Discuss

discuss@lists.centos.org

34753 discussions

[OT] rsa host key change
by Ed Donahue 17 Apr '09

17 Apr '09

Anyone ever come across a linux server host key changing with out a reboot, sshd restart, change in negotiating (SSHv1, SSHv2), and different DNS name or IP address? I have a server on RHEL4.4 that changed its host key. Red Hat Enterprise Linux ES release 4 (Nahant Update 4) openssh-server-3.9p1-8.RHEL4.15 2.6.9-42.ELsmp uptime 944 days Started getting the eavesdropping message from a login that is supposed to login w/o a password. And always did before today.

3 2

CUPS oddity
by Niki Kovacs 17 Apr '09

17 Apr '09

Hi, I've been setting up a few printer servers with CUPS. Our public libraries here all run 100% Linux (CentOS 5), so what I do is simply install the printer on one of the machines (with a static IP) and then configure CUPS so it can act as a printer server for Linux clients. It took me some time to figure this out, and I remember pulling my hair out for a few sunny afternoons, but I eventually got it working. One thing kept me stuck for days, until I finally found the answer on debian-administration.org. My server machine has an IP of 192.168.1.252/255.255.255.0. There's a Brother laser printer attached to it. The CUPS configuration file is edited so machines on the 192.168.1.* network can use it, and I can also access the CUPS web interface from anywhere. But here's what puzzles me. In the default cupsd.conf configuration file the 'Listen' directive looks like this: # Only listen for connections from the local machine. Listen localhost:631 Listen /var/run/cups/cups.sock Here's the oddity: I *have* to specifically add the machine's IP, like this: Listen localhost:631 Listen 192.168.1.252 Listen /var/run/cups/cups.sock If I don't add this, machines on the network are unable to connect to the server. Can anybody explain this strange behaviour to me? Cheers, Niki

3 3

CentOS 4 dkms-ndiswrapper
by Agile Aspect 17 Apr '09

17 Apr '09

Hi - I'm trying to get wireless running on CentOS 4.7 on a dual core laptop (latitude-e4500) for an employee. I'm having trouble with building dkms-ndiswraper-1.54-1.el4.rf - enclosed are the errors messages. The kernel was rebuilt to disable CONFIG_4KSTACKS. The errors prior to rebuilding the kernel are identical to the errors after rebuilding the kernel and rebooting - minus the error regarding the CONFIG_4KSTACKS. Basically I just typed make in the directory. I'm using dkms this time - I've never used it before. I'm under the impression that all I had to do was type 'make' in the directory in the ndiswrapper directory. Any help would be greatly appreciated! -- Agile make -C /home/praveen/rpmbuild/BUILD/kernel-2.6.9/linux-2.6.9 M=/usr/src/ndiswra pper-1.54-1.el4.rf make[1]: Entering directory `/home/praveen/rpmbuild/BUILD/kernel-2.6.9/linux-2.6 .9' MKEXPORT /usr/src/ndiswrapper-1.54-1.el4.rf/crt_exports.h CC [M] /usr/src/ndiswrapper-1.54-1.el4.rf/crt.o In file included from /usr/src/ndiswrapper-1.54-1.el4.rf/crt.c:16: /usr/src/ndiswrapper-1.54-1.el4.rf/ntoskernel.h:289:1: warning: "for_each_possib le_cpu" redefined In file included from include/linux/sched.h:15, from include/linux/module.h:10, from /usr/src/ndiswrapper-1.54-1.el4.rf/ntoskernel.h:22, from /usr/src/ndiswrapper-1.54-1.el4.rf/crt.c:16: include/linux/cpumask.h:374:1: warning: this is the location of the previous def inition In file included from /usr/src/ndiswrapper-1.54-1.el4.rf/crt.c:16: /usr/src/ndiswrapper-1.54-1.el4.rf/ntoskernel.h:649: error: field `lock' has inc omplete type /usr/src/ndiswrapper-1.54-1.el4.rf/ntoskernel.h: In function `raise_irql': /usr/src/ndiswrapper-1.54-1.el4.rf/ntoskernel.h:686: warning: implicit declarati on of function `mutex_lock' /usr/src/ndiswrapper-1.54-1.el4.rf/ntoskernel.h: In function `lower_irql': /usr/src/ndiswrapper-1.54-1.el4.rf/ntoskernel.h:708: warning: implicit declaration of function `mutex_unlock'make[2]: *** [/usr/src/ndiswrapper-1.54-1.el4.rf/crt.o] Error 1 make[1]: *** [_module_/usr/src/ndiswrapper-1.54-1.el4.rf] Error 2 make[1]: Leaving directory `/home/praveen/rpmbuild/BUILD/kernel-2.6.9/linux-2.6.9' make: *** [modules] Error 2 -- "We are drowning in information and starving for knowledge." -- Rutherford D. Roger

2 1

Re: [CentOS] Dual-boot with WinXP, CentOS already installed
by Kai Schaetzl 17 Apr '09

17 Apr '09

Sorin Srbu wrote on Fri, 17 Apr 2009 09:16:33 +0200: > Apparantely Windows can bork up after a while if the system files aren't on > C:. No, that is a myth. You just cannot put the system drive on an extended partition, it has to be a primary partition. If that is the first non-Linux partition Windows will name it C, anyway, as it just ignores partition types it doesn't know. But, as I said, the letter doesn't matter. I was thinking the Windows installer will see the linux partitions and > try to name them C: and D: etc, thus Windows will be installed on E: or F:, > which might not go down well with some programs. No, see above, but really, it doesn't matter at all. Windows isn't that dumb. > > > >> Would I be better off disconnecting the drive containing CentOS and > >> reconnect it when I'm done installing Windows. Then boot with CentOS > rescue > >> and reinstall grub to the Windows-hd MBR? > > > >Why? If you want to put them on separate *hardware* then you don't have to > >reinstall anything. Just make sure that the CentOS drive is the first boot > >drive. > > Windows drive numbering, see above. This way Windows doesn't see linux at > all. No, I meant: why then reinstall grub? No need for it. You just change grub.conf. If you want to use a second disk because of drive numbering: no, you don't need to do that. If you want to do it for other reasons, like space: go ahead! Kai -- Kai Schätzl, Berlin, Germany Get your web at Conactive Internet Services: http://www.conactive.com

1 0

Re: [CentOS] DKMS and new(er) Nvidia-drivers
by MHR 17 Apr '09

17 Apr '09

On Wed, Apr 15, 2009 at 1:08 AM, Sorin Srbu <sorin.srbu(a)orgfarm.uu.se> wrote: > > Diagonal thin black lines originating from upper left corner for starters. > Then if I open a gui, whatever sort, that window gets those lines too. Menus > are unreadable because of this, but slightly more readable if I move the > mouse pointer over the menu entry. The lines tend to go away for a short > while if I log off and log on again. Weird thing is that the lines are > always diagonal and tend to always originate from the upper left corner of > whatever window. > > This is with gnome mind you, and *supposedly* this one is the most stable of > all the desktop environments. Haven't tried with KDE and xfce. Can't tell > for sure if it's gnome or the Nvidia drivers specifically, but I'm leaning > towards the drivers. > > It's not a hardware issue, as I've run rhel3 on the same machines w/o any > artifacts. > > The hardware's two-three year old Asus mobo with a single-core AMD x64 and a > rather feisty Nvidia Quadra gfx card. Don't have the exact details right > now, but it should give a hint or two. All the P4-machines, as well as the > i7-boxes, seem to work fine with dkms. I've never seen this problem at all - running AMD 64x2 7750, 4Gb memory with GEForce 7200gs card, nvidia driver nvidia-x11-drv-173.08-1.beta.el5.rf.x86_64 (from rpmforge - duh). I've used a 19" CRT, 17" 1280x1024 flat panel and, currently, 22" 1680x1050 flat panel (Emprex - cheap, but works great). I tried the L&G nvidia driver last year some time, but I decided I'd rather have the rpmforge working dkms driver than the L&G, even though it's a beta driver. When I upgraded to 5.3, that was the first time the driver was actually rebuilt for the kernel since 5.0. HTH mhr

2 1

Re: [CentOS] Dual-boot with WinXP, CentOS already installed
by David G. Miller 17 Apr '09

17 Apr '09

"Sorin Srbu" <sorin.srbu(a)orgfarm.uu.se> wrote: > Hi all, > > I'd like to deploy a solution with dual-booting systems where CentOS 5.3 is > already installed and WinXP will be installed to a separate disk. > > I found > http://apcmag.com/how_to_dual_boot_linux_and_windows_xp_linux_installed_fir… > and it seems straight forward enough, although the description is for Ubuntu. > > The problem as I see it, is that the how-to differs from how CentOS looks in > /etc/grub.conf and the boot-loader in Ubuntu with respect to making grub work > again after the Windows install. > > I found Tldp.org mentioning dual-boot plenty, but most or all articles listed > are using lilo as a boot-loader, which seems a bit obsolete and besides I > can't quite translate the instructions from lilo to grub. 8-/ > > Basically, what would I need to change in the how-to from apcmag.com above?? > Thanks for any pointers. As noted in other replies, Windows not not play well with anything else. Your best bet would be to disable/remove/disconnect the drive with CentOS and then do the Windows install to the other drive. Once Windows is installed, you can reconnect the CentOS drive and then easily edit grub.conf to boot Windows using the "chainloader" directive. This approach will only work if there are no primary partitions on the CentOS drive that Windows recognizes. Windows (just like DOS) assigns drive letters to partitions in drive number order starting with the primary partitions and then moving on to the extended partitions. The bottom line is that you want the first Windows partition on the non-CentOS drive to get assigned drive letter C: when Windows boots. As long as Windows doesn't recognize the partition type (e.g., ext3), no drive letter gets assigned. The problem with suggestions to just install Windows and then either use another boot loader or repair the grub installation is that you are stuck doing that work through Windows. My experience has been that Windows isn't as easy to work with for making changes to the boot loader and any approach that overwrites the Windows boot loader may leave you with Windows not being bootable. The approach I described above leaves the Windows boot loader in place although you may need to edit C:\boot.ini to make sure it can still find Windows when both disks are "present." Here is the grub.conf that I use on my laptop. The internal drive boots to either CentOS or Windows while Fedora 10 and Ubuntu are on an external USB drive. # grub.conf generated by anaconda # # Note that you do not have to rerun grub after making changes to this file # NOTICE: You have a /boot partition. This means that # all kernel and initrd paths are relative to /boot/, eg. # root (hd0,2) # kernel /vmlinuz-version ro root=/dev/hda6 # initrd /initrd-version.img #boot=/dev/hda default=0 timeout=5 splashimage=(hd0,2)/grub/splash.xpm.gz hiddenmenu title CentOS (2.6.18-92.1.22.el5) root (hd0,2) kernel /vmlinuz-2.6.18-92.1.22.el5 ro root=LABEL=/ rhgb initrd /initrd-2.6.18-92.1.22.el5.img title CentOS (2.6.18-92.1.18.el5) root (hd0,2) kernel /vmlinuz-2.6.18-92.1.18.el5 ro root=LABEL=/ rhgb initrd /initrd-2.6.18-92.1.18.el5.img title Fedora (2.6.27.19-170.2.35.fc10.x86_64) root (hd1,0) kernel /vmlinuz-2.6.27.19-170.2.35.fc10.x86_64 ro root=UUID=232029fe-c524-47b4-a5d6-d45c0ce7e56b rhgb nomodeset initrd /initrd-2.6.27.19-170.2.35.fc10.x86_64.img title Ubuntu (2.6.27-7 x86_64) root (hd1,4) kernel /vmlinuz-2.6.27-7-generic ro root=LABEL=uSlash initrd /initrd-2.6.27-7-generic title Windoze rootnoverify (hd0,0) chainloader +1 My grub.conf points to Windows on a partition on the internal drive. Yours would point to Windows on the "second" drive (probably hd1). Likewise, your boot.ini will need to point to Windows with something like: multi(0)disk(0)rdisk(0)partition(1)\WINDOWS="Microsoft Windows XP" changing to: multi(0)disk(0)rdisk(1)partition(1)\WINDOWS="Microsoft Windows XP" Note the change -----^ I'm assuming your Windows disk would end up being /dev/hdb in a Linux world. If your system is using SATA drives or the Windows drive would be something other than /dev/hdb, you may need to make some adjustments to what I have suggested. Cheers, Dave -- Politics, n. Strife of interests masquerading as a contest of principles. -- Ambrose Bierce

7 8

Re: [CentOS] Dual-boot with WinXP, CentOS already installed
by Kai Schaetzl 17 Apr '09

17 Apr '09

Sorin Srbu wrote on Thu, 16 Apr 2009 14:35:40 +0200: > Googled some more. Realised /boot/grub/menu.lst *is* /etc/grub.conf... Duh! You want to change /boot/grub/grub.conf, nothing else ! > > Also some people say it's better to have Windows installed to the first > harddrive and the first partition (so that C: is the where it should be on > 1st hd/1st partition). Doesn't matter. > > Would I be better off disconnecting the drive containing CentOS and > reconnect it when I'm done installing Windows. Then boot with CentOS rescue > and reinstall grub to the Windows-hd MBR? Why? If you want to put them on separate *hardware* then you don't have to reinstall anything. Just make sure that the CentOS drive is the first boot drive. Kai -- Kai Schätzl, Berlin, Germany Get your web at Conactive Internet Services: http://www.conactive.com

3 4

Re: [CentOS] Dual-boot with WinXP, CentOS already installed
by Kai Schaetzl 17 Apr '09

17 Apr '09

Sorin Srbu wrote on Thu, 16 Apr 2009 16:03:53 +0200: > Generally speaking, which one is the easiest and/or safest if time is an issue > and you want to things fast and streamlined doing this, grub-install or backup > and restore MBR. That is assuming I don't screw up when I do either and > mistype or something. > > Or are both things different sides of the same coin kinda' thing? I think I prefer the grub-install way because it's easier to remember what you have to do and harder to mistype. Using dd works 100%, though, if done correctly. Even if there is something wrong with grub-install or grub in the live disk, dd is hardly affected. But, as I said, if you use two disks and pull the first out you do not have to reinstall/restore the MBR anyway. Kai -- Kai Schätzl, Berlin, Germany Get your web at Conactive Internet Services: http://www.conactive.com

2 1

Re: [CentOS] Dual-boot with WinXP, CentOS already installed
by Kai Schaetzl 17 Apr '09

17 Apr '09

Sorin Srbu wrote on Thu, 16 Apr 2009 14:08:57 +0200: > Basically, what would I need to change in the how-to from apcmag.com above?? It's like that tutorial says. Forget about backing up grub.conf. This is nonsense. You want to make a backup before changing it, just in case, but not because of the XP installation. After the installation you have to reinstall grub with "grub-install" (from a live disk) and add the chainloader stuff to grub.conf. mount /dev/sda2 /mnt/linux (if that is where you linux resides) grub-install --root-directory=/mnt/linux /dev/sda (if you want to install it to the partition use sda2 instead of sda, then the partition has to be active, so that you can boot from it) Instead of using grub-install you can backup the mbr before the installation and then restore it afterwards. You can easily get bad results from that if you mistype, though. Kai -- Kai Schätzl, Berlin, Germany Get your web at Conactive Internet Services: http://www.conactive.com

3 5

Re: [CentOS] rt3 3.8.2 latest version as rpm for C5?
by Mark Pryor 17 Apr '09

17 Apr '09

Jim, --- On Tue, 4/14/09, Jim Perrin <jperrin(a)gmail.com> wrote: > From: Jim Perrin <jperrin(a)gmail.com> > Subject: Re: [CentOS] rt3 3.8.2 latest version as rpm for C5? > To: "CentOS mailing list" <centos(a)centos.org> > Date: Tuesday, April 14, 2009, 5:12 PM > On Tue, Apr 14, 2009 at 3:07 AM, > Rainer Traut <tr.ml(a)gmx.de> > wrote: > > Hi, > > > > I know, epel has rt3 in version 3.6.x. > > And - sadly - for version 3.8.x the perl module > dependencies are way > > ahead of what I can find in epel or rpmforge. > > But we really need some of the 3.8.x series features. > > > > So has anybody maybe built those modules + rt3 v3.8.x > for C5? > > Any hint? > > I started to run down this rabbit hole and had most of it > built via > mock and one of dag's cpan-> rpm scripts. After a few > rounds of > builds, it became immediately apparent that some of the > requirements > directly conflict with the perl rpm shipped by centos and > RHEL. > Specifically the Encode bits required by RT 3.8 are much > newer than > what is provided in the centos perl rpm. I wasn't willing > to do that > level of replacement when building RT because I didn't need > the > feature sets in 3.8 and I wasn't aware of what else that > might impact > for other users. After trying for 2 hours, I ran into 3 brickwall-issues. Like you found, there are two base perl modules which require newer versions than that which comes with 5.3. These modules are File::Temp and Encode (core) cpan2rpm can package them OK, but there are fatal transaction check errors when installing. Mr. JV is the author of RT3 and DBIx::SearchBuilder. The SearchBuilder module also stopped me (briefly) with a dependency on DBD::Oracle. To grab this from CPAN requires the supporting Orable Libs. I altered the spec and removed any Oracle depends. These 3 issues were enough to cause me to give up. -- Mark

2 1

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Discuss