Re: [expert] DNS cache to disk (Is it possible?)
On Fri, 21 Apr 2000, you wrote: > Hi > > I have installed bind/named on my system and configured it to act as > cache-only. However, I have seen that the information is stored IN MEMORY. Ok, > this is not good news for me because I have a dial-up account and the servers > of my ISP seem to be quite slow (even slower than my local DNS). Do you know of > any way to make named store DNS information on disk? That would be a real DNS > cache. Maybe I should use 3rd. party programs. > > Moreover, when named is started, everything is fine, but... > > named[365]: listening on [127.0.0.1].53 (lo) > named[365]: Forwarding source address is [0.0.0.0].1024 > named: named startup succeeded > named[377]: Ready to answer queries. > named[377]: sysquery: sendto([128.8.10.90].53): Network is unreachable > > What does this last line mean? What the hell is that "sysquery"? > > Thanks! > > -- > Joan Ensesa > [EMAIL PROTECTED] You can use Squid to proxy for named and tell it to cache to disk. The simplest way to configure this (in terms of effort) is to download webmin-0.78-1mdk.noarch.rpm from a cooker mirror and then use netscape to open http://127.0.0.1:1 with login name root and root password for password, then click the "servers" tab and configure squid... Perhaps there is an option under BIND 8 which may also be configured from that same screen. Civileme
Re: [expert] Apartment network wishlist
On Sat, 26 Feb 2000, you wrote: > Hello! > > I have a question on a higher level, not so much about mandrake in particular. Sorry >if it is too "newbie". > > I want to build a little network in my apartment, and I would like some pointers as >to how. > > NOW: > I have two computers, going on three. > > 1) A windows/mandrake workstation, OC:d, so I don't want this to run all the time > 2) A linux server, rather old and slow, but fully capable of SSH and serving files > 3) An old compaq (486/33) with should be a terminal. > > I'm on a 10Mbit student housing network, and this is rather choked(tm). I have two >IPs, so communication over this using my hub is possible, albeit slow. I'd like some >extra bandwidth of my own, since I want to use the fileserver for my windows files >and I want to tunnel some communication over ssh without doubling my load on the >external network. > > I WANT: > 1) The linux server to be the gate to the external net, posing as >web/ftp/file/ssh-server. > 2) The windows machine to be able to use ICQ etc. with it's own IP. > > How do I do this? I probably need another NIC, but that's ok. But will the passing >of traffic through the linux machine pose any problems, like ICQ not beeing able to >connect (I've seen this happen when behind firewalls). > > /Adam Skogman Using the ip masquerade you can accomplish this easily. Here's how The linux server needs two NICs say eth0 connected to your local hub and eth1 to ONE IP. You will not need the other. Do not make any other external connections or you can be easily exploited. OK here is the code for the minimal setup You can set up a script file, type it in manually or put it in /etc/rc.d/rc.local ipchains -P forward DENY ipchains -A forward -i eth1 -j MASQ echo 1 > /proc/sys/net/ipv4/ip_forward ICQ, licq, everybuddy, gAIM, all work Netscape works... kfm as an ftp works beautifully. FTP requires "passive mode" The port commands otherwise point to the wrong places. If you are a fan of FTP command lines, use them from the linux server toward the outside. www.linuxdoc.org has the ipchains-HOWTO available in a variety of formats. You might also find it at /usr/doc/ipchains-x.y.z on your installation, but I am told not to bother to learn it because it will be replaced in the imminent kernel 2.4 release. I use it all the time(the configuration of a masquerading server, and ipchains). It is a wonderful way of closing ports, rejecting certain IPs, port-forwarding things you want other servers to handle But most important your internet IP cannot be used to reach you unless you want to be reached. Naturally ipchains and a masquerading server and all the known virus and exploit filters cannot make a winbox behind the masquerading firewall secure. Perhaps secure from winnuke, but some bonehead will still find a hostile site with MSIE or get an Excel spreadsheet in email from a trusted source and end up toast. The takeover from Netmeeting won't work Aggghhh. Some of my users have windows in dual boot situations but the NetBIOS ports are dead locally (yep, no Samba) and onto the internet, so their windows are without network services locally and without support as well (and for that matter, their Network Neighborhoods aren't loaded with the info for the system) So if they want to get beyond their boxes, they need to use the friendly linux installations on their desktops. Civileme
[expert] 2 AIR Experiences
1. FIC PAG 2130 with AMD K6-2 and 128M RAM Built-in Cyberblade (Trident) video and "SB-compatible:" sound on the SouthBridge VIA MVP4 Chipset GRaphical install worked. Sound reported "device or resource busy" Tried three more installs with different BIOS settings for enabling/disabling sound. Tried Lother, which also saw the sound, but could not make it audible. Video was twitchy. 640x480 at 16 bit depth was all the install would do. After playing with the graphical configurator for a while, I managed to get a flickering 1024x768 at 32 bit. Strangely enough, Xconfigurator on LM 6.1 working with this board found a non-flickering 1024x768 in 32 bit depth the first time. 2. S3 Trio 64V+ with 2 Meg... Vanilla sound card (ESS1688) Cyrix 686MX PR-200 processor on an Antec Motherboard using the 430 TX intel chipset. Stayed in text install. Froze "in second stage install" and this was from a boot off the CD-ROM. CD was a vanilla 24X and HDD was WD24300. Only other hardware present was an RTL8029 NIC. This one was running SuSE 6.1. 2a. Trident 9440 video substituted Came up in graphics this time, but froze "in second stage install" again. 2b. Switched to a Cyrix 686-PR200+ Processor and dropped the bus speed to 66, making it essentially a 686-pr166+ Graphical install started working. This is probably an artifact of the TX chipset overclocking the PCI bus when it is pushed past 66MHz (but why did SuSE work?). I have had more difficulty with Cyrix Processors under Mandrake than I can properly deal with. I still have two machines that won't update (Cyrix GXi-S types) because, while Mandrake will install, the update packages claim "this is for a different architecture". Moreover, in the linuxmall ads, it is posted that there are issues with Mandrake and Cyrix processors and that while a patch exists, it was never made available on the Mandrake sites. I have machines running 686MX-PR200 processors running L-M 6.1 using terrible chipsets, but not the Intel 430TX, which seems to be in a category by itself. Strangely, the TX Chipset can be used to advantage at 75MHz bus speed with a K6-233 running at 3x bus speed, and LM 6.1 runs there as well, but it fails with a 686MX running at 2x 75MHz on the 430TX. I am beginning to wonder if some instructions do a "not quite back to back" bus locking that evokes a version of the 686 coma bug which shows up only at the higher clock speed with the timing of the intel 430 TX chipset for the bus. Anyone have any similar experiences to share? Civileme -
Re: [expert] 128 mb mem
Adrian Saidac wrote: > I agree with all of you flame or not. > I really need an answer not a status of other systems. > Given the fact that Red Hat/Mandrake is keeping a long silence about > this make me believe that there is a problem somewhere. Why is showing > only on certain systems - THAT'S the mystery!! > Well, I don't and won't use ABIT boards, so it is unlikely I will be able to reproduce your error EXCEPT: I set this system for OS/2 My 119M shrunk to 63 I activated the memory hole at 15-16M and my memory magically shrunk to 14M! Look at your BIOS. Hit DEL when it is starting up, check the menus out for these rather standard settings, OR add append mem="M" to /etc/lilo.conf and as root # /sbin/lilo # shutdown -r now And if you cannot understand that it is NOT safe for an OS to test the memory directly, then you probably have never heard of memory-mapped I/O or shadowing, nor do you realize that linux runs on systems that routinely do those things consider that m68klinux may not be able to utilize the same binaries, but it can deal with the same source, and so can linuxppc, alpha linux, etc. Your conjecture that linux cannot handle more that 64M is ... perhaps a leap made without extensive knowledge. And yes I have a reason I don't use ABIT boards. I have never implemented one in linux without a hassle. I decided I would rather not live with the choices made by the ABIT design engineers, which impress me as one kludge on top of another. YMMV, of course. Civileme
Re: [expert] More Printer Woes!
WH Bouterse wrote: > I am presently using L-M 6.1 Helios > Thanks for the response! > > My printtool config is as follows: > Name lp > Spool Directory /var/spool/lpd/lp > Printer Device /dev/lp0 > Input Filter DeskJet 850 > > The ASCHII "directly to port" test prints the proper line > The other tests go to the que and wait > > #'lpr filename' ques and does nothing > #'lpr -Plp' filename , the same > > #'lpc stat' gives > lp: > queuing is enabled > printing is disabled > no entries > no daemon present > > Even when 'lpd restart' is done and even rebooting the machine > 'lpc stat' shows "no daemon present" and "printing disabled" > > A least I have seen where the problem "may" be but don't know what to do > next? > > Thanks for any help or ideas! > > William Bouterse > Juneau Alaska FIrst, change the print device to the HP 550c, ... 6xxc series Then see if you have the same problem Lots of 850s and 870s were shipped which will not respond to the language they are spec'd for. Civileme Same problem here in Kotzebue. That was the fix.
Re: [expert] Frontpage 4.0 Extensions.
I have stayed out of this until now, but I have to vent. Why are we trying to put a failed system on a good one? http://www.insecure.org/sploits/Microsoft.frontpage.insecurities.html Posted April 23 1998. I test them occasionally. They still seem to work rather well. Too well in fact. Why? Because FrontPage is a Brain-damaged proggie that insists on su type access to your site. Isn't that just lovely? And FP extensions give it to them! Don't get radical about many things, but here I am sending mail out of my root account because I didn't want to wait to get to a regular user account. FrontPage extensions, brrr. But if you are going to try to use the garbage like FrontPage from the other end, at least go to linuxberg and get Demoroniser. It is 8.5K and does a wonderful job of cleaning up Microsoft errors and HTML incompatibilities. I am not disparaging your efforts, mind you. I am asking you to re-think your PURPOSE. Civileme Civileme Herve Amoussou wrote: > Kirk Whiting a écrit : > > > I have recently installed the frontpage extensions on a test machine. Is there > > a good how-to yet? I am having trouble getting the admin to setup new accounts. > > It seems to eat all the processor and goes nowhere. > > > > any help would be appreciated. > > > > Kirk > > The FP extensions are troublesome with Apache. On windows 98 i have a FP web site > powered by Apache win32 that works very well. u just have to use a single main > page in FP with HTML links only.Currently trying same config on LM60
Re:[expert]Suggestion for Mandrake
Ummm, well I don't want to rain on anyone's parade but it should be known that at least one pentium class processor is mis-recognized and about half the updates will not install because "package is for a different architecture". I refer to the Cyrix MediaGXi-S Here are all the results I have P-100OK IBM/Cyrix 586waiting for report rise mP6no processor available 6x86 OK k6-2OK CeleronOK AMD 586Test this week P-IIOK Cyrix MediaGXMisrecognized by update packages initial install OK Cyrix MediaGXM No Processor available IDT C6OK AthlonNo processor available 6x86/M-IIOK IDT C6-2No processor available Anyone who wants to use these test results is free to do so, of course. These are actual installations of Venus. I have NOT tried a 486, but someone should just to complete results. Civileme
Re: [expert] Boot problems
Ummm, That's THREE. Seagate drive plus L-M 6.0 plus somekinda BIOS Shuttle HOT591 with VIA Chipset and Seagate 38641 Locked up on install... Worked fine With IBM IDE drives Used K6-2 300 Worked fine with Seagate and Festen J-Mark J-5TXC (Intel TX Chipset)with L-M 6.0 CD and Seagate ST38641 and Seagate Looked good on install Used 6x86-200 MX Got running on the disk. Logout from K Desktop produced Kernel Panic on reboot when it tried to remount any partition on the Seagate 38641 That same Seagate worked fine in an install and is still cooking along like a champ on a Shuttle HOT 599 (SiS 530 chipset) And now yours/ with I would guess a ST2850. I would suggest you try a new install without the Seagate. I think we might be triggering some undocumented diagnostic microcode on the drive electronics. Otherwise, I, too, am totally mystified. Also, the BIOS most likely plays a role in whatever is happening. Civileme John Aldrich wrote: > As I mentioned in my previous message, I installed Mandrake on my dual-ppro > 200mhz system (192 mb ram and about 9 gb drive space total.) > > I had a problem after I'd installed with the system hard-locking after the > second or third boot on checking /dev/hda. There is no other file system on > the computer, just ext2 and swap. I have about 50 mb swap space on /dev/hda2 > and about 75 mb swap on /dev/sdax (don't recall which partition...I *think* > I put it on /dev/sda2) > > Anyone got any suggestions? I had no problems installing. The system is > mostly running off the scsi card (Advansys Ultra SCSI ABP940-UA) and > Mandrake was installed off the CDROM just fine. The first hint of trouble > came when it locked up. > > Thanks... > John
Re: [expert] Unsupported Disk?
H, Well, I had the channels separated, but I went back and tested with the CD as IDE0 Slave, IDE1 Master, and IDE1 Slave--same results. Also tried UNDERClocking the HOT-591 to 66MHz FSB (the PCI Bus runs asynchronously to the Front Side so this showed no improvement, as expected) I looked at the SOURCE, (Usually the definitive authority) and the differences between 5.3 (which worked) and 6.0 (which didn't) were unrevealing. One item I was unable to chase down in the source I had was the function ideGetDevices(&ide) I cannot find the source for this function, just a header in scsi.h. I have the complete source on the way, but I wonder if this function was changed between the two sources, perhaps to accommodate the 'new' Ultra 33/66 ??? Anyone else looking at this? Michael Moore aka Civileme > I had probs with a Seagate ST36xxx Medalist drive: The installation would > bomb as soon as I started to install packages. Fixed by (a) moving cdrom > onto different IDE chain, so hard disk is on its own, or (b) stopped > oveclocking my Mboard. > > Michael Moore wrote: > > > > Linux Mandrake version 6.0 Power Pack > > > > Installation CD in Panasonic PD Drive > > AMD K6-2 on Shuttle HOT-591P with latest VIA BIOS patches installed > > Processor Clocked at 3 x 100 MHz. > > Jaton/Trident 875 AGP Video card > > Yamaha Opl-3SAx generic sound card > > 64 Mb SDRAM PC100. > > Seagate ST38641 HDD (Note: Has more than 1024 cylinders ) > > > > System was reliably running Windows98 and L-M 5.3 in a dual-boot > > configuration > > > > Reached Second Stage Install after booting from CD--When SERVER was > > selected, it froze > > Restart--Same point reached--When WORKSTATION was selected., it froze > > Restart--Same point Reached--CUSTOM got to the Disk Druid/fdisk/Done > > screen. Both Disk Druid and fdisk froze. > > > > Restart, now with L-M 5.3 Boot disk. Used fdisk to set up the > > partitions as I wanted them, then back to L-M 6.0 > > > > Restart--Same point reached, but now I chose "Done". Installation > > froze. > > > > Restart--Now new CD, a plain vanilla 36X proven with LM 6.0 installation > > on another system. Results unchanged. > > Restart--Now using an IBM TravelStar 3.0 Gb 2.5" HDD--worked all through > > the installation!!! > > Restart--Now with IBM DTTA 351010 10Gb HDD, also with more than 1024 > > cylinders. Worked again to perfection. > > > > anyone? ideas? flames? offers to buy a Seagate Drive for $1 US? > > > > Michael Moore
[expert] Unsupported Disk?
Linux Mandrake version 6.0 Power Pack Installation CD in Panasonic PD Drive AMD K6-2 on Shuttle HOT-591P with latest VIA BIOS patches installed Processor Clocked at 3 x 100 MHz. Jaton/Trident 875 AGP Video card Yamaha Opl-3SAx generic sound card 64 Mb SDRAM PC100. Seagate ST38641 HDD (Note: Has more than 1024 cylinders ) System was reliably running Windows98 and L-M 5.3 in a dual-boot configuration Reached Second Stage Install after booting from CD--When SERVER was selected, it froze Restart--Same point reached--When WORKSTATION was selected., it froze Restart--Same point Reached--CUSTOM got to the Disk Druid/fdisk/Done screen. Both Disk Druid and fdisk froze. Restart, now with L-M 5.3 Boot disk. Used fdisk to set up the partitions as I wanted them, then back to L-M 6.0 Restart--Same point reached, but now I chose "Done". Installation froze. Restart--Now new CD, a plain vanilla 36X proven with LM 6.0 installation on another system. Results unchanged. Restart--Now using an IBM TravelStar 3.0 Gb 2.5" HDD--worked all through the installation!!! Restart--Now with IBM DTTA 351010 10Gb HDD, also with more than 1024 cylinders. Worked again to perfection. anyone? ideas? flames? offers to buy a Seagate Drive for $1 US? Michael Moore