Re: Booting Linux from windows 2000
I don't know if you've found a suitable answer for this yet or not, but I have done what you ask about in the subject line. You use dd to make a copy of the boot sector for the linux partition, you drop in the root directory of the C: drive and then edit boot.ini to include that file. Then it works just fine. The only pain about this setup is that anytime you change your linux boot sector (at least with lilo) you have to use dd again to create your boot sector file again and then copy that into the root on the c: drive again so that the changes are reflected. If you would like an example of the boot.ini entry you would need, let me know I might be able to dig one up on one my systems. vec - Original Message - From: Nathan E Norman [EMAIL PROTECTED] To: debian-user [EMAIL PROTECTED] Sent: Saturday, November 09, 2002 1:41 PM Subject: Re: Booting Linux from windows 2000 On Sat, Nov 09, 2002 at 02:39:33AM -0500, Scott Henson wrote: On Fri, 2002-11-08 at 14:18, Pigeon wrote: It is true that Windoze doesn't like changes to the MBR. To hack the Win98 MBR I had to include code to put the original MBR back after the hack had done its work, then make Windoze reinstall the hacked version after it had done its check. That relied on having DOS available underneath and probably would be much harder in 2k. It's not something I would really recommend! No, Windows 95/98/ME all do just fine without the MBR. You can have lilo or grub steal it and windows wont even figure it out. Now Windows NT/2000/XP all need the MBR and wont boot without it. Under these you have to use the windows boot loader to load grub or lilo then they continue to boot linux. Best way is probably to make a grub boot floppy though. Not true. Win2k boots fine from grub, which is installed in the MBR here. -- Nathan Norman - Incanus Networking mailto:nnorman;incanus.net prepBut nI vrbLike adjHungarian! qWhat's artThe adjBig nProblem? -- alec flett @netscape -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Microsoft's plans to kill open source: TCPA
- Original Message - From: Josh Rehman [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Sunday, November 03, 2002 11:01 AM Subject: RE: Microsoft's plans to kill open source: TCPA It's really hard to control the container because people are smart, and A person is smart; people are stupid. Agent K, in M.I.B. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: VERY OT: mass installation on XBox
Tom writes: Man, and you guys call this the land of freedom? The Housing Associations he is talking about are private associations of homeowners. He agreed to abide by their rules when he bought his house. -- yeah, well...it used to be the land of freedom. It's still closer to freedom than most places in the world but that seems to be rapidly changing. Housing Associations *were* a good idea for *some* people, but they have gotten *WAY* out of hand. There are housing associations near my neighborhood that force you to have a tree in your front yard. How asenine is that? Believe it or not, if you don't have a tree in your front yard, they can leagally take your home and property from you. The very notion of property ownership in the US has become, at best, a joke. Even if you do have it paid off, they city in which you live can come and take it from you because you aren't conforming to their beautification ordinances or other suppoed 'reasons' like that. The county in which you live can come take it from you because they need it for county business. The state in which you live can come and take it from you for just about any reason they want. And as far as the federal gov was concerned it was theirs to start with. People say 'oh that never happens' but it happens all the time. It just isn't an epidemic so the masses tolerate it. Sad indeed. A very nice quote was sent to me today which sort of sums up the shituation in the US right now as far as I'm concerned: Beware the leader who bangs the drums of war in order to whip the citizenry into a patriotic fervor, for patriotism is indeed a double-edged sword. It both emboldens the blood, just as it narrows the mind. And when the drums of war have reached a fever pitch and the blood boils with hate and the mind has closed, the leader will have no need in seizing the rights of the citizenry. Rather, the citizenry, infused with fear and blinded by patriotism, will offer up all of their rights unto the leader and gladly so. How do I know? For this is what I have done. And I am Caesar. - Julius Caesar -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT Any web site that teaches how to make LAN cable connection
Even that is incorrect. The pins that ethernet uses with RJ45 is 1/2 and 3/6, not 3/5 or 4/6. Straight Through: One End:Other End: 1-1 2-2 3-3 6-6 Null Cable One End:Other End: 1-2 2-1 3-6 6-3 And if you really want to be thorough then on the straight through 4 - 4, 5 - 5, 7 - 7, 8 - 8 and on the null cable 4 - 5, 5 - 4, 7 - 8, 8 - 7 vec - Original Message - From: Mike Dresser [EMAIL PROTECTED] To: Elizabeth Barham [EMAIL PROTECTED]; [EMAIL PROTECTED] Cc: list debaun debian-user@lists.debian.org Sent: Wednesday, March 27, 2002 9:44 AM Subject: Re: OT Any web site that teaches how to make LAN cable connection There's straight-thru and crossover - for crossover, they should be reversed. If the original is BLUE-RED-GOLD then the matching one should be GOLD-RED-BLUE (example is for explanatory purposes only). Straight-thru is literally that ; the wire on both ends should look the *exact* same through the RJ-45. Straight-Thru: 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 (bottom) (bottom) Cross Over: 1 2 3 4 5 6 7 8 8 7 6 5 4 3 2 1 (bottom) (bottom) Elizabeth Please don't follow either of these wiring diagrams. It's very important that pin 3/5(or 4/6, depending on how you look at it) are swapped on both ends. You want: (T-568A) White/green, green, white/orange, blue, white/blue, orange, white/brown, brown Sometimes you'll see (T-568B, preferrred for new installations) White/Orange, Orange, White/Green, Blue, White/Blue, Green, White/Brown, Brown. Both will work the same. For crossover: Do above, one side the orange/blue/green/brown setup, and the other side the green/blue/orange/brown setup. If you don't swap the 3/5(4/6) pins, your cable will might work, but you'll get massive crosstalk, especially over distance or in an electrically noisy environment. Please see: http://www.escape.ca/~droopy/ethernetcables.html http://www.incentre.net/incentre/frame/ethernet.html Also, see: http://catalog.belkin.com/IWCatSectionView.processs?IWAction=LoadMerc hant_ID=Section_Id=522 http://www.startech.com/cable/networking.htm Because that's a LOT easier and more reliable way to do it! If you are making your own cables, proper cable testers cost far too much money. I make do with an old Fluke 620, but it isn't rated for 100mbit or gigabit. Hope this helps, Mike Dresser -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: VPN on Kernel 2.4.18
- Original Message - From: Jeff [EMAIL PROTECTED] To: Debian User debian-user@lists.debian.org Sent: Wednesday, March 06, 2002 10:53 PM Subject: Re: VPN on Kernel 2.4.18 Dave Scott, 2002-Mar-06 15:30 -0800: Thanks Jeff. Wow, I thought this was going to be an easy task. :( Surely there must be thousands of others that have done just this. |-| |-|Client(9x,2K)| | |-| | |DSL Modem | | | | | |-| | |-Internet| | |-| | | VPN | | |Public IP | |--|Nat || | |Firewall--|--|Workstations| | |--Debian 2.2 W2.4Kernel---| || | |--| 192.168.0.10-200 | NAT |192.168.0.1 | | | | | || | |---Windows NT 4.0 Server| |-|---PP2P Installed---| || 192.168.0.2 Need up to 6 VPN connections to the NT Server Is this possible. Or is there just a better way to go about this. They don't have any money for Cisco or Hard Firewall. At first I was going to use 2.2 Kernel because I read if you recompile the kernel and install ipfwd you can GRE multi connections across, but now I just read that 2.4 hasn't been configured to allow multi connects across, but the date of the article is old. Oh how confusing this is. Cheers -Dave Nice work on the ascii diagram! :-) Personally, I wouldn't use GRE for a VPN for the reasons I stated on my previous response. I'd say you have 2 solutions worth considering: Um, OK..When was the last time you setup PPTP through a firewall. It doesn't exactly work like just forward 1723 thorugh your firewall and your all set. PPTP uses GRE protocol number 47. IPSec uses protocol 50. Does ipchains support 'forwarding' this through the firewall? Nay, I say...been there done that. If you are using a 2.2.x kernel and you are running ipchains, the ONLY way to make this work is: 1: If you have more than one address on the public net, and you aren't doing policy based routing or 'true NAT' then the ONLY ip you can use for the forwarded PPTP connections and GRE is the one that is actually being masqueraded for the internal machines (e.g. internal machines internet connection appear to come from address X on the remote host, then you had better do your port forwarding through ip X and not any other or it won't work). 2: Download and install John Hardin's VPN kernel patch. Patch your kernel source and follow the directions on building your kernel in the linux VPN how-to. 3: Now that you have setup tcp port 1723 to forward to the internal PPTP server AND you have built a custom kernel with support for pptp masquerading, you download and install the latest version of ipfwd which is a small utiltiy used only to masquerade protocols unknown to ipchains (namely in this case protcol numbers 47 and 50). Also set this up according to the VPN howto (e.g. bash # ipfwd --masq 192.168.0.99 47 ). After that is done, you might try it and see if you can get a connection. The fine point that the VPN howto leaves out is that the reply traffic from the internal vpn box must come from the same IP as the external address you are connecting to. Also, the fine points of actually doing this wih a linux 2.2.x kernel is what is kind of the pain. The message I'm responding to seems to do the same thing; glosses over all issues in doing this with a 2.2.x kernel. vec 1. Use PPTP. This works and the encryption is adequate, unless anyone with some resources thinks you're hiding something special :-) You'll need to open port 1723 on your firewall for the incoming connections. Just don't use MS-CHAPv2 for auth here, there's a well-known vulnerability with it. So, the remote clients will run the native PPTP to connect directly to the NT server, having turned on the PPTP support to recieve connections. Oh, and you'll actually need to port forward through the firewall since you have NAT. Or, you could run the Linux pptpd server on the firewall to terminate the connections there and just route normally on the private side. I like the later. 2. Use Freeswan. This is an IPSec solution that is much more secure, but more complex too. There are interworking issues between Linux and Windows but there is some good information on that on the web (a google search for freeswan windows will turn up some of them). So, in this scenario you'd run the Freeswan server on the Firewall and terminate the connections there. Be sure to allow port 500 to connect to the firewall, and I think protocol ESP as well. That's what IPSec runs over. You'll find everything you'll need on the Freeswan website, and you'll need to patch the kernel too. That's all I can think of at the moment. Having only time
Re: portfw to multiple machines, same port
The best source of examples that worked as a sweet starter template for me can be found at: http://www.ecst.csuchico.edu/~dranch/LINUX/TrinityOS/cHTML/TrinityOS-c.html If you are running ipchains, it's a killer place to look. I plan to check it out again when it has iptables support in it to see if he has anything new. vec - Original Message - From: Xeno Campanoli [EMAIL PROTECTED] To: Michael P. Soulier [EMAIL PROTECTED] Cc: debian-user@lists.debian.org Sent: Saturday, March 02, 2002 8:59 PM Subject: Re: portfw to multiple machines, same port Michael P. Soulier wrote: From IPMASQADM(8): EXAMPLES Redirect all web traffic to internals hostA and hostB, where hostB will serve 2 times hostA connections. Forward rules already masq internal hosts to outside (typical). ipchains -I input -p tcp -y -d yours.com/32 80 -m 1 ipmasqadm mfw -I -m 1 -r hostA 80 -p 10 ipmasqadm mfw -I -m 1 -r hostB 80 -p 20 Do I still need to set up ipchains for packets coming back out, or does this take care of all of it? Another thing I'm similarly stuck on is portforwarding into a single FTP server. Do you just: ipmasqadm portfw -a -P tcp -L $external_ip 20 -R $DMZFTP_IP 20 ipmasqadm portfw -a -P tcp -L $external_ip 21 -R $DMZFTP_IP 21 or do I also need to put in some ipchains stuff defining the exiting packets? Also, can I use both portfw and mfw in a configuration, for instance mfw with the web servers and portfw with the ftp server? TIA Mike On 01/03/02 Xeno Campanoli did speaketh: As near as I can tell from the documentation I've read so far, you can't (in 2.2.x) ipmasqadm portfw a port to multiple servers of the same port. For instance if I want to go from the ip address on my cable connection to four separate webservers, say one an apache, one a boa, a dhttpd and a roxen, all of which have their own separate purposes, I just can't do this it looks like without getting multiple external ip addresses using portfw. It also looks like I in fact might be able to do this with mfw, which is apparently not recommended. Anyhow, I'm stretching beyond my ability here anyway for now. The one answer that does seem to be reasonable is to specify 80 for a front end webserver and then access the other webservers on other ports, so that the apache could be 81, the roxen 82, the boa 83. Is this fairly typical? I'm not keen on playing too radically, at least not this season. TIA for any feedback. Sincerely, Xeno -- http://www.eskimo.com/~xeno [EMAIL PROTECTED] Physically I'm at: 5101 N. 45th St., Tacoma, WA, 98407-3717, U.S.A. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- Michael P. Soulier [EMAIL PROTECTED], GnuPG pub key: 5BC8BE08 ...the word HACK is used as a verb to indicate a massive amount of nerd-like effort. -Harley Hahn, A Student's Guide to Unix Part 1.2Type: application/pgp-signature -- http://www.eskimo.com/~xeno [EMAIL PROTECTED] Physically I'm at: 5101 N. 45th St., Tacoma, WA, 98407-3717, U.S.A. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Help!!! undelete for ext3fs!!!
I did this, once but it was a pain. It was under ext2 instead of ext3 but that shouldn't matter in this case. It also required knowledge of some of the contents of the file and wasn't very useful for binary files. You essentially boot without mounting the filesystem on which you want to 'undelete' (or you unmount after boot if not the root partition and after killing all daemons that use the target filesystem). Then you grep the device (eg /dev/sdb1) for some part of the file you want and have it spew so many bytes before and after what you are looking for so that you can see where the start/stop of the file is. Once you have that you use a comand (dd? or something) to copy those bytes from that location on the device to a file on an already mounted file system and poof, you have it back. I know this is rather vague but you get the idea of the approach. I wish I could give you more detail but I did this around 4 years ago and haven't really needed to do much like it since. vec - Original Message - From: Sebastiaan [EMAIL PROTECTED] To: Cheryl Homiak [EMAIL PROTECTED] Cc: debian-user@lists.debian.org Sent: Friday, March 01, 2002 12:19 AM Subject: Re: Help!!! undelete for ext3fs!!! High, On Thu, 28 Feb 2002, Cheryl Homiak wrote: I just deleted something I didn't want to delete; won't hurt my system, just destroyed some important records I was keeping. Is there any way to undelete in ext3fs? Ouch, no idea. And if there is a way, but you had to have it pre-set up before the catastrophe occurred, I'd still like to know about it so I will have a safeguard in the future. You can setup some kind of Trashcan yourself. Make a directory on your system somewhere and make an alias for rm: alias rm = 'mv ? /tmp/Trashcan' not sure about the ? which should be the filename. Or write a script for this. If you are on a multiuser system, you definitely have to write a script, which moves the deleted files into subdirs in the Trashcan which are unreadable by other users. Greetz, Sebastiaan -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apropos does nothing
Perhaps you should read the entire thread (ironically) before replying to the initial message. Besides it doesn't change the fact that I have always found it to be a waste of a command and was merely stating it. vec - Original Message - From: Colin Watson [EMAIL PROTECTED] To: debian-user@lists.debian.org Sent: Tuesday, January 22, 2002 5:05 AM Subject: Re: apropos does nothing On Mon, Jan 21, 2002 at 09:05:26PM -0700, Vector wrote: It is a waste of a command in the first place. Perhaps you should read 'man apropos' (ironically) before dispensing wrong advice. -- Colin Watson [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apropos does nothing
It is a waste of a command in the first place. Use which, locate, and find instead. If locate gives you nothing run updatedb to build the file location database, then put in cron every night and you're all set. vec - Original Message - From: Jeffrey W. Baker [EMAIL PROTECTED] To: debian-user@lists.debian.org Sent: Monday, January 21, 2002 8:29 PM Subject: apropos does nothing Somehow I have an unstable installation where apropos doesn't do a damn thing: [EMAIL PROTECTED]:/var$ apropos ls ls: nothing appropriate. [EMAIL PROTECTED]:/var$ apropos apropos apropos: nothing appropriate. [EMAIL PROTECTED]:/var$ ls /usr/share/man/man1/ls.1.gz /usr/share/man/man1/ls.1.gz [EMAIL PROTECTED]:/var$ mkwhatis bash: mkwhatis: command not found How do I get it to work? -jwb -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apropos does nothing
man -k pthread heheh - Original Message - From: Jeffrey W. Baker [EMAIL PROTECTED] To: debian-user@lists.debian.org Sent: Monday, January 21, 2002 9:28 PM Subject: Re: apropos does nothing On Mon, 2002-01-21 at 20:05, Vector wrote: It is a waste of a command in the first place. Use which, locate, and find instead. If locate gives you nothing run updatedb to build the file location database, then put in cron every night and you're all set. Okay fine. But on slackware I can do 'man -K pthread' and get every manual page on the whole system that mentions pthread. What is the equivalent funtion on Debian? And please dont say: for page in `find /usr/share/man`; do zgrep whatever $page; done; 'cuz that just ain't what I'm looking for. -jwb -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: BUSINESS PROPOSAL
Somebody missed the joke...heheh vec - Original Message - From: Eric G. Miller egm2@jps.net To: debian-user@lists.debian.org Sent: Wednesday, November 14, 2001 1:50 AM Subject: Re: BUSINESS PROPOSAL On Tue, 13 Nov 2001 23:02:37 + benfoley [EMAIL PROTECTED] wrote: just send me a bag of whatever you're smoking. This is a well known scam. -- Eric G. Miller egm2@jps.net -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Tomcat ?? Where to get it
it is part of the apache project and can be obtained at: http://jakarta.apache.org vector - Original Message - From: Shane Broomhall [EMAIL PROTECTED] To: Debian-User (E-mail) debian-user@lists.debian.org; Suse-Linux-E (E-mail) suse-linux-e@suse.com; Linux-Users (E-mail) linux-users@linux.nf Sent: Saturday, September 29, 2001 9:32 PM Subject: Tomcat ?? Where to get it Hi All, I have to learn how to use a Java based web portal that will run on most web servers. To run on Linux it needs to have Java on Apache, I have been advised that it is better to use something called Tomcat. Could someone please direct me to where I can find out some more information on Tomcat. Thanks in advance. Shane Broomhall Brisbane Australia [EMAIL PROTECTED] [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
dmi and Mobo monitoring
I am running potato on a new dual AMD box with Tyan's K7 mobo. I need to monitor the temperature inside my box and send notifications (snmp trap or email or something) when it reaches a certain level. This is regular DMI/LDCM stuff under windoze and I am having great difficulty finding something on this for linux. Can anyone out there give me some pointers?? Does there need to be support for chipset in the kernel for this to work? Thanks, vec
Re: BIND9 on Debian
Which version of bind9? Is it the beta? I'm running bind 9.1.3 on potato and I haven't had any problems yet. It seems to be serving all the domains I have on it (about 100) and it reloads for me every time with notifications. vec - Original Message - From: Doug Fields [EMAIL PROTECTED] To: debian-user@lists.debian.org Sent: Friday, September 21, 2001 9:21 AM Subject: BIND9 on Debian Hello, I'm using BIND9 on Debian Woody with Kernel 2.2.19 with several patches (primarily freeswan). I'm finding that it doesn't work reliably in the following circumstances: 1) It doesn't seem to do a proper reload of zone files with rndc reload 2) After several rndc commands, the server can't be shut down (even with /etc/init.d/bind9 stop); it must be killall -9'd. 3) It doesn't send out NOTIFY messages with an rndc reload - but it does seem to when it starts up The reason I'm using BIND9 is to take advantage of the views feature to allow my internal names visible behind my firewall, and the external ones (a subset of the internal ones) visible outside the firewall. I would appreciate any suggestions. Thanks, Doug -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: BIND9 on Debian
I too am running smp with a 2.2.19 kernel. The biggest difference is that I'm not using the package. I built it from source. I'm specifying -n 2 on the command line as well so that it uses both CPUs and in the log it says it's using 2 cpus. So I would probably say try building it from source like I did and see if that helps. vec On Fri, 21 Sep 2001, Doug Fields wrote: Hello, I'm sorry; it's the 9.1.3 packaged up in Debian/Woody distribution. I've noticed the bug list (as suggested by Sean) mentions problems with threading and/or multiprocessors, but Vector's response (below) indicates it runs stable for him, although I am running woody kernel-source-2.2.19 on SMP boxes. Perhaps I will try another kernel without FreeS/WAN, but that would not be a very tenable position in the short term for me, and the other patches I can't imagine affecting it (they are the 2.2.20pre-10 3ware driver, the latest Adaptec 7xxx driver, and the latest eepro100 driver). Thanks for any additional thoughts, Doug At 11:24 AM 9/21/2001, Vector wrote: Which version of bind9? Is it the beta? I'm running bind 9.1.3 on potato and I haven't had any problems yet. It seems to be serving all the domains I have on it (about 100) and it reloads for me every time with notifications. vec - Original Message - From: Doug Fields [EMAIL PROTECTED] To: debian-user@lists.debian.org Sent: Friday, September 21, 2001 9:21 AM Subject: BIND9 on Debian Hello, I'm using BIND9 on Debian Woody with Kernel 2.2.19 with several patches (primarily freeswan). I'm finding that it doesn't work reliably in the following circumstances: 1) It doesn't seem to do a proper reload of zone files with rndc reload 2) After several rndc commands, the server can't be shut down (even with /etc/init.d/bind9 stop); it must be killall -9'd. 3) It doesn't send out NOTIFY messages with an rndc reload - but it does seem to when it starts up The reason I'm using BIND9 is to take advantage of the views feature to allow my internal names visible behind my firewall, and the external ones (a subset of the internal ones) visible outside the firewall. I would appreciate any suggestions. Thanks, Doug -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Unidentified subject!
Did you read the how-to mentioned in the last reply yet? vetor Quoting Jenner Almanzar [EMAIL PROTECTED]: You're right, what i'm trying to ping out if the default gateway. How can i configure the eth0? jenner -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Unidentified subject!
Did you read the how-to mentioned in the last reply yet? vector Quoting Jenner Almanzar [EMAIL PROTECTED]: You're right, what i'm trying to ping out if the default gateway. How can i configure the eth0? jenner -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Using different DNS per ISP
You could just find a nameserver out there somewhere to use that doesn't restrict lookups in the way your ISP's do (I'm not sure why they're doing that anyway.) There are plenty of them out there and they would probably work just as well (obviously the closer/faster the traffic to the DNS server you use the better performance you will see.) vector On Tue, 26 Jun 2001, Marcus wrote: I'm using two different ISPs depending on time of day, and cost. Problem is that they will only accept their own nameserver under /etc/resolv.conf. Is there a way around this, or do I need a script which changes resolv.conf before dialing? Thanks, Marcus
Re: net mask
Or to make things clearer, since he's asking in the first place he probably doesn't know what /29 means...so ie: 255.255.255.248 vector On Mon, 25 Jun 2001, Alson van der Meulen wrote: I guess he meant 213.201.43.208-213.201.43.215 then it would be [EMAIL PROTECTED]:~$ netmask 213.201.43.208:213.201.43.215 213.201.43.208/29 On Mon, Jun 25, 2001 at 10:30:36PM +0200, [EMAIL PROTECTED] wrote: hi, the address your supplied is a C-class address. This needs netmaks 255.255.255.0 However, the -15 you put behind it is unclear to me, I never have seen that after a ip-number. Anyways: A-class 1.x.x.x-127.x.x.x netmaks 255.0.0.0 B-class 128.x.x.x-191.x.x.x 55.255.0.0 C-class 192.x.x.x-223.x.x.x 255.255.255.0 The remaining part os D-class and used for multicasting apart from the 255.x.x.x which is tbe broadcast range. Netmasks geiven her are ofcourse the most general options and can be different if a ip-blok is subnetted. Andor 191.x.x.x 55.255.0.0 C-class 192.x.x.x-223.x.x.x 255.255.255.0 The remaining part os D-class and used for multicasting apart from the 255.x.x.x which is tbe broadcast range. Netmasks geiven her are ofcourse the most general options and can be different if a ip-blok is subnetted.
Re: locale error on potato
I have the same problem with locale on a perl module I was trying to install. I posted to this list twice, no response. Someone else posted about a local problem and someone replied with the same 'locale-gen' response. locale-gen can not be found on my potato system, so if find the answer please share! vector On Sun, 24 Jun 2001, Jimmy Richards wrote: On Sun, Jun 24, 2001 at 07:29:40PM +0200, Stefan Bellon wrote: Hi! Since I switched back from sid to potato and installed the Ximian Gnome Desktop I often get the following warning message: perl: warning: Setting locale failed. perl: warning: Please check that your locale settings: LANGUAGE = (unset), LC_ALL = (unset), LANG = en.ISO8859-1 are supported and installed on your system. perl: warning: Falling back to the standard locale (C). The locale setting on potato seems to be different than that one sid. So, how do I correct this? TIA. Greetings, Stefan. Hi, I don't know if potato has this command or not, but you might try running 'locale-gen' and see if that fixes it. HTH, Jimmy Richards -- Stefan Bellon * mailto:[EMAIL PROTECTED] * http://www.sbellon.de/ PGP 2.6 and GnuPG (OpenPGP) keys available from my home page 99% of accidents occur at home ... stay safe, go out lots! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Hardware
http://www.tomshardware.com vector - Original Message - From: Antonio Alberto Lobato [EMAIL PROTECTED] To: debian-user@lists.debian.org Sent: Thursday, June 21, 2001 7:31 PM Subject: Hardware Hi all ! Where do I find good and free downloads (or on line) of books, manuals, tutorials or guides about hardware ? In English or Portuguese. I think it would help me in my linux learning. Thanks Tom -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Dual processor PIII 866 versus single Pentium 1.4Ghz
For multiprocessor support, intel is not as horrid as you proclaim it to be. No question, AMD rocks. But there are not many (any at all?) dual cpu boards out there and the support for more than one processor with AMD is still quite immature. For mid-range and up servers, multiple CPU's can be a big win and intell is still the best choice for that type of solution. I'm quite pleased that AMD has pulled ahead with a faster bus I anxiously await a solid dual AMD cpu solution. As soon as comes along I will no doubt convert all my multiprocessor intell systems to AMD. vector On Wed, 20 Jun 2001, John wrote: Ill tell you right now. DONT GET INTEL CPUS. they are horrid now-a-days. AMD chips are based off of newer architecture, have 33% more on-chip chache, and can run at even 266mhz system bus, as opposed to only 133mhz p3. Athlons also have 3 Floating point pipelines, as opposed to p3's 1. all of that, and AMD chips are WAY cheaper. On Wednesday 20 June 2001 06:29, [EMAIL PROTECTED] wrote: This is not a debian specific question, but I know there are some hardware experts out there that may be able to help me and I would much appreciate it :-) I am considering upgrading our server at work. I don't fully appreciate the performance advantage / disadvantages of dual processors and SCSI hardisks. Here is what I have been recommended to get: 2x Intel Pentium III 866 Mhz Intel TUPELO (STL2) Motherboard Dual Processor capable Adaptec AIC-7899 dual channel SCSI controller 2x 512 Mb Intel Certified ECC Registered Memory 2x Cheetah Seagate 36Gb LVD Hard Disks + network cards, floppy, CDROM, etc. Basically, my question is: What would the difference in performance be between this configuration versus say a single P4 1.4Ghz? This server will unfortunately have to run Windows NT (as the proprietary software requires it). It is basically a database server with 8x 486s acting as essentially dumb terminals. They will run a basic version of the proprietary software and all the processing will be done by the server. 1. What do people think about whether it is worth spending the extra money for dual processors? Does Win NT fully utilise dual processors? It is nearly $1500 (Aus) for the motherboard! But a P4 1.4 GHz is much the same price (I think), but the mother board would be cheaper then. 2. Is it worth spending the extra money on a SCSI controller and hard disks? Here is what the specs say on the SCSI HD: Formatted Capacity: 36.4GB Interface: 68-pin Ultra2/SCSI Data Transfer Rates: 160MB/s Average Seek Times: 5.4ms Buffer Size: 16MB Rotational Speed: 1rpm Height (inch/mm): LP (1.0/25.4) Here is the specs on a 40Gb Western Digital IDE 7200 RPM: WD CaviarTM 40GB EIDE Hard Drive WD400BB Transfer mode: 100 MB/s Average Read seek: 8.9ms So the SCSI spins faster and has lower average access times. I suppose this means that it would be of benefit when you are talking about a large database server with multiple terminals connected. It is just hard to work all this out from the specs. Can anyone speak from experience on these issues? Especially regarding the dual processor versus single but faster processor. Thanks for help. Mark.
Re: Dual processor PIII 866 versus single Pentium 1.4Ghz
Sweet! I'll check it out and let you know what it's like... I was once a big fan of Tyan boards but had to get away from them for a while because they weren't releasing boards that were meeting my needs. Thanks for the heads up! vector Tyan has released a dual Athlon motherboard. According to Pricewatch you can get the board with *two* 1.2 GHz Athlon MP CPU's for just over 1000USD. The motherboard has onboard Ultra 160 SCSI (Adaptec AIC-7899W) controller and dual onboard Ethernet controllers (3Com 3C920's). Full specs at http://www.tyan.com/products/html/thunderk7.html I'm saving up. :)
Re: COM21 is killing me with ARP
Thank you Bryan, I couldn't have said it better. Besides, it is not clear from the trace provided that all ARPs are coming from the gateway anyway. vector - Original Message - From: Bryan Andersen [EMAIL PROTECTED] To: Patrick Colbeck [EMAIL PROTECTED]; Sebastiaan [EMAIL PROTECTED]; Vector [EMAIL PROTECTED]; debian-user@lists.debian.org Sent: Thursday, June 14, 2001 2:53 AM Subject: Re: COM21 is killing me with ARP Patrick Colbeck wrote: It really doesnt matter that there is a whole class B address space as you should only get arped when someone om the same class B needs to know your mac address. Once the arping device has your mac address it should cache it so it doesn't have to arp for it again for a long time. All the other people on the calls B shouldn't be trying to find your MAC address as they theoretically should only be talking to your service providers DSLAM. Actually it does matter. When Joe user turns off their box it nolonger can answer requests for it's ethernet adderess. This means a bunch of requests for it's arp address. So when someone scans the network you get bombarded by arp requests, and the caches naturally gets trashed durring this. If you know a provider does this you can realy hose up their network by bombarding them with random addresses in their network space. To keep from having this trash a network the router really needs to have enough cache entries to store all hosts on the network. Many routers just can't handle that for a class B network. They really should break their network up into class Cs. It sounds like somebody has screwed up at the service provider configuring their routers they have probably:- i) Configured a really small arp cache timeout value so the service provider router is permanatly having to re arp for the mac addresses of all the DSL modems or ii) Configured a static route via a broadcast interface (eg etherent) on the cental router. This is a really bad thing as instead of just arping for the next hop address the router will arp every time it needs to send a packet to any address on the network the route is for to try and determine the gateway to that address. This is a really good way to crucify network performance , static routes pointing at interfaces rather than next hop addresses should only be used on point t point networks (leased line etc). -- | Bryan Andersen | [EMAIL PROTECTED] | http://www.nerdvest.com | | Buzzwords are like annoying little flies that deserve to be swatted. | | -Bryan Andersen|
Re: COM21 is killing me with ARP
The fact that your provider is using an etire class B address space for a single broadcast network is what should be making you nervous. In most network architecture schemes, the ip's are divided into blocks and thus the amount of broadcasts being received by individual hosts are greatly reduced. I would want to see more of the dump to actually determine if this is really a problem. ARP's are used to gain the MAC address of a system with an IP that is on the same network as you are. With a class B address space of potentially 65,533 hosts on the same broadcast network you can expect to see *MANY* ARP requests! Is this ATT, sounds like something they would do... vector On Wed, 13 Jun 2001, Sebastiaan wrote: On Wed, 13 Jun 2001, Angus D Madden wrote: On Wed, Jun 13, 2001 at 12:01:23PM +0200, Sebastiaan wrote: a couple of days ago I installed a COM21 cable modem. Although I can internet without problems, the modem itself is sending me endless ARP requests, while my computer does not answer them. I analysed the data with I have a COM21 and I have the same problem. You'll notice that the TX light on your NIC will never stop. It never used to be a major problem with my ne2k-pci NIC (until the NIC got toasted for one reason or another). After that I switched to a 3c905*, which seemed to work great but would go dead after about 15 minutes (presumably because of the arp bombardment). I hope it did not get toated because of the COM21: I only like fried chips. I have no trouble with the interface (MACE, PowerMac) and the link is pretty stable until now. Thanks for the info, Sebastiaan My solution was to reset with interface with a cron job every 15 minutes. It's a total rat-fsck solution, but it works. This makes me nervous. 212.127.*.* is my ISP cable modem network. What is this, can I stop it? I'm not sure it comes from one source. If I tcp dump I see many 'arp who has xxx.xxx.xxx.xxx tell xxx.xxx.xxx.xxx' and the ip addresses jump around all over the place. Not sure how to stop it. g
