but currently all
the install/config scripts are in Perl. I guess it could also just be
an add-on package under contribs, thereby allowing someone to write it
in whatever language they want... I really wanted it to become a core
part of the LTSP install. It seems like several people have already put
in a lot of time to get a grip on this problem and that it might be
ready for prime-time.
Thoughts?
Anyone interested? Or did I just volunteer myself? :-)
Jason
___
Have big pipes? SourceForge.net is looking for download mirrors. We supply
the hardware. You get the recognition. Email Us: [EMAIL PROTECTED]
_
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For
Re: [Ltsp-discuss] RE: Question on old 2.09 pre and XF86 resolutionsettings
From: Charles Marcus [EMAIL PROTECTED] OK, I've been through the older docs (forgot they were all still up on the downloads page), and I tried changing the lts.conf file for the individual workstations, but the changes are not doing anything. I've also tried changing both XF86Config-4 files (XF86Config-4 is the one referenced in lts.conf) - there is one in /etc/X11 and one in /tftpboot/lts/ltsroot/etc (which one is used??), but the changes are not affecting anything - hitting ctrl-alt-+ (or ctrl-alt--) doesn't do anything - and yes I did restart the X-server every time. Restarting the X-server on the workstation is not enough. You must reboot the workstation entirely so that the lts.conf file is read again and the startup scripts prepare the configuration files correctly and lauch the correct X server. If you meant restarting the X-server on the LTSP server, then that is irrelevant. Someone else already pointed out that the files affecting the workstation are located under /tftpboot/lts/ltsroot/etc. I know that XFree 4 isn't supposed to be as complicated as 3, but does it still require all of those modeline entries? I found an older XF86Config file (no -4 on it), that had all of the modelines in it, but there are _none_ in either of the XF86Config-4 files. It doesn't matter so much what is sitting in the files in /tftpboot/lts/ltsroot/etc as what the startup scripts *do* with them. Look at /tftpboot/lts/ltsroot/etc/rc.setupx and /tftpboot/lts/ltsroot/etc/rc.local. These are where the magic happens. These are the scripts that read your lts.conf file specifications and translate them into configuration files on the workstation (in RAM!). Look to rc.setupx for the XFree86 setup parameters and how it handles the options pertaining to resolution. Pay special attention to quoting. Note that unless you are specifying the XF86Config file to use explicitly through the lts.conf option XF86CONFIG_FILE, then any file you put in /tftpboot/lts/ltsroot/etc will just be ignored and a default one will be built according to the other lts.conf options and defaults. Also, an important part of debugging the problem will be seeing the resulting XF86Config* files created on the workstation in the RAM-only /tmp filesystem. To do this, you must be sitting at the workstation and hit (Ctrl-)Alt-F1 to get to the console. Then you can do a 'cat /tmp/XF86Config | less' to read the file. Make sure it is handling your lts.conf options correctly. Another possibility for the workstations not obeying workstation-specific lts.conf entries is that the workstations do not know their own name, so they cannot identify their section of the lts.conf file. To fix this you need to add the 'use-host-decl-names on;' option in the appropriate context in dhcpd.conf (and restart dhcpd). Jason ___ Have big pipes? SourceForge.net is looking for download mirrors. We supply the hardware. You get the recognition. Email Us: [EMAIL PROTECTED] _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] VMWare test environment setup
[I'm posting this message back through to the list as it contains some useful info for others using VMware virtual ltsp clients. Those interested, please be sure to read the quoted message below as well.] Jim, Got it. The instructions are at VMWare's site: http://www.vmware.com/support/reference/common/guest_linux_xf864.html At the very bottom of the page: As it starts, X prints (WW) VMWARE: No Matching Device Section for instance (BusID PCI:0:15:0) found. and then (EE) No devices detected. This happens because the VMware SVGA adapter is detected as a secondary display controller (secondary to a separate ISA VGA card). You need to be sure to include the BusID line in the Device section when you modify XF86Config. See sample section in Installing the Precompiled Object File above. This is exactly the problem I had. I looked in rc.setupx and found the X4_BUSID ltsp variable and used it and it worked! :-) Jason Jason Bechtel wrote: Jim, Unbelievable... I thought I'd checked *everything*. Turns out I would probably have been fine with VMWare's host-only networking if I'd looked one step further to see that my version of rpc.mountd was compiled with tcp_wrappers support! Argh! All I had to do is add a line to hosts.allow for 'mountd:'... sigh ...live and learn. Well, now it's motivated me to turn on and customize ipchains (a.k.a. SuSEfirewall). So, I'm somewhat more secure than I would've been without going through all this trouble. So, I'm just doing bridged networking on the main eth0 interface afterall. Thanks for all your advice with the dummy0, though. I did give it a shot. Now X4 is complaining about 'No devices detected.' I haven't looked into it yet, but if you think I'll hit a dead end and you've been down this road before, please drop me any hints you might have. Thanks, Jason ___ Have big pipes? SourceForge.net is looking for download mirrors. We supply the hardware. You get the recognition. Email Us: [EMAIL PROTECTED] _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] light windowmamgers
Or, if you want something more customizable or if you don't have the Gnome/gtk+ libs installed, I wrote some simple Tcl/Tk scripts to do basically what xalf does (very nice, btw.) before I knew about xalf. http://uniqsys.com/~jbechtel/tk_start.tgz Admittedly more crude, but another option. Options are good. :-) Jason *** Try xalf (you can get it at www.freshmeat.net). Then launch your program like this: xalf -s mozilla or xalf -s xcalc The above will launch Mozilla (the second one will launch xcalc) but xalf will display a splash screen while it's loading. _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] LTSP on new-world macs
Fredrik, See the main website of LTSP for a link to a solution by Skip Gaede: March 28, 2002 - PPC Macs as LTSP Clients http://www.ltsp.org/ http://www.ltsp.org/contrib/index.php Jason Date: Fri, 26 Apr 2002 10:40:52 +0200 From: [EMAIL PROTECTED] We're investigating the possibility to use our iMacDV:s as thin clients connected to a x86 terminal-server running Mandrake. I've read through the mailing-lists and documentation but haven't found any information. The LTSP for nubus-macs are, as far as I know, not applicable to iMacs as they are pci-based. The iMacs are good as thin clients as they have a boot-rom built in. Have anyone tried this? Is it possible? Which parts of the client-software nneds to be converted to ppc? I already have a working installation of Mandrake 8.2ppc on the iMacs -- can I use the kernel and/or other parts of this installation? Regards, Fredrik Hegardt _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] VMWare test environment setup
Jim (and list),
I know that you've successfully set up a test environment for LTSP using
VMWare. I followed your instructions from prior posts to the lists, but
I've hit a wall. Hopefully you (or someone else on the list) can
help... (Please pardon the length of this mail.)
I'm using LTSP 3.0 from RPMs:
ltsp_core-3.0.0-1.i386.rpm
ltsp_kernel-3.0.1-1.i386.rpm
etc.
I've downloaded the lancepci and the pcnetfastiii bootable floppy images
from rom-o-matic.net and created a VM in VMWare that uses one of those
files as it's floppy 'drive' (both seem to work).
I've setup host-only networking in VMWare on the unused private subnet
192.168.64.0. My host OS (SuSE 7.2 w/ lots of updates) is .64.1 and the
client is .64.2.
I already figured out one trick regarding the ISC DHCPD included with
VMWare (2.0.4 build-1142). When it starts up it delivers this message:
Configured subnet: 192.168.64.0
Setting vmnet-dhcp IP address: 192.168.64.0
Opened: /dev/vmnet1
Recving on VNet/vmnet1/192.168.64.0
Sending on VNet/vmnet1/192.168.64.0
And the client tries to TFTP from 192.168.64.0! Strange, but VMWare
seems to trick the client into communicating with 192.168.64.0 for
DHCP... So, I had to use the next-server option to tell it where the
server *actually* is:
next-server 192.168.64.1;
So, now the kernel is downloaded and lots of 's follow as the initrd
is loaded and the boot process continues... Now the client tries (and
gets stuck at) the NFS mount:
Running dhclient
Mounting root filesystem: /opt/ltsp/i386 from: 192.168.64.1
(null)mount: nfsmount failed: Bad file number
NFS: mount program didn't pass remote address!
mount: Mounting 192.168.64.1:/opt/ltsp/i386 on /mnt failed: Invalid argument
ERROR! Failed to mount the root directory via NFS!
Possible reasons include:
1) NFS services may not be running on the server
2) Workstation IP does not map to a hostname, either
in /etc/hosts, or in DNS
3) Wrong address for NFS server in DHCP config file
4) Wrong pathmname for root directory in the DHCP config file
Kernel panic: Attempted to kill init!
Knocking these off one at a time...
NFS svcs. are running:
# exportfs
/usr/opt/ltsp/swapfiles
192.168.64.0/255.255.255.0
/usr/opt/ltsp/i386
192.168.64.0/255.255.255.0
# ps -ef | grep nfs
root 482 1 0 Mar25 ?00:00:00 [nfsd]
root 483 1 0 Mar25 ?00:00:00 [nfsd]
root 484 1 0 Mar25 ?00:00:00 [nfsd]
root 485 1 0 Mar25 ?00:00:00 [nfsd]
root 488 1 0 Mar25 ?00:00:00 /usr/sbin/rpc.mountd
--no-nfs-version 3
# ps -ef | grep portmap
bin321 1 0 Mar25 ?00:00:00 /sbin/portmap
# netstat -l | grep nfs
udp0 0 *:nfs *:*
The workstation IP maps to a hostname in /etc/hosts:
192.168.64.1ltsserv
192.168.64.2virtual
Just in case this wasn't enough, I set up a BIND9 server for my private
domain (ltsnet) to make sure, so DNS is also there and it still doesn't
work:
# nslookup -sil 192.168.64.2 192.168.64.1
Server: 192.168.64.1
Address:192.168.64.1#53
2.64.168.192.in-addr.arpa name = virtual.ltsnet.
# nslookup -sil 192.168.64.1 192.168.64.1
Server: 192.168.64.1
Address:192.168.64.1#53
1.64.168.192.in-addr.arpa name = ltsserv.ltsnet.
Bind is configured to listen on both localhost and the private interface:
listen-on { 192.168.64.1; 127.0.0.1; };
To eliminate possibilities 3 4, here's my
/etc/vmware/vmnet1/dhcpd/dhcpd.conf file:
# We set domain-name-servers to make some clients happy
# (dhclient as configued in SuSE, TurboLinux, etc.).
# We also supply a domain name to make pump (Red Hat 6.x) happy.
#
allow unknown-clients;
default-lease-time 1800;# 30 minutes
max-lease-time 7200;# 2 hours
subnet 192.168.64.0 netmask 255.255.255.0 {
#range 192.168.64.128 192.168.64.254;
option broadcast-address 192.168.64.255;
option domain-name-servers 192.168.64.1;
#option netbios-name-servers 192.168.64.1;
option log-servers 192.168.64.1;
#option routers 192.168.64.1;
option domain-name ltsnet;
option subnet-mask 255.255.255.0;
group {
filename /lts/vmlinuz-2.4.9-ltsp-5;
#filename /lts/vmlinuz-2.4.9-ltsp-lpp-5;
use-host-decl-names on;
next-server 192.168.64.1;
option root-path 192.168.64.1:/opt/ltsp/i386;
host virtual {
#hardware ethernet 00:50:56:A4:10:40;
hardware ethernet 00:50:56:CA:56:C3;
fixed-address 192.168.64.2;
}
}
}
So, is there something wrong with 'mount', either in the LTSP binaries
or in the SuSE binaries?
TIA for any
Re: [Ltsp-discuss] True thin clients in biz environment?
Tommy, It sounds to me like you consider true thin clients to only be expensive proprietary hardware solutions. I think that this is a mistake. For the same reason that your proprietary WinTerm boxes are now useless under a new operating environment, your Neoware boxes will soon be unable to adapt to your ever-changing network. What's wrong with PCs? Nothing says that PCs have to boot from a slow, clunky, failure-prone floppy. The only thing you need are some network-bootable NICs in some vanilla PCs and there's the cheapest, most versatile, most upgradeable *true* thin client you can find. Just because it doesn't have the chic form factor, doesn't make it any fatter on cost or performance... The open PC architecture is your friend. Every component is replaceable. Video not the best? Buy 40+ video cards at $40 each instead of 40+ new thin clients at $400 each. Just like proprietary software, proprietary hardware can only limit you and lock you into a solution. Jason PS: Please, no HTML email to the list. From: Tommy Eaton [EMAIL PROTECTED] Date: Fri, 12 Apr 2002 07:37:55 -0400 Subject: [Ltsp-discuss] True thin clients in biz environment? Hi, I have recently decided to replace our current 40+ user network with LTSP. We are using Citrix, Exchange, and Winnt 4.0 TSE. However, it order to accomplish this, I need to find a thin client that I can make boot to the LTSP, otherwise we are stuck in M$ land. Most of our existing thin clients are old WYSE WinTerms that I hope to replace with a new product line. I have been looking at Neoware's (www.neoware.com http://www.neoware.com/ ) Capio 508, but have been unable to get it to go. My question is this: Are any of you currently in an organization that is utilizing LTSP? If so, what are you using for thin clients? How are you accomplishing your setup? Using PCs that boot to the floppy is not an option. We must purchase the true diskless (yet still have the option to add drives at will) thin client hardware. Thanks in advance. _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Writing to named piped over NFS
Date: Wed, 6 Mar 2002 12:57:35 -0600 (CST) From: Lachlan Dunlop [EMAIL PROTECTED] Before moving my application to ltsp thin clients, it has been working great! With the touchscreen connected to the serial port on the server (but it is very limiting). So I have connected my touch screen to the com port on the thin client and it collects the data. It then writes each screen touch to a named pipe which is located in the /tmp/swapfiles directory. The application never reports an error, but there is no data in my pipe. I am running ltsp 3.0 on SuSE 7.0. I have swap turned on for the client. I can create files in the swap area no problem. Is there a different solution for this. Should I throw the named pipe stuff out when running over NFS? This is a fundamental process that ltsp would be awesome for if I could get the data across the network in realtime. Lach, My understanding of named pipes is that the part that makes them work resides in the kernel. A named pipe only works for communication between two apps running on the same running copy of a kernel! Each app opens the named pipe, which actually accesses a resource in the kernel. So, named pipes functioning over NFS doesn't make any sense. I don't think it could ever work. As for your touch-screen app... If it could be made to run on the server, but to display on the client (on the touch-screen), then it could write to a named pipe *on the server* and thereby communicate with whatever other process it needs to. If the app needs to run on the client in order to read the screen touches, then perhaps it could somehow (UDP? TCP?) send the screen touch info back to a process running on the server, which would then write the data to a named pipe *on the server*. Jason _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] NFS security issues
LTSPers, Following up the discussion on rw NFS and security issues... If you must use remotely-mounted/centrally-located directories in a hostile environment, might I recommend AFS instead of NFS. AFS allows for user-level authentication using Kerberos. It was formerly proprietary, but after IBM bought the company mainly developing it, they open-sourced it. See the following links for various projects implementing AFS: http://openafs.org/ http://www.stacken.kth.se/projekt/arla/ http://oss.software.ibm.com/developerworks/opensource/afs/ You can also *buy* AFS. Also, note that AFS also has a Windows(TM) port, though I've heard that it is not yet totally stable. Jason _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] [ OT ] KNOPPIX - Live Linux Filesystem On CD
If you can read German http://wwwknoppernet/knoppix/ Jason _ Ltsp-discuss mailing list To un-subscribe, or change prefs, goto: https://listssourceforgenet/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on ircopenprojectsnet
[Ltsp-discuss] Re: Windows DHCP?
From: Martin [EMAIL PROTECTED] Date: Wed, 20 Feb 2002 10:28:31 - Made the change as suggested... it now goes to the correct server... BUT despite only specifying the Root Path as below, the client tries to mount /opt/ltsp/i386000. I've tried putting it in quotes and various other things, but it always appends 000 to the end! Has anyone else seen anything similar? Not that I want you to have to continue living with that MS DHCP server, but... What about `ln -s /opt/ltsp/i386 /opt/ltsp/i386000`? Will NFS export a symlink? If not, would a hard link work where a soft link wouldn't? And, if all of your clients are going to be served this way and no form of symlinking will work, just move (rename) the whole directory to /opt/ltsp/i386000! Then, for the sake of any scripts or future LTSP upgrades, create the link in the other direction so that the original directory exists and points to the new one. Jason _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: Need config help please - 1 card server, separate dhcpserver
From: Dyg It [EMAIL PROTECTED] Date: Wed, 20 Feb 2002 09:15:54 -0600 My network uses 192.168.1 addresses instead of the default LTS default (192.168.0) with a subnet mask of 255.255.255.0. I have a linksys cable/dsl router which also acts as my hub and dhcp server for my network. Its address is 192.168.1.1. I have a machine with ONE ethernet card that I want to run as a Linux Terminal Server (@ 192.168.1.254). I don't want to add a second ethernet card or move dhcp to the LTS - it won't be running all the time. I need to keep the linksys as my dhcp server. There's no reason you can't run multiple DHCP servers on your network. You should be able to restrict the LinkSys to a subset of your C-class (for dynamic addresses). Then set up DHCP on your Linux server and have it handle the MAC addresses for the NICs in your LTSP workstations in another part of the C-class (static addresses). Have you already tried this? If so, did the workstations reliably choose to take the response from the DHCP server which had a specific entry for their MAC address? I'm not sure if this fits in with your plans (needing to keep the router as the DHCP server), but Jim is right that the DHCP server in that router is probably not capable enough to handle LTSP workstations. Jason _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] X Font Server - bad news?
Date: Tue, 12 Feb 2002 11:41:10 -0800 From: Marcus Smith [EMAIL PROTECTED] Subject: [Ltsp-discuss] X Font Server - bad news? So... for me the Big Question is: How do I go about getting pretty, anti-aliased fonts without XFS? Apparently, one could configure XFS to handle more clients as Jim said, but he makes a good point about the associated risk of all clients freezing(!). And, as far as I know, I don't think there's any difference between serving fonts out of XFS and X getting them directly from the filesystem (whether that fs is local or NFS mounted). The configuration is basically the same. Just as XFS is configured (for RH and SuSE, the file is /etc/X11/fs/config) to contain a list of paths which it searches consecutively, like this catalogue = /usr/X11R6/lib/X11/fonts/misc:unscaled, /usr/X11R6/lib/X11/fonts/75dpi:unscaled, /usr/X11R6/lib/X11/fonts/100dpi:unscaled, /usr/X11R6/lib/X11/fonts/Type1, /usr/X11R6/lib/X11/fonts/URW, /usr/X11R6/lib/X11/fonts/Speedo, /usr/X11R6/lib/X11/fonts/CID, /usr/X11R6/lib/X11/fonts/misc, /usr/X11R6/lib/X11/fonts/75dpi, /usr/X11R6/lib/X11/fonts/100dpi, X itself can be likewise configured like this Section Files FontPath /usr/X11R6/lib/X11/fonts/75dpi:unscaled FontPath /usr/X11R6/lib/X11/fonts/local FontPath /usr/X11R6/lib/X11/fonts/misc:unscaled FontPath /usr/X11R6/lib/X11/fonts/75dpi:unscaled FontPath /usr/X11R6/lib/X11/fonts/100dpi:unscaled FontPath /usr/X11R6/lib/X11/fonts/Type1 FontPath /usr/X11R6/lib/X11/fonts/URW FontPath /usr/X11R6/lib/X11/fonts/Speedo FontPath /usr/X11R6/lib/X11/fonts/75dpi FontPath /usr/X11R6/lib/X11/fonts/100dpi ... EndSection This is taken from X4, but the same applied to versions 3.3.x. I'm not sure where the anti-aliasing comes in, but I was under the impression that it was a feature of the X server itself and not of XFS. Please, someone correct me if I'm wrong about that. One of the things that I learned when working with X3.3.6 is that changing the order of the paths can make a big difference in readability. If netscape displays with teeny tiny fonts, you can increase the font size in each user's netscape prefs (bring) or you can put the ':unscaled' fonts first and put 100dpi before 75dpi both in the unscaled lines and afterwards. This is a good first step. The best thing to do is to install some TrueType fonts, however. These look nice and it's not hard to setup. You can use xfstt or some other dedicated TT font server, but I believe xfs and X support them natively nowadays. As always, here's the relevant HOWTO: http://www.linuxdoc.org/HOWTO/mini/FDU/index.html. Oooh! I just looked and it seems to have been updated sometime this month. There is a nice big section on X4.x and Anti-aliasing, too! It seems that KDE 2.x supports it because Qt 2.3.x supports it. Gnome 2.x will also support it if GTK 2.0 supports it. But individual apps must also be anti-aliasing aware through the toolkits they are built against. So, even if you use X4.x and your X driver supports the X-RENDER extension and you use KDE 2.2.2, still not all of your apps may be using those anti-aliasing features. They link to this HOWTO at Trolltech for getting anti-aliasing with Qt (KDE): http://trolls.troll.no/~lars/fonts/qt-fonts-HOWTO.html. So, in my opinion, ditch XFS even though you might've been able to get away with it. :-) Jason _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] printing terminal screen
Julius, I am glad I could share something helpful for you. A couple things... 1) I'm sure it was just a typo, but the main program is xwd, not xdm (just so we're all straight on that). It sounded from your text like you were using the right app, though. 2) I was recalling that script off the top of my head, so 'gs' might really be in there, but I didn't include it in my recommendation. It definitely makes sense, however, if you do not have a PostScript printer and your lpr filters do not call 'gs'. However, this may solve your gs in a pipe problem. Let lpr use gs and leave it out of the script itself. 3) I'm sorry I forgot to mention this up front, but I did not come up with this solution myself. Someone at Unique Systems http://www.uniqsys.com showed it to me and I don't remember who it was and I don't know if they figured it out or dug it up from USENET... Anyhow, glad it helped! Jason On Sun, 10 Feb 2002 09:28:24 -0500 (EST) Julius Szelagiewicz [EMAIL PROTECTED] wrote: Jason, thank you. your suggestion to use xdm | convert | gs | lpr is pure beauty of a simple and strigthforward solution. xdm is pretty flexible, convert wil convert anything and if i ever get ghostscript to work in a pipe i'll have it all, well mabe not all but most. thank you _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: [Ltsp] printing terminal screen
Date: Fri, 8 Feb 2002 10:00:00 -0500 (EST)
From: Julius Szelagiewicz [EMAIL PROTECTED]
Subject: [Ltsp-discuss] Re: [Ltsp] printing terminal
screen
the users that i moved from dumb terminals to ltsp boxes
instead
of being extatic are grumbling that they can not print
gnome terminal
screens. they actually have a need to do it and they do
access the main
app from terminal emulation program. any advice? tia,
julius
Jim's solution sounds good and I would do that if it's
possible. If not though, I have another solution which is
more general. Create a script using a few shell commands
and make a button on the desktop that executes the script.
Then put a printer or similar icon on it and tell your users
that it's the print window tool. When they click the
button, the cursor changes and then they click on the window
to print. The cursor changes back to normal, they hear two
beeps and the print job is submitted. The script might look
something like this:
xwd | convert some options I can't recall | lpr
-P${PRINTER}
See the man page for convert. On Red Hat it's part of the
ImageMagick package. Then you'll have to define each users'
PRINTER variable in their .profile (or equivalent) file
using their $DISPLAY or $USER variable and whatever method
you choose to map users/workstations to printers. You would
have to do this for Jim's solution as well.
Now you have a print window tool that works for all
programs! The only downside is that the 'xwd' command dumps
the *entire* window, not just the terminal contents.
Jason
_
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] DHCP question
Wolfgang- This is what happens when you use the 'default-lease-time -1' option? That would answer my question, then. If there is no true lease per se, and the only record of the address assignment is in the dhcpd.conf file, then even merely telling dhcpd to reread its conf file would be enough to handle the case of changing NICs in a client. So, if this is true, I think we could safely add (properly nested, of course) the 'default-lease-time -1' option to the dhcpd.conf.example file that is shipped with LTSP. Jim, It's ultimately your decision. How to you feel about this step? Does it make sense? Is it worth it for the rare case when the DHCP server cannot be relied upon? Jason Date: Mon, 4 Feb 2002 18:28:03 +0100 (CET) From: Wolfgang Schweer [EMAIL PROTECTED] Subject: Re: [Ltsp-discuss] DHCP question On Mon, 4 Feb 2002, Jason Bechtel wrote: Okay, but what about just sending it the signal to reread its configuration file (SIGUSR1, SIGUSR2, or whatever it prefers)? Will this also be enough to release the infinitely assigned address? If so, then we have an excellent, foolproof and tight solution to one of the small problems that currently exist with LTSP. I don't administer any DHCP servers at the moment, so I can't test this. At work (school) I've got to manage three DHCP servers. None of them has any entry in the servers leases file concerning clients with fixed addresses, only clients with dynamically assigned addresses get entries. Infinite lease time means IMHO, that the client will no longer ask for renewal. Wolfgang _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] DHCP question
David- Since many (probably the great majority of) LTSP environments are not done with uniform hardware in the workstations, one needs to specify different video, NIC, and perhaps kernel file settings for different workstations. That's not so true with version 3. The autodetection seems to work fairly well. Good point for NIC and video hardware settings, but what about different X settings for old monitors? Don't you run into this with all your doorstop workstations? Also, I know X4.2 fixes the S3 driver, but before 4.2 came out, how did you specify the XF86_S3 driver instead of 'auto' for your S3 clients? What about printers? If by address recycling you are talking about having LTSP workstations receive effectively randomly assigned IP addresses (and therefore hostnames via DNS/hosts) then the functionality of the lts.conf file is broken. No, that's not what I meant. One of the fundamental advantages of LTSP is that the terminals can be (and generally are) old equipment. This means that they are inexpensive, but prone to failure. When a terminal does fail (as I expect and plan for), I want to be able to tell someone to grab a spare from the closet and stick it on the desk. I want terminal installation to be so simple, even a pointy-haired boss can't screw it up. This lowers failure cost and downtime, which I consider a major selling point for LTSP. Okay, that makes more sense... :-) If the station losing its address is not going to return to the network, then there is less of a problem, unless several fail together (power spike, bad luck) and you run out of addresses (as you explain in your example with the C-class network). I'm still not convinced how you avoid *all* workstation-specific settings, though. Oh, wait, I think I get it. You *do* specify the workstation-specific settings, but by MAC address. And you just don't care which IP/hostname the thing gets. If you keep monitors with clients, this solves the X settings problem. But what about local printers? Hmmm... I think I'm getting confused. Can I see your dhcpd.conf file and your lts.conf file? ...all of the LTSP setups I've been involved in have had entries that match a MAC address to a fixed IP address, hostname, and kernel (either through bootp or DHCP). ... But for LTSP clients, there is no dynamic assignment whatsover. This is fine, but it increases network management, something I'm trying to get away from. I see. I'd never considered trying for that level of automation with LTSP. I just assumed that LTSP clients required a lot of configuration in various services and that if I wanted automation, I would have to find more automated ways of performing the configuration steps (never considering the option of trying to remove those configuration steps). I agree, one catch-all default is a good idea. We simply disagree regarding the usefulness of this specific default. I still argue that, in a normal work environment, dynamically assigned IP addresses with a one week lease provides a good approximation of static IP addressing without needing any oversight. I don't like the feeling of an arbitrarily chosen expiration time that approximates a good solution... I'm holding out hope for a lock-tight solution. :-) If you have a pool of LTSP workstations and are handling IP addresses dynamically... 1) I'd love to hear how you are pulling it off. How do you assign hostnames? Are they necessary? You can use nsupdate at login time to map hostname=username (this one has problems, of course), or you can simply name them ws001, ws002, etc, based on the last octet of the IP address. But if the hardware/location changes for the same IP/hostname, how do you handle the things I mentioned above (X settings, printers, etc.)? 2) Why bother? Just set aside a block of addresses to be assigned statically to your LTSP clients. What do you gain with dynamic address allocation if you always have the same number of clients? See my comments above. I am assuming that, since they are doorstops I rescued from the trash, the terminals are going to fail regularly. When they do fail, I want replacing it to be a no-brainer; in particular, I don't want to have to give anyone a password so they can muck around in my dhcp config or lease files. All you have to do is replace the MAC address in dhcpd.conf and restart the DHCP server. It's a 5-minute procedure at the most that you *might* have to perform once in a week. I agree that it does cost a finite amount of $$$ (or ¤¤¤), but it's not that bad. It sounds like you have workstations that all have the same (or all modern and equally capable) monitors and no local printers. In this situation, you might be able to get away with dynamically assigning addresses to your clients. But I think most people cannot depend on this uniformity. It seems to me that the default for LTSP should be to allow for the flexibility. For
Re: [Ltsp-discuss] IceWM questions...
mslicker wrote: It is very important to correct misleading presumptions made by authors of articles in the media. If the context is given that its two approaches for an administrator, versus a comparison of two technologies that are apples and oranges (in the broader context), then this article would be much improved in quality for all. It would not mislead and cause more misunderstanding of how UNIX GUIs are put together. I agree that one should try to correct statements that one finds misleading. I do not fault you in the least for speaking up. As I said, you are absolutely right (that from an academic perspective they are apples and oranges). We just had different grounding assumptions about the context of the discussion. Speaking of context, what article(s) are you talking about? Did all of this start over an article? I just searched (don't get excited everyone, I looked through manually) the archives on Geocrawler for the roots of this thread. It seems to have started with the subject line Help: Server reboot when out of memory! and some suggested using IceWM instead of KDE (this was on Red Hat 7.2 w/ 30 clients) to spare memory on the server: http://www.geocrawler.com/archives/3/10022/2002/1/300/7614590/ http://www.geocrawler.com/archives/3/10022/2002/1/300/7610093/ If this was the context of the discussion, then I don't see where you are coming from talking about authors of articles in the media and so on. This mailing list is certainly public and could influence people's decisions, but it's not like anyone performed an unfair lab test pitting IceWM against KDE on Red Hat 7.2 and then got it published on Slashdot! Frankly, I think we need a script that strips out and refined the KDE and GNOME systems for LTSP or light desktop use. I may likely work on such a thing in the coming months as this is an issue that related directly to my LTSP network. I think this is a great idea! :-) I think you would find much appreciation on this list for such a script. Many people want to have the bells and whistles of Gnome or KDE without all the bloat, but no one is willing (rightly so, IMHO) to bear the tedium of cutting away all of the fat. If such a script were available and customizable (to control the extent of the crippling), I think it would be one of the most popular utilities for use with LTSP! It does sound like a lot of work, though. Perhaps someone else here can help contribute to the effort. I say, go for it! Jason _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Root file system - Kernel Panic
Sabarim, Just a shot in the dark, but it might be a UNICODE problem (one end speaking UNICODE and another not). It looks somehow like a problem I had between Samba on Linux and the SMB implementation on an AS/400 where the problem turned out to be UNICODE. Sorry, I can't give you any way to debug the problem. My recommendation would be to get DHCP working on Linux instead. Jason From: sabarim [EMAIL PROTECTED] Date: Tue, 5 Feb 2002 20:02:27 +0530 The 'Root Path' given in the DHCP setup is x.x.x.x:/opt/ltsp/i386 , whereas the path displayed on the error line is x.x.x.x:/opt/ltsp.i386000 Here I'm not able to figure out where the the 000 gets appended ... _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] DHCP question
Date: Thu, 31 Jan 2002 20:44:39 +0100 (CET) From: Wolfgang Schweer [EMAIL PROTECTED] On Thu, 31 Jan 2002 [EMAIL PROTECTED] wrote: You can change your entries in your dhcp.conf file for: default-lease-time and max-lease-time to a new and larger amount. having set a fixed address for a ws, you can grant infinite lease time with this entry: default-lease-time -1 Jim and LTSP list, Should this be part of the automatic and/or documented DHCP settings when configuring LTSP (default, everywhere)? It seems like nothing is risked by giving infinite leases to fixed address workstations... The only question is, what happens when you replace a network card in a station and want to reassign the IP address to the new MAC address? And the answer seems to be that you would need to update the DHCP file anyway, and restart the DHCP server (or at least get it to reread its conf file with some sort of signal). My only remaining concern (no way to test it myself here) is if this would be enough to forget the infinite lease of that IP address to another MAC. Can anyone verify this? It would be a great addition to the polish and stability of LTSP. Jason _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] problem with X and monitor (was: problems with ltsp)
Date: Thu, 31 Jan 2002 11:25:34 -0500 (COT) From: John Jairo de la Rosa [EMAIL PROTECTED] Subject: Re: [Ltsp-discuss] problems with ltsp David, in your answer you tell me that the monitor can't handle the the signal coming from the card, but when I use windows 98 or Red Hat on this machine , there is no problems with the monitor, so I think the problem is by the LTSP, isn't it? John- The problem is that the settings that the X-server is using are telling the video card to produce a signal that the monitor is incapable of handling. If you choose to call this an LTSP problem that is your choice, but you would be facing the same problem with X-Windows and this video card and monitor combination even if you were not using LTSP. I would call it a X-Windows vs. monitor problem. When you had Red Hat installed locally on this machine, you probably easily avoided the problem because it came with a nice Xconfigurator/XF86Setup program to detect your video hardware and perhaps your monitor and it provided a ton of good settings options for X-Windows to work with. For the other monitor, can I configure the lts.conf file of this way?: [ws02] XSERVER=XF86_svga X_MODE_0=800X600 X_COLOR_DEPTH=16 RUNLEVEL=5 but I don't know how it's configured the refresh rate for having a slower refresh rate. It sounds like you are using XFree86 version 3.3.6 (your XSERVER setting starts with XF86_), in which case, you're best off installing Red Hat locally again (or putting the video card on an existing Linux box) and configuring XFree86 using XF86Setup or Xconfigurator. Then, copy the /etc/X11/XF86Config file to your LTSP server as something like 'XF86Config_stupid_monitor'. Then you can just reference that file in the lts.conf: http://www.ltsp.org/documentation/ltsp-3.0.0/ltsp-3.0.html#AEN1016 (XF86CONFIG_FILE option) NOTE: If you aren't using X336 (i.e. version 4.x), then this shouldn't be happening because 4.x does almost everything automatically. But since it is, I would switch back to 3.3.6, just to take more manual control of the situation. I think in the parameter X_VERTREFRESH, but I don't know which are its values. http://www.linuxdoc.org/HOWTO/XFree86-Video-Timings-HOWTO/index.html http://www.linuxdoc.org/HOWTO/XFree86-HOWTO/index.html And once you get a display you can at least read, use xvidtune to perform fine adjustments. Which values can I put in XSERVER? Read, please! http://www.ltsp.org/documentation/ltsp-3.0.0/ltsp-3.0.html#AEN1016 http://www.ltsp.org/documentation/ltsp-3.0.0/ltsp-3.0.html#AEN1529 http://www.ltsp.org/documentation/ltsp-3.0.0/ltsp-3.0.html#AEN1535 and if you still need more info, http://www.xfree86.org/support.html I believe in an earlier message that you indicated that you thought the problem might be with this setting. But if you are getting as far as you are, then this setting is fine. Otherwise you would get an error about devices. please, send me a example of this configuration because Tomorrow I must deliver the solution for this problem Remember, you're getting this helpful advice *free of charge* from a community of volunteers. That's all I'll say to this. Jason PS: Jim is constantly repeating this on the list to people having problems configuring X on the station... Start in RUNLEVEL 3 and manually type /tmp/start_ws so you can see the errors on the console (Shift-PgUp to go back). You should be in RUNLEVEL 3 the whole time you are debugging this. _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] HOWTO's seeking maintainers
LDP Weekly News lists HOWTO docs without maintainers: http://lwn.net/2002/0124/a/ldpwn.php3 I realize that this is a bit off-topic, but the reason I mention it is that one of them is the NFS-Root Mini-Howto, which is pretty important to LTSP booting. I thought about offering to maintain it, but then thought that perhaps someone on this list would be better suited. If anyone is interested, the link above says the following: If you're willing to become the maintainer for one of them, please join the LDP discussion list at lists.linuxdoc.org and post a message indicating your interest. Jason _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] IceWM questions...
Matthew, From a scientific viewpoint it naturally makes more sense to compare the actual window manager portions of each system. But why would any system administrator responsible for dozens, perhaps hundreds, of corporate Linux desktops take the time to customize KDE so that it runs faster and consumes fewer resources and doesn't provide too much rope for the users to hang themselves with just to find that the next release of KDE changes where all its config files are stored, the organization of its modules, etc.? Why would the sysadmin change distributions to Caldera (assuming most are not already using it) to get a decent preconfigured KDE? IceWM is sitting there waiting to be used effectively in its out of the box configuration on any distribution. It has all settings in one configuration file. It allows for individual configuration files *if needed* but doesn't force them on you. It does provide desktop manager type features (toolbar, menu, hotkeys) without actually allowing users to manage (alter) the desktop themselves. You can always give them this ability with add-on utilities and personal configuration files. http://www.icewm.org/index.php?page=utilities Don't get me wrong. Your point is absolutely correct. KWM and IceWM are probably very comparable in their speed and their ability to put decorations and controls on windows and to place them sensibly on the screen. But no one runs just KWM. The point that myself and some others are making is that the reality of the situation (what matters) is what users and administrators actually do with the software. And in my opinion, I would rather install a fast simple system that doesn't overload my server and then add functionality as needed. I don't think anyone is blaming KWM for KDE's bloat. They're just part of the same package. Jason PS: XFce 3.8.14 is out http://www.xfce.org/. Date: Thu, 24 Jan 2002 01:48:13 -0800 (PST) From: mslicker [EMAIL PROTECTED] To: Michael H. Collins [EMAIL PROTECTED] Well then if you can compare Windows Managers (WMs) to desktop systems then why not compare KDEwm to ICEwm? Let me tell you something, without the underlying desktop componant architecture (and applications that use it), you will use vastly more system resources as you add simultaneous users and applications. KDE and GNOME tend to be slow on most major distrobutions because the distrobutions are configuring them in a very heavy manner. The same number of things configured without a desktop system would be completely unthinkable in terms of performance. Try Caldera's 3.1 desktop to illustrate how fast KDE can be, if better refined. Red Hat, Mandrake, and SuSE stuff as much flashy garbage in their distrobutions as they can so they can make money on CD sales.everyone upgrading to see all the new things. Caldera's focus is on business systems and is therefore the only refined distrobution I am aware of. They do not focus of stuffing as much on a CD as possible and quick upgrades for CD sales.. Unfortunately, their model isn't as profitable in the short run. Most impression is, probably a lot of you who are advocating ICEwm as a replacement for KDE or GNOME are actually using both desktop systems and not realizing it. You can get similar performance just by optimizing all the stuff that gets loaded with you default desktops. The KWM (KDE's default Window Manager) by itself is in fact very fast, I'd even think it might be faster than ICEwm. Or, try replacing KWM with ICEwm and then see if KDE runs any faster or slower? What I'm saying is, you are blaming the wrong thing for you desktop's slow performance. I gaurantee it. --Matthew _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Help: Server reboot when out of memory!
From: OFFRAY LUNA [EMAIL PROTECTED] Date: 21 Jan 2002 21:27:59 -0500 Hi Mohamed, Yes, you will find usefull to swicht to another desktop enviroment as IceWM in order to reduce the load of the server. I would probe making a custom config of IceWM to solve the most of the taks of the users (launch StarOffice, etc) and then I would make this custom IceWM as the default desktop and I would see the load of the server. I suppose I should mention to anyone using IceWM with LTSP that I've written a couple little scripts to improve the experience. One thing that the new KDE and Gnome have which is nice is the application startup feedback (like Windows' busy cursor, but they have many more ways of showing it). This lets users know that their application is starting so they don't go and click the button 20 more times. The feedback script runs out of another script which creates lockfiles to suppress multiple launches. The URL is http://www.uniqsys.com/~jbechtel/tk_start.tgz And for anyone NOT running public access terminals (library, kiosk, internet cafe) I have also written a pop-up window for use with IceWM's built-in email notifier. It includes a button to launch netscape mail (replace command with whatever is appropriate for your users) directly from the notification window. The URL is http://www.uniqsys.com/~jbechtel/tk_newmail.tgz You can try also disable memory intensive apps like screen savers. Actually, I don't think screen savers are so memory intensive as they are bandwidth intensive. Unless you are running them local to the clients, you should *NEVER* use screen savers with your terminals. They will drop your network to its knees. Good luck! Jason PS: Last time I checked, KDE and Gnome were both at around 25MB resident memory per session and IceWM was at 1.4MB. _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] User Profiles?
Hi David, KDE has configuration files like any desktop environment. You can always create a test user for each profile and set up the desktop exactly as it should be for members of that group. Then, add some scripting at some point during login, but before KDE is launched (Xsession, for example), to make sure that $USER has the correct KDE configuration files in their home directory. But this is messy and when KDE3.0 comes out, you'll probably have to revise all of your prototype environments and perhaps your scripting as well. I would suggest switching to IceWM. You can still use your KDE apps (like Kword), but you have about 5% of the memory footprint of KDE. Also, IceWM is smart enough to check the existence and executability of the items in its menus and toolbar items before mapping them. It will simply skip over programs that the user cannot run. Then, you can just assign users to groups that correspond to certain programs and make the programs group executable only. Then each user only receives the option to run programs that he or she has access to run and management becomes separate from the particulars of the configuration files of the desktop environment. Jason From: Rose, David [EMAIL PROTECTED] Date: Tue, 15 Jan 2002 11:05:00 -0500 Hello. Just installed LTSP and got it working. So far, everything looks awesome! A question that I have is, I would like to make 'profiles' -- both with regard to rights to the file system and to having a common desktop appearance / certain programs available to the user. I understand how to make groups and give the group rights to the file system; however, for creating 'groups' of users which have a certain KDE configuration I don't have this answer. If I were to have 10 graphic artists who were to use KWord and GIMP; 5 finance people who were to use gnumeric and kword, etc.; how would I set up the LTSP server so I could easily manage who gets what apps (and menu)? Citrix has userprof.man file which states which programs a person can use, and which icons are displayed. Does LTPS? _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] DHCP Configurator
James, I don't know about having incompetent users being able to muck around with your LTSP client setups, but... :-) If you are referring to the webmin module developed by the Termserv project which only modifies lts.conf, then you are right, it does not allow you to add stations in DHCP or DNS (yet). But there are already very good webmin modules for DHCP and Bind(DNS) that come standard with Webmin. If you are interested in helping to extend the Termserv webmin module to handle DHCP as well as just lts.conf, please see the following links and join. Your help would be welcomed. Indeed, there should eventually be a way to add workstations as you say: supply the MAC, station name, mouse type, NIC driver, and video driver all in one form in the Webmin module and let the module handle the details. http://termserv.berlios.de/index.html http://termserv.berlios.de/ltsp-module/ http://developer.berlios.de/projects/termserv/ Jason From: James Vasey [EMAIL PROTECTED] Date: Tue, 15 Jan 2002 20:07:40 - Is there, or is anyone working on a method of quickly allowing a non-competent user to quickly add workstations, and configure them from within a web browser i.e. the user enters the mac address, AT/PS2 connector, driver, etc? I know there is a webmin for the LTSP server side, but haven't seen one for setting up clients? If anyone is working on one, let me, perhaps work together toward a solution? James Vasey - desperate, confused, and trying to supply LTSP to a school with a network tech from mars. _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: Xmovie player
DK- When I needed it, the MPEGs were soundless as were the clients. Even if you could get a player to redirect the sound to the workstation, the lag would probably be intolerable. I would run Xmovie locally on the station to eliminate the lag problem. If you can get sound and local apps working, I don't see why xmovie wouldn't work. According to the Changelog, version 1.9: Migrated to GCC-3.0, Linux 2.4.7, libc-2.2. SuSE 7.2 and 7.3 both come with version 1.5.4, so I think this version is recent and stable enough. As for distortion of the picture, if you are running the app remotely, you have to expect this. Xmovie is generating the frames on the server and then they are being sent over an already busy LAN to be displayed on the client. I wouldn't expect any good performance with this setup. If your clients are fast (you mentioned PIII), then running the movie softare locally would probably improve the situation dramatically. There's also XAnim: http://xanim.va.pubnix.com/home.html XAnim is a program for playing a wide variety of animation, video and audio formats under X11. Supports Unix, VMS, Amiga, and W95/NT. Formats include AVI, MPEG, Quicktime, Animated GIF, SGI Movie, Replay, IFF, FLI, FLC, DL, Amiga MovieSetter, and JFIF. Unlike Xmovie, XAnim seems to provide the option of specifying the sound device on the command line. So, if you absolutely must run it on the LTSP server, you could tell it to use a device that redirects sound to the client. I would still expect problems with video quality when running XAnim remotely, though. Jason PS: A quick search for 'VCD' on freshmeat.net found Nonton VCD at http://oss.mdamt.net/vcd/ which claims the following: * VCD seek * Full Screen * Multiple Video size * Sound level control * i18n support vcare wrote: I used the latest version of Xmovie but it gives me problem of lib files so i downloaded the older version , but it is enable to give me sound on diskless client as well as the picture quality is also distorted _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Alan Cox digs LTSP
Sorry if I'm not the first to mention it (I subscribe in digest mode), but Alan Cox mentioned LTSP very favorably in his interview with KernelTrap. I picked it up from Slashdot and followed the following link: http://kerneltrap.com/article.php?sid=490 Congratulations to Jim and everyone who has helped make LTSP the success that it is! As AC says, _Em_powered by Linux! _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Natsemi Nic ends with a ERROR
Last time I needed it (for a netgear), the natsemi Linux driver was a proprietary, closed source driver. So, it can't be compiled into the kernel and has to be loaded explicitly. Jason Subject: Re: [Ltsp-discuss] Natsemi Nic ends with a ERROR From: [EMAIL PROTECTED] Date: Thu, 10 Jan 2002 02:16:09 +1100 You need to specify this in the /etc/dhcpd.conf file. Take a look at the instructions-3.0.html file on the ltsp.org site. As the natsemi based NICs are PCI NICs (probably the Netgear FA311 or FA312, or the Linksys something) shouldn't it be automatically detected by LTSP? _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Wireless Etherboot
Samarth- I don't know anything about wireless Etherboot, but you already have Linux installed on the hard drives. You could save yourself a lot of trouble and use the existing X-Windows capabilities on the local machines. Just replace the current way of starting X-Windows with a call like the following: X -query bigserver Then just enable XDMCP on the server (in gdm/xdm/kdm). Admittedly, this does not give you the benefits of a truly diskless setup, but I don't believe there currently exists a slam dunk solution for using LTSP with laptops. You do get what I would consider the most powerful feature of LTSP... centralized updates and administration of applications. Jason From: [EMAIL PROTECTED] Date: Fri, 4 Jan 2002 14:38:30 -0500 I have 10 laptops, Compaq LTE 5400's on which I am running Slackware. The configuration is as follows: Pentium 150 80 MB RAM 2GB HDD Cirrus Logic Adapter GD7543 1MB VRAM Lucent Wavelan Wireless Cards. These laptops run X in a 800x600 mode at 16-bit color. The performance takes a big hit when it comes to running browsers and other applications at the same time. I am planning on setting up a couple of systems as workstations connecting to a PIII-500 512MB RAM server. This will allow a performance comparison and based on the users responses, all the other laptops can also be setup the same way. I am trying to find the wireless etherboot image for these wifi nics but haven't really found one. If anybody in the list has tried this out previously, any suggestions on where to look will be greatly appreciated. _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: slow and strange mouse movement (was: NFS, 2.4.16 other)
Vladimir- I don't recall which version of X your workstation is running, but have you checked out the XFree86 documentation on their website? It's sometimes helpful: http://www.xfree.org/3.3.6/mouse.html http://www.xfree.org/4.1.0/mouse.html I'm also not sure which video card (X driver) you are using. You should also check out Configuration options and/or Driver options for your particular driver. At the bottom of the following pages are links to information for the various drivers: http://www.xfree.org/3.3.6/index.html http://www.xfree.org/4.1.0/index.html Sometimes you can disable hardware acceleration (no_accel, no_imageblt, no_linear) or switch between the hardware and software cursor (sw_cursor, hw_cursor). These options are hardware dependent, so check the documentation first. Oh, and you say that your settings work with the other PC, but does this other PC have the same hardware (video, mouse) or not? Sometimes mice sort of work with the wrong settings. See the Configuring your mouse section under the first two links provided: http://www.xfree.org/3.3.6/mouse4.html#20 http://www.xfree.org/4.1.0/mouse4.html#20 Jason PS: If none of this helps, please report back to the list with your exact mouse hardare type and your motherboard make and model information. There have been problems in the past with mice where it turned out to simply be an incompatibility between the motherboard and the mouse. Date: Wed, 12 Dec 2001 11:08:32 +0200 From: Vladimir N.Velychko [EMAIL PROTECTED] I think the problem with your mouse is not something having to do with your server, it's in the lts.conf, you will have to change mouseprotocol and device, maybo from ps2 to the serial ports. But i do not know the exact lines right now... It was in one of my previous letters: --- In my case X start but mouse cursor like frozen - movement is very slow and strange :( lts.conf: * [ws001] X_MOUSE_PROTOCOL= MouseSystems X_MOUSE_DEVICE = /dev/ttyS0 ... * XF86Config: * Section InputDevice Identifier Mouse0 Driver mouse Option Protocol MouseSystems Option Device /dev/ttyS0 EndSection * With the same config lines and mouse all work fine on the other PC. Any suggestions - welcome. --- Of course I try other mouse protocols without success :( _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Tuning advice
jhazer, There have been many discussions about this before, but I know the archive search is broken, so here's a quick summary... To figure out what your video card can do, use this formula: pixels = x_dim * y_dim bytes/pixel = bit_depth/8 VRAM required = pixels * bytes/pixel = x_dim * y_dim * bit_depth / 8 So, if you want 800x600 with 16-bit color, your card needs 960,000 bytes of RAM (basically, 1MB of video RAM). Now, what really concerns me is the monitor. Those old heavy 14 monitors just suck to configure X for. They do not accept a wide range of signals and often being close but not quite right results in headache-causing flicker or distortions. If you have tons of time to play around with the thing, be my guest, but if you just want to get your station up and running, look for a newer monitor. If you choose to stick with the little 14 monitor, here are two approaches to tweaking the display: 1) Get it working so that you can at least see the screen somewhat and read text. Then run 'xvidtune' and play with the parameters until the screen gets better. It might not make sense immediately, but stretching and moving the screen around can improve flicker conditions. When it looks good, tell xvidtune to spit out the modeline and then program it into the lts.conf file for that station. 2) Pull out the bazooka: download the LTSP Enhancements package, listed on the LTSP Contrib page. A part of their package allows you to run X configuration tools locally on the station and will store the results back on the server for you. Read the docs. Jason From: [EMAIL PROTECTED] Date: Fri, 7 Dec 2001 19:05:04 Any advice on how can i fine tune a workstations display? Is it possible = for an isa generic vga and a 14 Paperwhite vga monitor to have a 800x600= resolution and 16 color depth? and oh btw, i tried XF86_VGA16 and it outputs an error that i have to put= vga16 or Mono in XF86Config. Need help on this also. _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] LTSP in a home environment?
Andy- I'm sure I won't be the only one to respond saying that you are overestimating the hardware demands for your purposes. My comments follow your text... From: Andy Berry [EMAIL PROTECTED] Date: Sun, 9 Dec 2001 12:52:09 -0700 I currently have two different computers, both M$ boxes snip Cost is no issue, as the ltsp server would be little or no added expense than the file server, and I have clients available, so I would be spending no more money, with the exception of the $75 or so it would take to get bootable network cards to be able to go diskless right away. I have Linux experience, so I am not worried about getting in a jam, and I will not try to switchover all at once, but get my server and one client going before totally changing over. I came up with a list of the following questions: 1) Is the basic premise even feasible?? Quite. I know at least one person on this list who has wired his home and installed flat-panel touch screens in the walls and is running the whole thing with LTSP. I don't have the link handy, though. 2) How sensitive is diskless to different client configurations?? I have two Pentium (P5-100/64M RAM) boxes that would be my first clients. If I was to upgrade one to current standards (PIII/256M), would I have problems? Do I have to keep video modes identical, or can they be different? Anything besides video that might be crucial? LTSP supports a completely heterogeneous environment. You can have different settings for each client machine all in one file. As for the client hardware requirements, your Pentium machines with 64MB sound like ideal thin clients! Seriously, do not waste any money upgrading them. The only reason you might consider an upgrade is if you want to run applications locally on the thin clients, but it sounds like you want to run them on the server so that central access to data is transparent. 3) What is a good configuration in today's world?? I know this is a FAQ, but nothing really addresses here and now. Assuming 3-5 clients and 3-4 users as the ultimate goal, would the server be OK with a PIV and 1GB or better RAM, and fast SCSI disks? Down the road, I expect my clients to slowly migrate up to the lowest of the currently available parts. Looking around, I could build a PIII-700/256MB client for ~$300+ monitor. Overkill? Should I just seek out the cheapest used P5 in my area? A PIV w/ 1GB and SCSI could power 20-40 thin clients. I'd say that PIII-700 with 256MB is looking like a good server for you... Maybe up the RAM to 384 or 512 just in case, though. It all really depends what kind of software you're going to be running. If you're planning on using KDE/Gnome, StarOffice, and Mozilla, then allocate a large amount of RAM for each session (~128MB) on the server. If you use a light-weight window manager instead and if you aren't using memory hogs for applications, then you can get away with less. Then be sure to account for the base RAM that the server will want to use. Again, the clients only need to run X-Windows (~16MB RAM) so don't think that they have to be speedy. LTSP's real strength is in running everything on the server so you can concentrate on paring down your clients. Many people on this list find joy in achieving the fanless client (no moving parts!). 4) Am I missing something obvious??? LTSP takes the concept of thin client computing to the extreme. It takes a little while for all the implications to settle in... :-) Jason _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] KDE upate! (i broke it)
Matt- I want to have the clock on the logon screen (it just looks nice) What do I have to put in the config file to make it a clock? Are you saying that the Control Center (kcontrol) running as root will not allow you to modify the KDM screen?? I'm using KDE 2.2.2 on SuSE 7.2 and it works fine and modifies the right file. If it's not working, why not use the information provided by Martin Herweg and do something like this: rm /opt/kde2/share/config/kdmrc ln -s /opt/kde2/share/config/kdm/kdmrc /opt/kde2/share/config/kdmrc or the other way around, so that regardless of which location kcontrol tries to modify, it will get the kdmrc file. Right now it has this LogoArea=Kdelogo I want it to be a clock Mine says LogoArea=Clock And can I make this a different clock (while I'm at it I digital one would be cool) If kcontrol doesn't provide the option, it probably doesn't exist. Jason _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Netscape problem on terminals
Hugo, Date: Wed, 28 Nov 2001 11:25:42 -0600 (CST) From: =?iso-8859-1?q?Prolinux?= [EMAIL PROTECTED] I need your help solving netscape run away on ltsp terminals, it runs away so often without freeing cpu. I am using Pentium 75 with 16Mb and 500Mb HD as terminals and I would like to either replace netscape with konqueror, mozilla or galeon, but with java support, or use netscape as a local application (if my current hardware could support it). Any suggestion? I've run into problems with Java and Linux web browsers myself. That's what usually causes Netscape to run away on Linux, I think. You will have trouble finding a combination of browser and Java versions to correctly load the website you are looking at. 16MB is barely enough to run X-Windows. You would need at least 32MB on each workstation (64 would be better) to run netscape locally. If you're using the 500MB HDs for swap, however, then 32MB is fine. I think the stations would run really slowly with only 16MB, though. They would be swapping constantly! I am using mandrake 8.1 and made some attempts to download different versions of jre (java runtime environment) to setup either konqueror or mozilla to use java plugins but without success, when I load web pages that need java support they come incomplete, but netscape works fine on them. Could someone try latinchat.com and tell me if chatting is posible with a different browser than netscape? I just tried the latest jre's from Sun and IBM with Mozilla 0.9.6 and the chat site Java Lite applet didn't load. I think you are stuck with Netscape, in which case there are a couple things you can do to fix the running away problem: 1) Install verynice. It is a dynamic process renicer. It will automatically give runaway processes a lower and lower priority (higher and higher niceness level) until they reach a certain threshold after which it will attempt to kill the processes off using various kill signals. Make sure you setup the /etc/verynice.conf file the way you want. It can be configured to ignore certain processes or to pay special attention to certain processes (this is what you want). The link is http://tam.cornell.edu/~sdh4/verynice/ and there is an RPM package and installation instructions. I've used this package with much success. 2) You could download my kludge of shell scripts from the LTSP contrib page http://www.ltsp.org/contrib/index.php and run the main script (sickem) out of cron. It is not nearly as elegant as verynice, but it also gets the job done. Jason _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
[Ltsp-discuss] Re: LTSP local devices
LTSP lovers, The link John Gay was talking about for Enhanced Network Block Devices (enbd) is http://www.pcxperience.com/enbd If anyone on this list would like to see a real solution to accessing removable media on LTSP workstations, *please* visit this site and pledge some money toward the completion of this project. None of the existing methods really work and this is a vital feature for many implementations of LTSP. Jason _ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
Re: [Ltsp-discuss] Local floppy access .
Ola,
gdm uses its own configuration, while kdm has historically used the
existing configuration files for xdm. This may have changed in more
recent version of kdm, but you can try putting this scripting in the
file /etc/X11/xdm/Xsession and see if it works.
Jason
Date: Thu, 22 Nov 2001 14:50:59 +0100
From: Ola Ketil Siqveland [EMAIL PROTECTED]
If I want to use KDE with KDM where shall I put these command
below can I use the same as for gdm
Ola Keitil Siqveland
Nancy Solano wrote:
2) Fix gdm. Use this scripting in /etc/X11/gdm/PreSession/Default
instead of what is currently in the docs:
# Check for remote floppy on LTSP workstation
if [ -b $HOME/.dev/fd0 ]; then
FLOPSTN=${DISPLAY%%:*}
FLOPSTN=${DISPLAY%%.*}
/usr/local/bin/nbd-client $FLOPSTN 1025 $HOME/.dev/fd0
fi
This alleviates the need to use the username to bind to the
terminal.
The DISPLAY variable contains all the information we need for
that. I
also used HOME instead of USER because a given user's home
directory is
not always /home/$USER and the HOME variable is available. It
also only
runs for users with the block device $HOME/.dev/fd0, so it
discriminates
based on user, but adapts to wherever they are logging in from! :)
"
Description: Binary data
Re: [Ltsp-discuss] Problem using LTSP and DSL
Oliver,
Welcome to the Linux community! And thank you for your detailed
description of your problem. That makes it much easier to understand
where you are coming from.
Now, there's no reason that you shouldn't be able to do this. The
main problem is that your networking needs to be straightened out
and then your DHCP configured properly.
First, if you want your clients on the same subnet as your DSL router,
then you do not need the two network cards. But since you already
have two network cards, you should stick with this setup. It's better
to keep the client traffic separate from the internet traffic.
Your two network cards (in the server) should be on different
subnets. Keep whatever DSL network settings are already working and
change the subnet with the diskless clients to something else. Let's
call the DSL subnet 192.168.1/24 and the clients' subnet
192.168.0/24. The /24 means a 24-bit netmask (255.255.255.0), which
is the most common for local networks with 192.168.x address spaces.
Now, get your Realtek card set up for the DSL subnet. You could use
192.168.1.1/24 and that should get you working. Make sure your
server (not running DHCP yet) can use this interface and establish
internet connectivity.
Now, your IBM NIC should be set up for the 192.168.0 network as
192.168.0.254/24. Now bring up this interface and run the command
netstat -rn
and look at the output. It should look something like this:
Kernel IP routing table
Destination Gateway Genmask Flags MSS Window
irtt Iface
192.168.1.1 0.0.0.0 255.255.255.255 UH0 0
0 eth1
192.168.0.254 0.0.0.0 255.255.255.255 UH0 0
0 eth0
192.168.1.0 0.0.0.0 255.255.255.0 U 0 0
0 eth1
192.168.0.0 0.0.0.0 255.255.255.0 U 0 0
0 eth0
127.0.0.0 0.0.0.0 255.0.0.0 U 0 0
0 lo
0.0.0.0 192.168.1.254 0.0.0.0 UG0 0
0 eth1
The identities of eth1 and eth0 may be reversed. That's okay. And
the last line for the default route should contain the IP address of
your DSL router where I've placed 192.168.1.254.
Now, your networking is setup and you need to configure DHCP to work
properly in your environment. You need to set it up with a 'subnet'
section for each subnet or it won't start. Do not declare a 'range'
or it will attempt to handle the IPs dynamically. You want to do
something like this:
option subnet-mask 255.255.255.0;
option routers 192.168.0.254;
subnet 192.168.1.0 netmask 255.255.255.0 {
}
subnet 192.168.0.0 netmask 255.255.255.0 {
group {
filename /lts/vmlinuz.rtl8139;
use-host-decl-names on;
host ws001 {
hardware ethernet 00:80:C8:D9:53:E4;
fixed-address 192.168.0.1;
}
host ws002 {
hardware ethernet 00:80:C8:D9:53:E5;
fixed-address 192.168.0.2;
}
host ws003 {
hardware ethernet 00:80:C8:D9:53:E6;
fixed-address 192.168.0.3;
}
}
}
Now, on starting DHCP. You don't want to just type 'dhcpd start'. I
suspect that this is invoking the daemon directly because it's in your
PATH. If you type
which -a dhcpd
you will probably be shown the path to the daemon itself:
/usr/sbin/dhcpd
The scripts in /etc/init.d/, which are named for the services they
manage, do far more than simply call the daemon. You should always
start and stop services using these provided scripts. So, now to
start dhcp, run
/etc/init.d/dhcpd start
If it doesn't start properly, send us your /etc/dhcpd.conf file and
the output of 'netstat -rn'.
Jason
Date: Thu, 22 Nov 2001 10:29:23 +0100
From: Oliver Krehan [EMAIL PROTECTED]
Hi all,
first I have to say, that I am using Linux now for only two months.
Because
of this, I am not very good in explaining problems to other and
better
users. I hope, I can give you a good explanation.
I have a problem using LTSP in combination with DSL.
First I set up LTSP_2.08 with all the standard parameters in the
config-files and everything worked fine.
The installation consists of a server with a IBM Fast Ethernet LAN
Card and
3 clients, all equipped with Realtek 8139-chipsets.
Then I have installed another LAN card for the DSL, a Realtek 8029.
The NIC
is configured properly and the DSL is running fine. The only problem
is
using both together.
Because DHCP is running, the DSL can not be initialized and vice
versa.
When I boot up my RedHat Linux 7.0 without DHCP and with no clients
attached and then start the DHCP, there is an error message saying,
that
both adapters use the same subnet. The IP-address of the IBM NIC is
[Ltsp-discuss] Re:
yossi, You can also do this from within your Tcl/Tk app. The command is wm overrideredirect . I haven't tested it with Gnome, but it should handle 'wm' commands like any other good WM. Tcl rocks! Jason Date: Sun, 11 Nov 2001 15:06:08 +0200 From: yossi [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: [Ltsp-discuss] Making the user interface simple Running 2.08 with 20 terminals and an app written in tcl/tk interfaced to a sql server. The users are mostly computer illiterate and the application only uses buttons to make it simple. I need some advice as to how to configure gnome so that the user only sees the tcl/tk application and cannot use any of the gnome controls nor the task bar not the start button, etc. Ideally when a user logs in, they receive one full screen window without resizing controls - that's it. " Description: Binary data
Re: [Ltsp-discuss] Desktop config
Grgicevic, It is possible to do this with KDE, but it takes time and it is annoying to have to write scripts to handle all this stuff. The easiest thing to do is to use something other than KDE. I would recommend IceWM or XFCE or another small window manager. IceWM can be completely controlled from one configuration file for all users. New users automatically inherit these settings and there are no desktop icons for hardware that would actually be on the server. And there is no messing up the desktop because the user has no controls over the desktop itself. I don't know about the German keyboard, but it should be doable. Finally, it's better for your server, too! KDE uses over 20MB per session!! IceWM is a cool 1.4MB... Jason From: Grgicevic Davor (VTG) [EMAIL PROTECTED] Date: Mon, 12 Nov 2001 13:10:35 +0100 Hi .. I am tring to costumize my LTSP server The rason for this is .. To get .. For all kids at start The same desktop and.. To put/remove neccesery programs from desktop . For instance cdrom or floppy .. Or other things..( default german keyboard, Desktop definitition . Ect. ) And to make default desktop .. For shool kids.. It should be automatic installed with deffinition of user And when kids mess there desktop , just rm -rf /home/$user/.kde And his original desktop is back :) Does somebody has expirience with this or some ifos/docu/HOWTO/tips ?? " Description: Binary data
