Re: [expert] Apache error in 2.4 kernel
Where did you get your version of Apache? - source from apache.org? - binaries from apache.org? - rpm from Mandrake mirror? What modules have you compiled in? What does you Apache config file look like? - Using the same one you used before - Old config with mods? - New config provided with distribution? What modules do you load? When does the error occur? - When you try to start Apache? - When delivering a static html page? - When executing a script? - When executing a cgi program? - Completely at random with no rhyme or reason? Where do you see the error? - On the console at startup? - In error_log file? - Delivered to the browser accessing a page? The more complete the context you can provide, the more clues people will have to try to help you. DP > [mailto:[EMAIL PROTECTED]]On Behalf Of [EMAIL PROTECTED] > Sent: Friday, March 09, 2001 12:40 PM > > I've been using Apache on 2.2.x kernels for over a year and never had a > problem with, now I built 2.4.1 and 2.4.2 kernels and Apache gives some mm > error "failed to acquires shared memory segment " or something similar. I > boot back into 2.217 and everything's smooth. What's the problem? > thanx
Re: [expert] Apache error in 2.4 kernel
If 2.2 was working, why did Linus work on 2.4? What is 2.5 for? My infant son is content suckling his mother, why should he be weaned? Seriously mattg, Although Linux is not as vlunerable to lifecycle issues as commercial operating systems, there are limits. As hardware advances, it just isn't practacle to upgade antiquated releases of code - support for the new hardware goes into new releases. It makes good sense to upgrade the underlying OS and keep things current and take advantage of hardware advances as well as advances in the underlying OS (memory mgmt, file system support, etc...) For your short sightedness, I must assume one of: - your applications are greener than you and won't live to see tomorrow - you have a cache of DPS-6 hard drives and will NEVER upgrade - PICK on a Honeywell is they way, the light, and the glory. I hate to break it to you, mattg, but tomorrow will come for some of us. If what you have created will never grow and never change then I have pity. DP on Monday 12 March 2001 10:30, goldengull.net administrator wrote: > if it was working without a problem...then why switch? > > mattg > > > -Original Message- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED]]On Behalf Of [EMAIL PROTECTED] > Sent: Friday, March 09, 2001 12:40 PM > To: [EMAIL PROTECTED] > Subject: [expert] Apache error in 2.4 kernel > > > I've been using Apache on 2.2.x kernels for over a year and never had a > problem with, now I built 2.4.1 and 2.4.2 kernels and Apache gives some mm > error "failed to acquires shared memory segment " or something similar. I > boot back into 2.217 and everything's smooth. What's the problem? > thanx
Re: [expert] timezone
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 look at /etc/sysconfig/clock try setting "UTC=false" and then doing something like... # rdate -s clock-1.cs.cmu.edu && hwclock --systohc DP On Monday 26 February 2001 05:34, Lou Baccari wrote: > Hello, > > I have this one system that continues to bug me. I can not get the system > time to stay correct. It's always 5 hour behind the correct EST time. I've > tried Drakconf, linuxconf, timetool, and I've even set up xntp, but after > a reboot it returns itself 5 hour behind. Any ideas??? > > Thanks, > > Lou. -BEGIN PGP SIGNATURE- Version: GnuPG v1.0.4 (GNU/Linux) Comment: For info see http://www.gnupg.org iEYEARECAAYFAjqazl8ACgkQCIwrOqxGmNuoiACeKBmqPGE1eM9osN8Jym4cocmG AEcAn27aLVmY4ig9Ektd/8/YNF8g7K9k =MQA6 -END PGP SIGNATURE-
Re: [expert] LM 7.2 new install and network probs
On Friday 09 February 2001 06:11, Andreas Müller wrote: > I did not find any entries > concerning lo in any startupservices. I've never installed on a system without a NIC so I'm not sure what you've got ... - option 1 - the preferred road (if /etc/rc.d/init.d/network is set up) - If you have these two files: /etc/sysconfig/network-scripts/ifcfg-lo /etc/rc.d/init.d/network try this... # service network stop (should bring lo down) # service network start (should bring lo up) If that works, then add "network" to SysVinit: # chkconfig --levels 2345 network on -- option 2 - the other road -- add ifconfig lo 127.0.0.1 to /etc/rc.d/rc.local
Re: [expert] LM 7.2 new install and network probs
What happens when you type... # ifconfig lo 127.0.0.1 up Are you able to ping now? On Thursday 08 February 2001 12:30, Andreas Müller wrote: > [root@localhost /root]# ifconfig -a > lo Linkverkapselung:Locale Schleife > LOOPBACK MTU:3924 Metric:1 > Empfangene Pakete:0 Fehler:0 Weggeworfen:0 Überlauf:0 Rahmen:0 > Verschickte Packete:0 Fehler:0 Weggeworfen:0 Überlauf:0 Rahmen:0 > Kollisionen:0 Sendewarteschlangenlänge:0 > > [root@localhost /root]# ping 127.0.0.1 > PING 127.0.0.1 (127.0.0.1): 56 octets data > sendto: Network is unreachable >
Re: [expert] Changes made in fstab at boot
Ok - it may be bad form to reply to my own message, but I'm tired and I must note that the change I just suggested is a kludge that will only work reliably if you have compiled the supermount module for both 2.2.x and 2.4.x kernels. A bit more creativity is required if you want it to work for "either or". I'll leave that one as an exercise for someone with more energy than I've got! DP David G. Powers wrote: > -- > Solution... > -- > If you changed /etc/rc.d/init.d/mandrake_every_time you could solve this > "problem" > > (both of these should occupy one line) > -[ ! -f "/lib/modules/$(uname -r)/fs/supermount.o" ] && [ -x > /usr/bin/perl ];then > > +[ ! -f "/lib/modules/$(uname -r)/fs/supermount.o" -a ! -f > "/lib/modules/$(uname -r)/kernel/fs/supermount/supermount.o" > ] && [ -x /usr/bin/perl ];then > > --
Re: [expert] Changes made in fstab at boot
For those of you that get it, please forgive me for being long winded - feel free to press the delete key now! ...But... If you're having difficulties with supermount settings when you change between 2.2.x and 2.4.x kernels, I hope this message is of some assistance. 1. Take a look at /etc/rc.d/init.d/mandrake_everytime 2. Find these lines ... if grep -q '^\/.*supermount.*' /etc/fstab && \ [ ! -f "/lib/modules/$(uname -r)/fs/supermount.o" ] && [ -x /usr/bin/perl ];then action "Disabling Supermount" /usr/sbin/supermount -i disable fi 3. Decipher the code... -- a) If /etc/fstab contains the string "supermount" (yeah- a bit over simplified, $(man grep) for the absolute truth) then... b) Check to see if the supermount module is not in "/lib/modules/$(uname -r)/fs/supermount.o". (also make sure /usr/bin/perl can be executed) c) if the module wasn't there, then disable supermount -- -- To understand c) above, look at the man pages for supermount $> man supermount The -i option indicates that modifications are to be made to /etc/fstab The disable option indicates that supermount entries are to be removed If you want to get really comfortable with how your /etc/fstab file is getting changed, try the supermount command without the -i option. This (as you saw in the man page viewed in step 4) will output changes to stdout instead of updating /etc/fstab. #> supermount enable #> supermount disable -- Summary -- Mandrake installs the supermount module in... "/lib/modules/$(uname -r)/fs/supermount.o The 2.4.x kernel places the supermount module in... /lib/modules/$(uname -r)/kernel/fs/supermount/supermount.o *** Notice the DIFFERENT PATH to supermount.o. *** When you boot into 2.4.x the code in /etc/rc.d/init.d/mandrake_everytime removes the supermount. The next time you boot into a 2.2.x kernel, there is no longer a supermount in /etc/fstab therefore the file is left untouched [ see a) above ]. -- Solution... -- If you changed /etc/rc.d/init.d/mandrake_every_time you could solve this "problem" (both of these should occupy one line) -[ ! -f "/lib/modules/$(uname -r)/fs/supermount.o" ] && [ -x /usr/bin/perl ];then +[ ! -f "/lib/modules/$(uname -r)/fs/supermount.o" -a ! -f "/lib/modules/$(uname -r)/kernel/fs/supermount/supermount.o" ] && [ -x /usr/bin/perl ];then -- (Thanks to Tom Brinkman for the details that I expounded upon) > Hey > Thanks for the reply, but my problem is not that the supermount.o module > is not found. It is, that the '/etc/fstab' file is rewritten. What I like > to have is these 4 lines: > > /mnt/fd0 /mnt/fd0 supermount fs=vfat,dev=/dev/fd0 0 0 > /mnt/cdrom /mnt/cdrom supermount fs=iso9660,dev=/dev/cdrom 0 0 > /mnt/cdrom2 /mnt/cdrom2 supermount fs=iso9660,dev=/dev/cdrom2 0 0 > /mnt/ls120 /mnt/ls120 supermount fs=vfat,dev=/dev/hdd 0 0 > > so I edit the file to hold these lines, but after the next boot, they are > changed to: > > /dev/fd0/mnt/fd0vfat > sync,nosuid,noauto,user,nodev,unhide 0 0 > /dev/cdrom /mnt/cdrom iso9660 > ro,nosuid,noauto,exec,user,nodev0 0 > /dev/cdrom2 /mnt/cdrom2 iso9660 > ro,nosuid,noauto,exec,user,nodev0 0 > /dev/hdd/mnt/ls120 vfat nosuid,noauto,nodev,user 0 0 > > What can I do, to prevent this to happen? > > Thanks in advance > Mogens Jæger.
Re: [expert] Article about Mandrakesoft's "new direction"
I got the impression that the Mandrake distribution efforts will continue - the company is just trying to find a revenue stream to sustain their efforts. It sounds like a healthy move to me - perhaps I am just too optimistic. DP > > http://linuxgram.com/newsitem.phtml?sid=108&aid=11659 > > Sounds like it may be time to look at SuSE or RedHat again.
Re: [expert] replacement use of inetd & GPIB useage
I believe you're looking for /etc/xinetd.conf and /etc/xinetd.d/* On Thursday 01 February 2001 18:23, you wrote: > 1) I believe you're looking for /etc/services > > Regards, > Aj > > --- richard <[EMAIL PROTECTED]> wrote: > > Hi, two questions > > > > 1) inetd appears not to be used with Mandrake , I > > use inetd.conf to > > redirect incoming telnet on port 23 to another > > application, if I start inetd > > incoming to port 23 still gets to the login prompt > > even when directed > > to an application called node..
Re: [expert] Unsuccessful Ping
Forgive me for asking, but what do you mean by a false identiy?
Isn't a congested network that doesn't have enough bandwidth for an ICMP
packet a problem?
Assuming the congestion problem is not persistent, you could either increase
the ping count from 4 and/or use "-i" to set a longer send interval, thus
increasing your sample duration. Just how long will you accept to type of
congestion you're concerned about? Your tolerence exceeds mine!
You might consider analyzing traceroute output, perhaps you can identify
"acceptable" lag within your subnet.
On Thursday 01 February 2001 09:03, you wrote:
> I'm using the following to check for an unsuccessful ping, but if
> the line is open and bottlenecked enough to return a hundred percent
> loss, then the ping will return a false identity. Any suggestions?
>
> ping -c 4 -R mail.aculink.net > search-opening-data 2>&1;
> MOKAY=`egrep -ie packets search-opening-data | gawk -F " " '{print
> $7}'`; if [ "${MOKAY}" = "100%" ]
>
>
> Ignore the omittance of the remaining "if" statement.
>
> Note: When you reply to this message, please include the mailing
> list/newsgroup address in Cc: and my email address in To:.
>
> *
> Signed,
> SoloCDM
Re: [expert] kernel panic on 2.4.1-pre10 can't find root partition
Thanks for the input. I tried it and I still can't boot with the new kernel - `make install` created the system map & boot image in /boot. - I removed the initrd line and made sure the image line pointed to the fresh image - I ran lilo just to be sure all was well - I rebooted and got the nasty kernel panic: VFS: Cannon open root device "307" or 03:07 Please append a correct "root=" boot option Kernel Panic: VFS: unable to mount root fs on 03:07 On Tuesday 23 January 2001 12:45, you wrote: > add a make install this will skip the step of you copy the stuff > manually...
Re: [expert] Disk drive monitor
Check out www.bb4.com - they've got a GPL app that can monitor all sorts of stuff and send alerts. If you can write shell scripts, it is simple enough to add your own monitors. If you want something quick, just set this script to run via crontab. You can modify the notfication even to suit your needs. I use the email address for my cell phone so I get an alphnumeric page. DP - diskmonitor.sh - #!/bin/sh # set this to a reasonable percentage warn_level=50 # this is where alerts will be sent [EMAIL PROTECTED] # check disk utilization using `df` for percent in $(df | grep /dev/ | cut -c 52-54); do if [ $percent -gt $warn_level ]; then # construct and send a message echo -e "subject: disk utilization warning on \ $(uname -n)\n\n$(df -h)"| sendmail $recipient # one warning should suffice break fi done - On Tuesday 23 January 2001 12:09, you wrote: > Hi, > > Short of writing one, does anyone kno of a disk drive monitor that pops up > a warning on your screen when your partition is almost full? > > l #!/bin/sh warn_level=50 [EMAIL PROTECTED] for percent in $(df | grep /dev/ | cut -c 52-54); do if [ $percent -gt $warn_level ]; then echo "warning $percent" echo -e "subject: disk utilization warning on $(uname -n)\n\n$(df -h)"| sendmail @recipient echo -e "subject: disk utilization warning on $(uname -n)\n\n$(df -h)" break fi done
[expert] kernel panic on 2.4.1-pre10 can't find root partition
I am attempting to upgrade my kernel from 2.2.17 to 2.4.1-pre10 and I just can't seem to get it right! Please help!!! -- Env -- Linux Mandrake 7.2 gcc version egcs-2.91.66 (kgcc) ReiserFS 3.5.26 on all partitions (except swap!) I can boot into 2.2.17 and the ReiserFS partitions mount with no problem -- Steps -- untar... make mrproper make menuconfig Tried RieserFS as a module and as a built-in make dep make bzImage make modules make modules_install copied System.map to /boot/System.map-2.4.1-pre10 copied arch/i386/boot/bzImage to /boot/vmlinuz-2.4.1-pre10 mkinitrd /boot/initrd-2.4.1-pre10.img 2.4.1-pre10 added entries to /etc/lilo.conf image=/boot/vmlinuz-2.4.1-pre10 label-linux-pre root=/dev/hda7 initrd=/boot/initrd-2.4.1-pre10.img read-only /sbin/lilo everything went fine - no error messages ... until -- Results -- When I attempt to restart, I get a nasty message: VFS: Cannon open root device "307" or 03:07 Please append a correct "root=" boot option Kernel Panic: VFS: unable to mount root fs on 03:07 I booted back into 2.2.17 and took a look at /etc/fstab. My "/" partition is definitely on /dev/hda7 I tried 'rdev /boot/vmlinuz-2.4.1-pre10` which returns: Root device /dev/hda7 I tried setting the kernel root option an boot time; root=/dev/hda7 - no go! Obviously I don't know what I'm doing. Can someone please point me in a direction! Thanks, DP
Re: [expert] Linux Server hardware sizing / designing
You should consider using Tomcat - http://jakarta.apache.org Tomcat is purported to be much more robust than JServ (but I've never used JServ so I can't compare). Regarding the size of your server; I hate to say this, but it depends on what your servlets are doing. I assume there is a database of some sort involved which will dictate storage requirements . I'd get a prototype up and running on a couple of sub $1K boxes (web/db) and thrash those to find your bottlenecks. You can start with the Apache 'ab' tool (man ab) DP On Monday 22 January 2001 06:00, Steve wrote: > I would like to run a Apach WWW server with JServ servlets on LM7.2. Could > someone point me a link how to build my server to support my users? Total > users are 1000 so simultanously it might be about 100... ? I don't know how > to calculate the hardware needs...
Re: [expert] telnet failure
Kyle, There are two prime places to look. # cat /etc/xinetd.d/telnet.conf # cat /etc/hosts.allow Also, what is your security setting? (# echo $SECURE_LEVEL) If your not concerned about security, log in as root and try typing # msec 0 This will put you in "honeypot" mode where script kiddies & crackers can come prance in and abuse you like a ___ (fill in your favorite target). As a rule telnet is not a great daemon to run - do you really want clear-text passwords on your network? You should really consider SSH as an alternative. DP > > Kyle Hargraves wrote: > > > Hi, > > > > > > with regard to Mandrake 7.2 (with all recent .rpms installed) > > > and with nothing in /etc/hosts.deny and either an empty > > > /etc/hosts.allow or appropriate entries in /etc/hosts.allow I cannot > > > telnet to a recently built Mandrake 7.2 box. I get the initial > > > "connection" then a rejection : > > > kyle@johnson kyle]$ telnet purcell > > Actually this is not correct. True - inetd must be assigned > a run level but from the above responses - OBVIOUSLY tcp > (telnet) services are installed. Afterall - I can telnet OUT > but some weir'ed Mandrake'ism is preventing the connection inbound > > Kyle
Re: [expert] XFree 4.0.1 vs 3.3.x
Try running Xconfigurator DP > > On Sat, 20 Jan 2001, Neal Lippman wrote: > > > I just did my first install of LM 7.2 from d/l'd iso images. I was > > > surprised in that I seem to be running XFree 3.3.x rather than > > > 4.0.1 which I expected to see.
Re: [expert] glibc upgrades and accidental reboot
Have you tried running `rpm --rebuilddb`? DP On Friday 19 January 2001 05:01, Chris Spackman wrote: > Is there some way to fix the > -devel-oldversion problem? I don't like the idea of rpm thinking that > something is installed when it actually isn't.
Re: [expert] building, using a boot image driver
> 1) Is this *Redhat 7.0 boot image* compatible with > Linux-Mandrake Complete 7.2? The RedHat 7.0 boot image isn't going to enable you to install Mandrake 7.2 > 2) Is the initrd command similar to dd command I read about ? dd just copies an image from one location to another. initrd builds a ram disk image to be used at boot. This image includes the kernel, modules, etc... > 3) How can I build a boot image floppy if I only have > a laptop running Windows 2000 Professional for the > download of new SCSI drivers? (sad I know) rawrite allows you to transfer an existing boot image to a floppy. You can find rawrite.txt in the \dosutils\ folder on the LM install CD. If you want to create your own boot image I don't know how you could do that with Win2K. You need at least win32 port of modutils and I don't think that exists. I've got a couple of comments... Did you try an "expert" install (Press F1 upon boot from CD)? You should be able to select the NCR53c8xx driver to at least get things installed. If the NCR driver doesn't work, you've got to get a boot image that includes the SYM53x8xx module. Perhaps the "other.img" (on the install cd in /images/) contains this on - I don't know. During the install there are several sessions running. You can use the CTRL-ALT-F(n) keystokes to switch between them. There is even a command prompt which you might be able to load the sym53c8xx driver from (attached) using insmod. I'm pretty green at this stuff - perhaps someone else has better info/ideas. DP
Re: [expert] To mail list experts re this list
There's got to be a better way! (Yes - I'll help!) If there are no volunteers (or enough resources) from the Mandrake camp, I will accept requests for assistance! DP On Thursday 18 January 2001 18:09, dave wrote: > ... no math genius, but I think that the volume of traffic > on this this could be less than half of what it is if the typical delay > were reduced to some reasonable level. Moreover, the reduced redundancy > would make it a much more valuable resource and reference.
Re: [expert] Postfix Mail Problem
What does your /etc/postfix/aliases file contain? Is there a default recipient for the root mail? DP On Wednesday 17 January 2001 16:53, John LeMay Jr. wrote: > .. using LM7.2 with Postfix. ... > ... all mail destined for root on the mail server is going to the catch all > address...
[expert] RPM keeps locking up
I attempted to install the mc updates via MandrakeUpdate and my system seemed to freeze. I restarted in runlevel 3, grabbed the rpms via FTP and attempted to update using: rpm -Fvh mc-4.5.51-7.1mdk.i586.rpm I see: # mc ##" # VM: rpm process killed # I got the rpm from: ftp://download.sourceforge.net/pub/mirrors/mandrake/update/7.2/ If I switch to a consol with X running, I see... # VM: X process killed # VM: kdeinit process killed # VM: kdenotify process killed # VM: rpm process killed # between the invokation of the rpm command and the kernel message "VM: rpm process killed" my system does not respond, not on any virtual terminal. I Alt-F(n) between them but the login prompt does not get around to requesting a password. Que malo - please help Thanks, DP On Wednesday 13 December 2000 20:55, Andy wrote: > > When I try to install rpms, the computer locks up and becomes unresponsive. > It does this in all runlevels. When running in very verbose, it goes > through the series of ### and then into just a #. After > about a minute it stops and prints "Killed "and it exits. It freezes > completely in KDE. Where should I begin to look to solve this? Any help > would be much appreciated. Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
[expert] multiple copies of posts multiple copies of posts multiple copies of posts
How many times should a message be sent? How many times should a message be sent? How many times should a message be sent? DP Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
