Re: [expert] stable kernel version ????????
--On Wednesday, October 24, 2001 2:35 PM -0300 "Alfredo C. Lopez" <[EMAIL PROTECTED]> wrote: > Which version of gcc do you have installed? > > I compile with gcc-2.96.x . > But I have egcs installed too. Check wich version returns kgcc -v gcc version 2.96-81 (Red Hat) Karl Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] stable kernel version ????????
--On Wednesday, October 24, 2001 9:16 AM -0600 David Oberbeck <[EMAIL PROTECTED]> wrote: >However, I discovered that the kernel will still not compile, > somehow the configuration header files are still fscked up. Well, to > be more precise - as you mentioned - the kernel compiles, but the > module compilation fails. > >Sorry, don't have a solution for this one, still an issue with > kernel 2.4.12-3 for me (anyone else have any suggestions?) For me the problem seemed to be in linking for only some (maybe just a few) of the modules. I had to go back to 2.4.9 to compile/link without this problem. On any given system, there's a good chance that if only a few modules have problems, they wouldn't be needed anyway. And if a module that failed is needed, its source can be downloaded and compiled separately. I didn't have a chance to try this... 'make -k modules' This will tell make to ignore errors as much as it can and go on. You will probably want to review the make output to see just what problems there were, and see if you can live with them. So use: make -k modules 2>&1 | tee file.to.capture.output then after it's done have a look at file.to.capture.output I didn't try this, but if you do I'm interested to know if it works. Karl Cunningham Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] Mount vfat partitions on 2.4.10
--On Friday, October 19, 2001 5:54 PM +0200 Jorge Giménez Mayorgas <[EMAIL PROTECTED]> wrote: > I build a clean kernel with 2.4.10 sources . Everything went ok except > for one thing . I can mount on start on even after vfat partitions . With > the kernel of lm 8.0 it works perfectly. I tried everything fat and vfat > support in the kernel , separated from the kernel as modules. Nothing > worked. On startup lm 8.0 complains about "/dev/xxx bad superblock " > when mounting local filesystems. The only one thing is to update mount > package. > Can anyone help me?. I had this same problem on 2.4.11 and 2.4.12 and found that the kernel compile of modules was bombing out before it had all the modules built, and the ones it didn't get to weren't being installed. See if you have the vfat.o module available for your newly built version. If not, then you may have this same problem. You can also look at the compile output when doing 'make modules'. If the output shows errors at the end, this is likely your problem. I went back to the 2.4.9 kernel and it built all modules just fine. Karl Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] bizarre remote access/network problem -Solved
I'm pretty sure it was one computer that was causing the problem. One of the symptoms was slow network traffic between other machines -- down to 50kbps sometimes on 100Mbps segments. Just powering up the offending computer and letting it sit at a bios screen would cause the problem. Power it down again and everything else went back to normal. It was a cheapie MB and power supply, and I never tried anything else after the ferrite core fixed it. Karl --On Friday, June 29, 2001 14:12 -0700 Scott Taylor <[EMAIL PROTECTED]> wrote: > If you have that kind of trouble with CAT-5 I would suggest getting > someone in to test the cables and find out where the noise is coming > from (it could be as simple as a kink or a mouse chewed it, etc.). I > don't think you should ever need to put a ferrite core on CAT-5, it's > already designed to reduce noise through it's twisted pair design. If > the cable is up to specification and the ends are terminated properly it > should be fine up to, at least, 50 meters. AFAIK, if you were to do > something like that on a Levington install, you would void the lifetime > warranty. > > Scott. > >
Re: [expert] bizarre remote access/network problem -Solved
I solved a similar bizarre LAN problem by threading one of the CAT-5 cables through a ferrite toroid core as many times as it would go. The symptoms were similar to yours and not consistent -- any time any hardware changed the symptoms would change. One of the machines on the network was putting noise on the line and messing up one of the switches. Putting a long cable (80') between that machine and the switch fixed it, but not wanting a mess of cables tried the ferrite core on that cable and it worked too. If you have more trouble, you might consider the ferrite core method. Karl --On Friday, June 29, 2001 2:34 PM -0500 brian <[EMAIL PROTECTED]> wrote: > In case anyone cares, and I appreciate Pierre Fortin and Scott Taylors > suggestions in troubleshooting this, but I thought i would post what the > problem actually was. > > Original problem was i couldn't ping all IP's of my work computers from > home. I thought this possibly a mandrake issue since I didn't have a > problem with any of my windows machines. But it wasn't. > > I inheritted a network to administer and have had to troubleshoot > problems on top of problems. I found first a bad port on a hub. I > replaced this with a switch. That got my main MDK box pingable. I tried > MDK 7.2, 8.0, OpenBSD, two different kinds of NICs on 3 different > machines. And nothing improved. Then I replaced two old managable > switches with new switches. That got my entire network pingable remotely > (strange that it always worked internally.) . > > The manageable switches had databases of MAC address that it could not > clear and would not find new hardware attached to any ports. Thats why > the existing Windows machines would work I suppose. > > Anyway, i learned alot and thought i would share. I could only test at > home and as the problems were intermittent I got a lot of false > positives. Maybe now I can get some sleep before the next problem > arrises ;) > > > -Brian
[expert] xinted make dependency problem?
I'm trying to compile xinetd without tcp-wrapper support. But I made the mistake of compiling it the first time with libwrap support. After that, make doesn't seem to do anything productive. running: # ./configure --with-libwrap produces the following output: [lots of discarded lines] checking whether to use libwrap... no [lots of discarded lines] # ./configure --without-libwrap produces the following output: [lots of discarded lines] checking whether to use libwrap... yes [lots of discarded lines] So it appears configure knows what I want. But when I run make, it doesn't recompile any sources or relink. The object files and executable are not rebuilt. If I remove all the object files to force it to recompile and link, all is ok. I'm new to this. Am I missing something? Thanks Karl Cunningham
Re: [expert] USB CameraMate?
There's a thread currently going on the kplug list about usb cable modems, cameras, printers, etc. The thread is called "USB Modem" Check the archives at http://www.ultraviolet.org/mail-archives/kplug.2001/ The archives aren't updated every day so you might have to wait to see it there. Or check out http://www.kernel-panic.org/ Karl Cunningham --On Saturday, June 09, 2001 12:06 + Dave Ayers <[EMAIL PROTECTED]> wrote: > I am new to the Mandrake ranks (and this list) and am running Mandrake 7.2 > on a Chembook laptop with a USB port. Said laptop bought with 7.2 > pre-installed from ASL. Before I get myself all tied in knots trying to > get a MicroTech USB CameraMate drive to work, any opinions on whether > there is any hope for this USB device with 7.2 and if so how to proceed? > > Dave Ayers > Quincy, Illinois >
Re: [expert] Is that sshd process really serving ssh?
You could have your script telnet to port 22 and check for the proper SSH response. Mine answers with: SSH-1.99-OpenSSH_2.3.0p1 Karl Cunningham This situation i'm trying to work with is like this: There is sshd >running on a box, and there is a script watching to make sure that it's >running. But if someone >logs in via ssh, another sshd process gets spawned for that connection. If >that person were to >somehow kill the sshd server process (the parent of the spawned sshd) then >s/he could continue on >with the connection but nobody else would be able to get in. The watchdog >would not know this >because it would see a sshd in the process list, which is all it is >checking for. But i would like > >it to be able to check if the sshd that it's seeing in the process list is >actually serving ssh >or if it is not serving ssh.
Re: [expert] Installation - Unsupported ethercard?
I am using a Linksys LNE100TX card, and it works fine with the tulip.o module. I'm not sure what the available modules are during install, but is there any reference to tulip.o? When you were trying to run ifup and ipconfig (did you mean ifconfig?), were you logged on as root? kc At 01:06 AM 4/11/2001 -0400, you wrote: >I have a LinkSys LNE100TX Ethernet Card. It is not on the list of cards >in the install of 7.2. It does, however, have drivers available. It is >also a descendant of the DECchip Tulip(dc21x4x) card. > >In my last install, I picked DONE for the network type when I couldn't >find the right card. I then installed the drivers, post installation, >but couldn't find programs like ifup and ipconfig. The DONE option >probably skipped the network install phase. > >This time around, I'm going to pick the DECchip Tulip(dc21x4x) card >during the install. That way, I'll get the networking packagesd >installed, then I'll update the driver with the newer code. > >Does this sound workable?
Re[2]: [expert] shell programming question -- solved
To John, Dan, Rusty -- Thanks very much for your help. In this case,
directing output to an interim file gave different results than piping it
all in one line, but trying it led me to the answer. I believe the
ambiguity is in which of the processes involved in the pipes have been
started in time for ps to catch them.
The following seems to work reliably:
temp=`ps -ef`
if [ `echo $temp | grep -c $0` -gt 1 ] ; then
echo "another is running"
else
echo "all alone"
fi
thanks again
karl
At 10:02 AM 4/9/2001 -0700, you wrote:
>Well, the first thing folks should do when trying to figure out what's
>going on in a script is capture the output and see whats going on.
>
>So, for example, one might do this:
>
>cnt=0
>while [ $cnt -lt 70 ] ; do
> ps -ef | egrep -v grep > /tmp/debug.1.$$
> grep $0 /tmp/debug.1.$$ > /tmp/debug.2.$$
> echo -n `wc -l /tmp/debug.2.$$`
> let cnt+=1
>done
>echo
>
>Then go look at the files and see that sometimes you'll catch
>things like:
>
> emacs myscript
> ./myscript
> vi myscript.info
>
>(as a dumb, contrived example)
>
>In my experience, you must be very careful when grepping for command
>lines - especially if they are scripts being executed, since all too
>often you also happen to have that script open in the editor!
>
>
>One thing I've seen is where the script saves its pid in a file,
>which you then check for. Of course, you have to watch for the
>critical section there (and there is no easy way to get rid
>of it entirely, so you just have to do the best you can) (Huh?
>What critical section??? (or, What's a critical section?) see
>below) Another option would be to make a server that would
>keep track of things like that and answer the question
>'Am I the only one of me wanting to run?' - but that seems like
>a LOT of work Easiest thing is probably just to try to make
>sure you don't run the script TOO often (10 times in a minute
>would be a bad idea ;-), and do a 'simple' check like we've been
>discussing.
>
>
>Ok, so what's a critical section? Its a section of code (usually)
>that, if you have 2 separate threads of control execute that code
>at the same time the results can (might) be erroneous. One of the standard
>examples is : static int i; crit_sec_oops(int addme){ int j; j=i + addme;
>i=j;}
>
>Now, if thread 1 comes through, calculates the sum, then thread 2 comes
>along, caluclates the sum AND manages to store it back into i, then
>thread 1 continues along and stores ITS version into i - you just lost
>thread 2's contribution to i.
>
>Trying to have a file as your locking means leaves you a critical
>section also, but we'll leave that as an exercise for the student ;-)
Re: [expert] shell programming question
Thanks for the suggestions. I'm getting inconsistent results, though. When I run the following, the most common result is 5 but once in a while there is a 2, 3, or 4. I'm not sure what's going on. cnt=0 while [ $cnt -lt 70 ] ; do echo -n `ps -ef | grep -v grep | grep $0 | wc -l` let cnt+=1 done echo I can test for the maximum value out of 20 tests and I'm sure it would get the job done. But at this point I'm curious why this happens. Any ideas? Thanks. Karl At 04:08 PM 4/8/2001 -0600, you wrote: >This works from command line... > ># if [ `ps -efw | grep -v grep | grep rerf |wc -l` -eq 0 ] ; then >echo 'yes' ; else echo 'no' ; fi > >In a shell script, make it more readable... >if [ `ps -efw | grep -v grep | grep rerf |wc -l` -eq 0 ] >then >echo 'yes' ># more statements here, etc. >else >echo 'no' >fi > >- >Also, analias command that I use *often* is > # alias psg > alias psg='ps -efw | grep -v grep | grep' > >This avoids seeing the 'grep' that is forked off by your request. > >Thanks...Dan
[expert] shell programming question
I'm writing a shell script that will have problems if another instance of itself is running, and I'd like to be able to trap that the user has started multiple instances. I've tried cnt=`ps ax | grep -c xyz` where xyz is the name of the script. It returns 3 when there is only one instance running. I think I understand why it's 3 (one for the script itself, one for it running ps and one for it running grep?). Will this always work? Is there a better way?
Re: [expert] openssh and commercial ssh (both v2) incompatible????
The keys are incompatible between the two. Openssh comes with a utility to convert from one to the other (don't remember what it's called). I use a commercial ssh client to log into an openssh server and had the same problem until I converted the keys. Karl At 08:55 AM 4/4/2001 -0700, you wrote: >Heeding the security advisories (a bit late, but heeding nonetheless), >I recently upgraded all my ssh's to the latest version. > >Then, I changed ISPs (not because I wanted to but because my ISP sold >out on us). > >SSH to my previous ISP had been working fine. > >However, I found I could not ssh in to my new isp. > >So, I tried downloading the latest commercial ssh. > >It worked fine. > >So, if you have an openssh client trying to connect >to a commercial sshd server, you MAY have problems. > >Once I (get the time to) figure out either who to report >this to, or how to fix it, I'll do either (or both ;-) > >Anyway, just thought I'd pass along the headsup...
Re: [expert] ping, ping, doesn't ping
Is either machine connected to the Internet? If so, can you ping out to somewhere on the net by IP? And try pinging from a web-based ping engine back to the machine connected to the Internet. At 03:46 PM 4/3/2001 -0400, you wrote: >At 12:15 PM 4/3/2001 -0700, you wrote: >>how do I get those? is the route command? > >route -n
Re: [expert] network performance problem
Just how slow is it? Is it consistently slow? Try unplugging the cable modem and see what happens. That'll tell you if it's involved. If you get _long_ delays when it's unplugged, then it's probably trying to go out to the dns there, and a local dns may help. You can also put entries in your /etc/hosts file for all boxes on the lan. Karl At 16:20 3/25/2001 -0800, you wrote: >A question for you lan experts: > >I have a M7.2 box with a cable modem. This is >connected to a local lan with various M$ boxes, via a >10/100 switch. Local lan is 192.168.0.n, assigned with >DHCP from the linux box. > >Everything works OK, but ftp & X on the lan are very >slow, with heaps of activity on the hub LED's. > >I'm not familiar with this area much, but have the >impression it may be dns related, as I'm not running >the linux box as a dns. I'm wondering if the 192* >addresses are having packets going via the DNS which >is an IP address at my ISP, causing the speed problem. > > >If anyone can shed any light on the cause (& hopefully >a solution) of the poor local lan performance I'd >appreciate it :-) > > >adTHANKSvance > >Brent Wood > >__ >Do You Yahoo!? >Get email at your own domain with Yahoo! Mail. >http://personal.mail.yahoo.com/
Re: [expert] apcupsd 3.8.1 and power on/off
Your asus mb probably will have a power screen in the BIOS setup where you can tell it what to do when the power comes back on after a failure. The choices usually are 1) Never come back on, 2) Always come back on, or 3) Come back to the previous state. The more sure approach is to remove the wire (usually green) that goes from the power supply to pin 14 in the ATX power connector. Clip the wire off at the connector and hardwire it to one of the ground wires from the power supply. With this wire grounded, the power supply will be on anytime ac is applied to it. Karl At 09:29 PM 3/21/2001 +, you wrote: >I have two linux boxes sharing a APC 650VA back UPS that are networked >together. One is an old AT based K5 75MHz that spends its days pushing >and mangling packets and dolling out IP addresses. The other is an ATX >based Althon system with an ASUS K7V m/b. Communication with the UPSes >is good and the machines have proven to be able to talk to each other >when the power is out and shut down properly. I even have them set up >to have the UPS cut the power after a certain amount of time until the >power comes back. > >Now here is the problem. The AT based machine with a real switch has no >problems coming back on when the power comes back, but the ATX machine >with its "soft" switch stays off. I was wondering does anyone know of a >way to have the ATX based system come back up automatically when the >power comes back?
Re: [expert] Mail delivery problems
Do you have reverse DNS lookups available for both the domain and subdomain? A lot of SMTP servers require this before they'll accept mail. Karl Cunningham At 01:34 PM 3/16/2001 -0800, you wrote: >When I get this message: > >|- Failed addresses follow: -| > <[EMAIL PROTECTED]> ... transport smtp: 550 relaying to ><[EMAIL PROTECTED]> prohibited by administrator > >How do I know which administrator is prohibiting it. I am sending a message >from my address at my domain, to my user account at a sub-domain that I >have created using another linux box. I assumed that I would be able to >send mail between the domain and the subdomain, but I can't.
Re: [expert] Network Card Problems
Consider setting the PCI IRQ's manually via BIOS rather than letting the mobo set them itself on bootup. And use setup programs that usually come with PNP ISA cards to set their IRQs. I've done this on the last few systems I've put together and have had a no problems at all with IRQs. Most newer BIOSes will let you do this and almost every PCI card uses IRQ A, so you can set them up however you like. Watch out for things like the Realtek-based NICs that can't reliably share IRQs. At least you know what's going on if you set them up manually. Karl At 02:48 PM 3/6/2001 +0100, you wrote: > >I had a similar problem lately. I had a defective mobo replaced, and after >that the soundblaster didn't work. After switching the soundblaster and the >NIC from their respective slots, the NIC didn't work... > >Then I put the NIC in a different PCI-slot, assuming the previous slot must >have been broken. Then it appeared to work, linux detected the NIC, the >modules loaded OK, everything went cool, untill I actually tried to use the >network. > >Further testing revealed that the replacement mobo had 2 total dead and 1 >cripple PCI-slot. After sending it back, and getting 2 more mobos (next >replacement came with defective PS/2 connections... Djees!) it finally >works. > >Bottom line: screw open the case, and try putting the NIC in a different >PCI slot, and see if it helps. If it does, either the mobo is defective, or >you got one of them PCI incompatibility things that occurs on the last (or >is it first) slot of a PCI-system... Which is completely normal, btw...
Re: [expert] Removing Xwindows
At 22:42 3/4/2001 -0800, you wrote: >openssh-askpass is not needed by openssh. The dependency should be the >other way around. Do an 'rpm -q --requires openssh' to be sure. I build >my own openssh RPM without any such dependency and they have always worked >fine. RPM reports the dependency as this.. openssh-askpass requires openssh-clients. Looking at the files included in the openssh-clients rpm, it includes its own ssh-askpass. And the openssh-askpass rpm seems to be completely an Xwindows package. Based on this, I think the dependency is probably fictitious. And to prove it, I removed -askpass (with --nodeps) and ssh client and server still work just fine. Karl
Re: [expert] Removing Xwindows
Thought I'd post an update... I had already removed a few packages and XWindows wasn't working, so I couldn't start it to remove some of the optional packages from there. So I started with a list from rpm -qa | grep X and proceded to remove things by manually recursing the dependency tree using rpm -e There were a few circular dependencies where I used --nodeps. Then went in and cleaned up the deadwood subdirectories left after the process. Unfortunately, rpm listed four libraries contained in XFree86-libs as needed by openssh-askpass, which is needed by openssh, which I definitely want to keep. I'm guessing that the openssh issue is a Mandrake-specific thing, if it is even a real dependency at all. Priobably compiling a generic openssh package would get around this, but ssh is working well and I don't want to break it. Any thoughts? Whole thing took about an hour. Not bad, really. Best news is that everything I want seems to still work. Karl
Re: [expert] Removing Xwindows
Thanks Shawn and John. Sounds like something for a rainy day. Shawn, I chose "server" during install. I'll let you know how it went after I do it. Thanks again, Karl At 09:33 3/2/2001 -0800, you wrote: >Sadly, reinstalling won't work. Mandrake just refuses to install without X. But never fear, it's >POSSIBLE to remove all offending packages (rpm packages). > ># rpm -qa |grep [xX] >will give you a bunch of packages, probably not all of which are X-related, but many of which >are. > >Here's what *I* would do: Go into X, fire up rpmdrake and remove all of your optional x programs >(xemacs, etc etc). Get rid of a few libraries too, if possible. It will blab about a bunch of >dependancies and do you want to remove those too, and you say yes. Mass uninstallation of >electives. That should make the list generated from the command above quite a bit shorter. Then, >with X not running, remove the rest. Read the manpage for rpm, but it's something along the lines >of ># rpm -e >and if it whines about depenancies, then go remove the dependancies first, then the . > >There must be easier ways to do it, i can think of one that involves a series of shell scripts, >but it's one way to do it. > >The moral of the story is that all of the X-related files you want gone are a product of installed >rpm packages, so "the correct way" to rid your system of them is to uninstall the rpm packages. > >Good luck, >j > >--- Karl Cunningham <[EMAIL PROTECTED]> wrote: >> Hi -- >> >> Does anyone have a procedure for removing the entire Xwindows system and >> everything under it? I have LM7.2, which installed X by default, but I >> don't expect to use it on this box and I'd like to remove the extra clutter. >> >> I can just remove the appropriate subdirectories, but I imagine there'd be >> trouble if I only did that. I can reinstall, but I'd rather not... >> >> Thanks, >> >> Karl Cunningham >> > > >__ >Do You Yahoo!? >Get email at your own domain with Yahoo! Mail. >http://personal.mail.yahoo.com/ > > >
[expert] Removing Xwindows
Hi -- Does anyone have a procedure for removing the entire Xwindows system and everything under it? I have LM7.2, which installed X by default, but I don't expect to use it on this box and I'd like to remove the extra clutter. I can just remove the appropriate subdirectories, but I imagine there'd be trouble if I only did that. I can reinstall, but I'd rather not... Thanks, Karl Cunningham
