On 5/7/08, Ralph Angenendt <ra+cento AT br-online.de> wrote:
Lanny Marcus wrote:
Scott: Great! If I can locate kernel-headers-2.6.18-53.1.14.el5 and kernel-2.6.18-53.1.14.el5 I can use the rpm -e command to remove them and then yum update again and that should update the kernel in her box. How do I locate them?
Why locate them? rpm -e takes the package *name*, not the package itself.
Yes. I woke up about 430 this morning and I realized that rpm can locate the file by itself. :-) However, I want to learn how to use find that Mark (mhr) mentioned!
[lanny@compaq1300 ~]$ su - Password: [root@compaq1300 ~]# rpm -e kernel-headers-2.6.18-53.1.14.el5 [root@compaq1300 ~]# rpm -e kernel-2.6.18-53.1.14.el5 [root@compaq1300 ~]# yum update Loading "priorities" plugin Loading "installonlyn" plugin Setting up Update Process Setting up repositories Reading repository metadata in from local files 263 packages excluded due to repository priority protections Resolving Dependencies --> Populating transaction set with selected packages. Please wait. ---> Downloading header for kernel to pack into transaction set. kernel-2.6.18-53.1.14.el5 100% |=========================| 258 kB 00:03 ---> Package kernel.i686 0:2.6.18-53.1.14.el5 set to be installed --> Running transaction check
Dependencies Resolved
============================================================================= Package Arch Version Repository Size ============================================================================= Installing: kernel i686 2.6.18-53.1.14.el5 updates 13 M
Transaction Summary ============================================================================= Install 1 Package(s) Update 0 Package(s) Remove 0 Package(s)
Total download size: 13 M Is this ok [y/N]: y Downloading Packages: (1/1): kernel-2.6.18-53.1 100% |=========================| 13 MB 03:36 Running Transaction Test Finished Transaction Test Transaction Test Succeeded Running Transaction Installing: kernel ######################### [1/1]
Installed: kernel.i686 0:2.6.18-53.1.14.el5 Complete! [root@compaq1300 ~]# exit
Notice that: I was able to remove both files, with rpm -e without any problem and I was able to run yum update, and it installed the new kernel, but not the kernel-header file.
[lanny@compaq1300 ~]$ uname -a Linux compaq1300.HOMELAN 2.6.18-8.el5 #1 SMP Thu Mar 15 19:57:35 EDT 2007 i686 i686 i386 GNU/Linux [lanny@compaq1300 ~]$
After rebooting the box, it is still on the original kernel, rather than on the new kernel.
I just installed kernel-headers-2.6.18-53.1.14.el5 again. Was that correct or not?
Probably very close to solving this now! Thanks much to everyone who has contributed ideas! Lanny
[lanny@compaq1300 ~]$ uname -a Linux compaq1300.HOMELAN 2.6.18-8.el5 #1 SMP Thu Mar 15 19:57:35 EDT 2007 i686 i686 i386 GNU/Linux [lanny@compaq1300 ~]$
After rebooting the box, it is still on the original kernel, rather than on the new kernel.
I just installed kernel-headers-2.6.18-53.1.14.el5 again. Was that correct or not?
Probably very close to solving this now! Thanks much to everyone who has contributed ideas! Lanny
Hi there do you have any mention of the new kernel in /etc/grub.conf?
you might find that the default kernel is still the original one in which case there would be a line like default=1 in grub.conf changing this to default=0 might bring up the new kernel on reboot i have an old dual processor box that boots from the previous kernel after updates for some reason which i haven't researched
hth
mike
Michael Simpson wrote on Wed, 7 May 2008 12:15:35 +0100:
do you have any mention of the new kernel in /etc/grub.conf?
Note, this is only a symlink to /boot/grub/grub.conf !
Kai
Michael Simpson wrote:
do you have any mention of the new kernel in /etc/grub.conf?
you might find that the default kernel is still the original one in which case there would be a line like default=1 in grub.conf changing this to default=0 might bring up the new kernel on reboot i have an old dual processor box that boots from the previous kernel after updates for some reason which i haven't researched
That's probably because your /etc/sysconfig/kernel contains:
# UPDATEDEFAULT specifies if new-kernel-pkg should make # new kernels the default UPDATEDEFAULT=no
Make the obvious change of "no" to "yes" if you want newly updated kernels to become the boot default.
On 5/7/08, Robert Nichols rnicholsNOSPAM@comcast.net wrote:
Michael Simpson wrote:
do you have any mention of the new kernel in /etc/grub.conf?
you might find that the default kernel is still the original one in which case there would be a line like default=1 in grub.conf changing this to default=0 might bring up the new kernel on reboot i have an old dual processor box that boots from the previous kernel after updates for some reason which i haven't researched
That's probably because your /etc/sysconfig/kernel contains:
# UPDATEDEFAULT specifies if new-kernel-pkg should make # new kernels the default UPDATEDEFAULT=no
Make the obvious change of "no" to "yes" if you want newly updated kernels to become the boot default.
--
Thanks for that!
It is a fedora 7 box which is responsible for internal dns and samba / backup so having it set this way has been kinda handy as i have tested the newer kernels on another similar but not as important box before altering the default in grub.conf.
This bit of knowledge is *very* useful for me at the present time as i have 2 CentOS 5.1 boxen (identical dell 2950s, dual quadcores, large raid 6 yadda yadda) one of which is in production for an electronic patient record for the NHS, the other for backup and testing of updates.
/me firing up vi /etc/sysconfig/kernel
cheers
mike
wrote on Wed, 7 May 2008 05:27:25 -0500:
I just installed kernel-headers-2.6.18-53.1.14.el5 again. Was that correct or not?
As you removed it earlier, yes. There doesn't seem to be any dependancy for it. AFAIK it's not necessary for operation, anyway, it's only necessary if something has to be compiled as it contains all the c header files.
Your current state is now not very clear. Please provide the following information:
rpm -qa kernel* ls -l /boot ls -l /boot/grub cat /boot/grub/grub.conf
Besides that I suggest you get a good book on using Linux on the console. If "Using Linux" by Que is still getting published I suggest that. They provide a wealth about Linux and the shell and you can use them for years. I have some rather oldish ones that still provide good value. You have now three Linux boxes or so if I got it right. You cannot do everything just "by desktop", you are a "Linux shop" now and you *have* to start learning about all the nice console tools sooner or later, better sooner ;-)
Kai
On 5/7/08, Kai Schaetzl maillists@conactive.com wrote:
wrote on Wed, 7 May 2008 05:27:25 -0500:
I just installed kernel-headers-2.6.18-53.1.14.el5 again. Was that correct or not?
As you removed it earlier, yes. There doesn't seem to be any dependancy for it. AFAIK it's not necessary for operation, anyway, it's only necessary if something has to be compiled as it contains all the c header files.
Your current state is now not very clear. Please provide the following information:
rpm -qa kernel*
[root@compaq1300 ~]# rpm -qa kernel* kernel-headers-2.6.18-53.1.14.el5 kernel-2.6.18-8.el5 kernel-2.6.18-53.1.14.el5 [root@compaq1300 ~]#
ls -l /boot
[root@compaq1300 ~]# ls -l /boot total 10672 -rw-r--r-- 1 root root 64556 Mar 5 11:58 config-2.6.18-53.1.14.el5 -rw-r--r-- 1 root root 62150 Mar 15 2007 config-2.6.18-8.el5 drwxr-xr-x 2 root root 1024 May 7 05:08 grub -rw------- 1 root root 2920899 May 7 05:08 initrd-2.6.18-53.1.14.el5.img -rw------- 1 root root 2196209 Sep 17 2007 initrd-2.6.18-8.el5.img drwx------ 2 root root 12288 Sep 17 2007 lost+found -rw-r--r-- 1 root root 80032 Nov 22 18:24 message -rw-r--r-- 1 root root 87584 Mar 5 11:58 symvers-2.6.18-53.1.14.el5.gz -rw-r--r-- 1 root root 83542 Mar 15 2007 symvers-2.6.18-8.el5.gz -rw-r--r-- 1 root root 904149 Mar 5 11:58 System.map-2.6.18-53.1.14.el5 -rw-r--r-- 1 root root 884787 Mar 15 2007 System.map-2.6.18-8.el5 -rw-r--r-- 1 root root 1791796 Mar 5 11:58 vmlinuz-2.6.18-53.1.14.el5 -rw-r--r-- 1 root root 1765428 Mar 15 2007 vmlinuz-2.6.18-8.el5 [root@compaq1300 ~]#
ls -l /boot/grub
[root@compaq1300 ~]# ls -l /boot/grub total 207 -rw-r--r-- 1 root root 63 Sep 17 2007 device.map -rw-r--r-- 1 root root 7616 Sep 17 2007 e2fs_stage1_5 -rw-r--r-- 1 root root 7456 Sep 17 2007 fat_stage1_5 -rw-r--r-- 1 root root 6720 Sep 17 2007 ffs_stage1_5 -rw------- 1 root root 816 May 7 05:08 grub.conf -rw-r--r-- 1 root root 6720 Sep 17 2007 iso9660_stage1_5 -rw-r--r-- 1 root root 8192 Sep 17 2007 jfs_stage1_5 lrwxrwxrwx 1 root root 11 Sep 17 2007 menu.lst -> ./grub.conf -rw-r--r-- 1 root root 6880 Sep 17 2007 minix_stage1_5 -rw-r--r-- 1 root root 9248 Sep 17 2007 reiserfs_stage1_5 -rw-r--r-- 1 root root 5427 Nov 22 18:24 splash.xpm.gz -rw-r--r-- 1 root root 512 Sep 17 2007 stage1 -rw-r--r-- 1 root root 104924 Nov 29 15:36 stage2 -rw-r--r-- 1 root root 7040 Sep 17 2007 ufs2_stage1_5 -rw-r--r-- 1 root root 6272 Sep 17 2007 vstafs_stage1_5 -rw-r--r-- 1 root root 8864 Sep 17 2007 xfs_stage1_5 [root@compaq1300 ~]#
cat /boot/grub/grub.conf
[root@compaq1300 ~]# cat /boot/grub/grub.conf # 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/VolGroup00/LogVol00 # initrd /initrd-version.img #boot=/dev/hda default=0 timeout=5 splashimage=(hd0,2)/grub/splash.xpm.gz hiddenmenu title CentOS (2.6.18-53.1.14.el5) root (hd0,2) kernel /vmlinuz-2.6.18-53.1.14.el5 ro root=/dev/VolGroup00/LogVol00 rhgb quiet initrd /initrd-2.6.18-53.1.14.el5.img title CentOS (2.6.18-8.el5) root (hd0,2) kernel /vmlinuz-2.6.18-8.el5 ro root=/dev/VolGroup00/LogVol00 rhgb quiet initrd /initrd-2.6.18-8.el5.img title Other rootnoverify (hd0,0) chainloader +1 [root@compaq1300 ~]#
Besides that I suggest you get a good book on using Linux on the console. If "Using Linux" by Que is still getting published I suggest that. They provide a wealth about Linux and the shell and you can use them for years. I have some rather oldish ones that still provide good value. You have now three Linux boxes or so if I got it right. You cannot do everything just "by desktop", you are a "Linux shop" now and you *have* to start learning about all the nice console tools sooner or later, better sooner ;-)
Kai: Thank you, again, for your time, ideas and help! You are absolutely correct, that I need to become as proficient as is possible, with the commands! I have "Running Linux, Fourth Edition" and ASAP, I will study the chapter(s) about the Command Line, which I know has all the power and none of the problems that GUI front ends sometimes have. Lanny
wrote on Wed, 7 May 2008 08:53:01 -0500:
-rw------- 1 root root 816 May 7 05:08 grub.conf
got changed 5:08 today. Are you sure you booted after that? What does uname -a show now?
default=0 timeout=5 splashimage=(hd0,2)/grub/splash.xpm.gz hiddenmenu title CentOS (2.6.18-53.1.14.el5) root (hd0,2) kernel /vmlinuz-2.6.18-53.1.14.el5 ro root=/dev/VolGroup00/LogVol00 rhgb quiet initrd /initrd-2.6.18-53.1.14.el5.img
The machine *is* booting vmlinuz-2.6.18-53.1.14.el5 and you should also be able to see all three available boot options. You should be seeing something like "booting CentOS in x seconds". You can interrupt this by pressing a key. You should then see all three options and the first one with CentOS (2.6.18- 53.1.14.el5) highlighted. If you continue with that I don't see a way it could not be booting into vmlinuz-2.6.18-53.1.14.el5.
What makes me wonder a bit is this: (hd0,2). If you set this system up new with CentOS 5 this would hardly be the case if you accepted default partitioning options. Is there another system, maybe Windows, on the disk or some external boot manager or so?
Kai
On Wed, 2008-05-07 at 18:30 +0200, Kai Schaetzl wrote:
-rw------- 1 root root 816 May 7 05:08 grub.conf
got changed 5:08 today. Are you sure you booted after that? What does uname -a show now?
Kai: Yes, I rebooted, very early this morning. I just tried it again. My wife had been using MS Windows XP. I rebooted the box into CentOS 5:
[lanny@compaq1300 ~]$ uname -a Linux compaq1300.HOMELAN 2.6.18-8.el5 #1 SMP Thu Mar 15 19:57:35 EDT 2007 i686 i686 i386 GNU/Linux [lanny@compaq1300 ~]$
default=0 timeout=5 splashimage=(hd0,2)/grub/splash.xpm.gz hiddenmenu title CentOS (2.6.18-53.1.14.el5) root (hd0,2) kernel /vmlinuz-2.6.18-53.1.14.el5 ro root=/dev/VolGroup00/LogVol00 rhgb quiet initrd /initrd-2.6.18-53.1.14.el5.img
The machine *is* booting vmlinuz-2.6.18-53.1.14.el5 and you should also be able to see all three available boot options. You should be seeing something like "booting CentOS in x seconds". You can interrupt this by pressing a key. You should then see all three options and the first one with CentOS (2.6.18- 53.1.14.el5) highlighted. If you continue with that I don't see a way it could not be booting into vmlinuz-2.6.18-53.1.14.el5.
What you describe above is how it works on my box and on my daughters box. They are working correctly. However, on my wife's box (compaq1300) there is a glitch, that is causing this not to work properly. On compaq1300, I do *not* have the three boot options (original kernel, latest kernel and Windows XP). I have two (2) options, if I interrupt grub: (a) the original kernel (b) other, which is Windows XP
As CentOS 5 boots, it shows, a number of times, the original CentOS 5 kernel, that is on the Install DVD I got last year. At *no* time, does it show the latest kernel booting.
What makes me wonder a bit is this: (hd0,2). If you set this system up new with CentOS 5 this would hardly be the case if you accepted default partitioning options. Is there another system, maybe Windows, on the disk or some external boot manager or so?
Yes. The 3 Desktop boxes are dual boot. CentOS 5 and Windoze XP. Grub is installed on all 3 boxes. Of the 3 boxes, 2 of them are working properly. :-) Have a nice evening and thank you! Lanny
Lanny Marcus wrote on Wed, 07 May 2008 13:13:13 -0500:
On compaq1300, I do *not* have the three boot options (original kernel, latest kernel and Windows XP). I have two (2) options, if I interrupt grub: (a) the original kernel (b) other, which is Windows XP
Ok, that explains it. I bet you see that different on your other systems. You either boot with the Windows Boot-Manager (which looks different from the CentOS one, so you should be able to easily see that) or with some Grub on *another* partition (not the boot partition on hd(0,2) which is the third partition on disk 1). In which order where the systems installed? Did you run into any trouble after installing the second one concerning the dual-boot scenario? Try to reminisce about the history of the system and what got installed when and how. And if you reinstalled grub (or fixed the Windows boot manager with fixmbr from the Windows recovery console) some time later for instance (and then to the wrong partition). That is the clue to understanding why it is different to your other systems. The Windows boot menu is called "boot.ini" and is in the root drive of the Windows installation (it's hidden in Windows). If that doesn't contain any mention of your CentOS, then try to mount the other unmounted partitions one by one and check which holds the other /boot partition.
Kai
On Wed, 2008-05-07 at 21:38 +0200, Kai Schaetzl wrote:
Lanny Marcus wrote on Wed, 07 May 2008 13:13:13 -0500:
On compaq1300, I do *not* have the three boot options (original kernel, latest kernel and Windows XP). I have two (2) options, if I interrupt grub: (a) the original kernel (b) other, which is Windows XP
Ok, that explains it. I bet you see that different on your other systems. You either boot with the Windows Boot-Manager (which looks different from the CentOS one, so you should be able to easily see that) or with some Grub on *another* partition (not the boot partition on hd(0,2) which is the third partition on disk 1). In which order where the systems installed? Did you run into any trouble after installing the second one concerning the dual-boot scenario? Try to reminisce about the history of the system and what got installed when and how. And if you reinstalled grub (or fixed the Windows boot manager with fixmbr from the Windows recovery console) some time later for instance (and then to the wrong partition). That is the clue to understanding why it is different to your other systems. The Windows boot menu is called "boot.ini" and is in the root drive of the Windows installation (it's hidden in Windows). If that doesn't contain any mention of your CentOS, then try to mount the other unmounted partitions one by one and check which holds the other /boot partition.
Kai: I am not using Windows Boot Manager. Grub comes up, as on the 2 boxes, where things are working properly.
Questions: (a) Can I copy /boot/grub/grub.conf on my box and replace that file on my wife's box, with my version? Would that work OK? Worth a try?
(b) Which files should I compare, between my box and my wife's box, the problematic one, to see if I can locate differences?
There was confusion on my part, when I installed Windows XP on my wife's box. Hers was the first one I installed Win XP on, which I'd never installed before and it ended up getting installed more than once. However, in general, I thought her box was the easiest, with regard to the WinXP installations. There may have been some partitioning issues also, since each box has 4 partitions on the Windows side (C, D, E & F). In general, it is *much* easier for me to install CentOS than to install Windoze. And, *much* faster. :-)
TIA, Lanny
Lanny Marcus wrote on Wed, 07 May 2008 15:10:58 -0500:
Kai: I am not using Windows Boot Manager. Grub comes up, as on the 2 boxes, where things are working properly.
Just to be sure, it's really grub? You get a somewhat blueish screen that says "booting centos in x seconds, press any key to see options" or so? I think there's also a CentOS symbol on it, but am not sure. We have to be absolutely sure about that. And if you select Windows from that boot screen, does that boot right into Windows or do you get another boot menu that lists only Windows?
Questions: (a) Can I copy /boot/grub/grub.conf on my box and replace that file on my wife's box, with my version? Would that work OK? Worth a try?
No, this wouldn't help, because the grub.conf that *we know of* is fine. It's just not getting used, because you are booting from another one. AFAIK, grub cannot embed a boot menu in the MBR (Master Boot Record), so that information must be coming from somewhere else. You have *two* grub.conf's (and two /boot partitions) on the machine AFAIS. You would have to *merge* the two: you need the options for booting Windows from the first one and all the other options from the second one. AFAIK, the MBR on your disk does not boot from hd(0,2), but from another partition. You have to find out which one that is and change the grub.conf on that partition accordingly. The caveat of this is that you would have to do this each time the kernel changes or you would need to change a bit more, so that this becomes the new boot partition. Another option would be to grub-install again and overwrite the current information in the MBR, so that it then boots from hd(0,2). I'm not confident enough about both options to talk you thru. Maybe I'm missing other possibilities why that happens, but the basic problem is that your machine does not boot from that hd(0,2), but with information from elsewhere.
There was confusion on my part, when I installed Windows XP on my wife's box. Hers was the first one I installed Win XP on, which I'd never installed before and it ended up getting installed more than once.
Did you install it after CentOS or before it? You will need to make a list of all partitions. Not sure what the best way to do this would be. Probably fdisk. Run fdisk, then type "p" (for printing the partition table), then leave it with "q". Be careful, as printing the table is only the least dangerous action in fdisk!
Kai
on 5-7-2008 3:19 PM Kai Schaetzl spake the following:
Lanny Marcus wrote on Wed, 07 May 2008 15:10:58 -0500:
Kai: I am not using Windows Boot Manager. Grub comes up, as on the 2 boxes, where things are working properly.
Just to be sure, it's really grub? You get a somewhat blueish screen that says "booting centos in x seconds, press any key to see options" or so? I think there's also a CentOS symbol on it, but am not sure. We have to be absolutely sure about that. And if you select Windows from that boot screen, does that boot right into Windows or do you get another boot menu that lists only Windows?
Questions: (a) Can I copy /boot/grub/grub.conf on my box and replace that file on my wife's box, with my version? Would that work OK? Worth a try?
No, this wouldn't help, because the grub.conf that *we know of* is fine. It's just not getting used, because you are booting from another one. AFAIK, grub cannot embed a boot menu in the MBR (Master Boot Record), so that information must be coming from somewhere else. You have *two* grub.conf's (and two /boot partitions) on the machine AFAIS. You would have to *merge* the two: you need the options for booting Windows from the first one and all the other options from the second one. AFAIK, the MBR on your disk does not boot from hd(0,2), but from another partition. You have to find out which one that is and change the grub.conf on that partition accordingly. The caveat of this is that you would have to do this each time the kernel changes or you would need to change a bit more, so that this becomes the new boot partition. Another option would be to grub-install again and overwrite the current information in the MBR, so that it then boots from hd(0,2). I'm not confident enough about both options to talk you thru. Maybe I'm missing other possibilities why that happens, but the basic problem is that your machine does not boot from that hd(0,2), but with information from elsewhere.
There was confusion on my part, when I installed Windows XP on my wife's box. Hers was the first one I installed Win XP on, which I'd never installed before and it ended up getting installed more than once.
Did you install it after CentOS or before it? You will need to make a list of all partitions. Not sure what the best way to do this would be. Probably fdisk. Run fdisk, then type "p" (for printing the partition table), then leave it with "q". Be careful, as printing the table is only the least dangerous action in fdisk!
Kai
Fdisk -l ( lower case L )should list ALL partitions and their respective devices without worrying about breaking something.
On Thu, 2008-05-08 at 00:19 +0200, Kai Schaetzl wrote:
Lanny Marcus wrote on Wed, 07 May 2008 15:10:58 -0500:
<snip>
You will need to make a list of all partitions. Not sure what the best way to do this would be. Probably fdisk. Run fdisk, then type "p" (for printing the partition table), then leave it with "q". Be careful, as printing the table is only the least dangerous action in fdisk!
Another way that I prefer is
sfdisk -l
It's output can be saved in a file and it can even make copies of things (like partition tables) that can be fed back into it to recreate a disk if the info gets wiped. "Man sfdisk" for some enjoyable reading... well, for old farts like me that don't mind man pages for find, etc. ;-)
Kai
HTH
On 5/7/08, Kai Schaetzl <maillists AT conactive.com> wrote:
Lanny Marcus wrote on Wed, 07 May 2008 15:10:58 -0500:
Kai: I am not using Windows Boot Manager. Grub comes up, as on the 2 boxes, where things are working properly.
Just to be sure, it's really grub? You get a somewhat blueish screen that says "booting centos in x seconds, press any key to see options" or so? I think there's also a CentOS symbol on it, but am not sure. We have to be absolutely sure about that.
Kai: On my box (dell2400) it is GNU GRUB v.0.97, with 3 choices (2 for Linux, 1 for WInXP). On my wife's box (compaq1300) it is also GNU GRUB v.0.97, but with 2 choices (1 for the original Linux kernel, 1 for WinXP).
And if you select Windows from that boot screen, does that boot right into Windows or do you get another boot menu that lists only Windows?
Ah. This is where the problem on her box probably comes from! Last November, when I installed Windows XP, in English, my native language, I did not understand something in Microsoft's English. There were four (4) partitions for Windows (C,D,E,F) and apparently when it was finished, I clicked incorrectly and it installed Windows again, So, it installed MS Windows XP onto the four (4) Windows partitions on her box! If Windows is selected on the GNU GRUB menu, then there is a 2nd boot menu, of Windows. The one to select is the 2nd from the top. I would redo her box completely, if she goes out of town, but then she will need to reinstall a lot of stuff. The installation of MS Windows on her box is a disaster! Lots of wasted space on the HD too.
Questions: (a) Can I copy /boot/grub/grub.conf on my box and replace that file on my wife's box, with my version? Would that work OK? Worth a try?
No, this wouldn't help, because the grub.conf that *we know of* is fine. It's just not getting used, because you are booting from another one. AFAIK, grub cannot embed a boot menu in the MBR (Master Boot Record), so that information must be coming from somewhere else. You have *two* grub.conf's (and two /boot partitions) on the machine AFAIS. You would have to *merge* the two: you need the options for booting Windows from the first one and all the other options from the second one. AFAIK, the MBR on your disk does not boot from hd(0,2), but from another partition. You have to find out which one that is and change the grub.conf on that partition accordingly. The caveat of this is that you would have to do this each time the kernel changes or you would need to change a bit more, so that this becomes the new boot partition. Another option would be to grub-install again and overwrite the current information in the MBR, so that it then boots from hd(0,2). I'm not confident enough about both options to talk you thru. Maybe I'm missing other possibilities why that happens, but the basic problem is that your machine does not boot from that hd(0,2), but with information from elsewhere.
There was confusion on my part, when I installed Windows XP on my wife's box. Hers was the first one I installed Win XP on, which I'd never installed before and it ended up getting installed more than once.
Did you install it after CentOS or before it?
When I began using Linux, I read that for dual boot boxes, MS Windows should always be installed first. So, I always install MS Windows first and then Linux.
You will need to make a list of all partitions. Not sure what the best way to do this would be. Probably fdisk. Run fdisk, then type "p" (for printing the partition table), then leave it with "q". Be careful, as printing the table is only the least dangerous action in fdisk!
Per Scott Silva's suggestion: [root@compaq1300 ~]# fdisk -l
Disk /dev/hda: 81.9 GB, 81964302336 bytes 240 heads, 63 sectors/track, 10587 cylinders Units = cylinders of 15120 * 512 = 7741440 bytes
Device Boot Start End Blocks Id System /dev/hda1 * 1 783 5919448+ 7 HPFS/NTFS /dev/hda2 784 5347 34503840 f W95 Ext'd (LBA) /dev/hda3 5348 10586 39606840 8e Linux LVM /dev/hda5 784 1566 5919448+ 7 HPFS/NTFS /dev/hda6 1567 2920 10236208+ 7 HPFS/NTFS /dev/hda7 2921 3733 6146248+ 7 HPFS/NTFS /dev/hda8 3734 3747 105808+ 83 Linux /dev/hda9 3748 5347 12095968+ 8e Linux LVM [root@compaq1300 ~]#
Trying to remember what I did on approximately 24 November 2007 is difficult. Probably, after shrinking the Windows partition with QTParted, to allow space for the CentOS installation, I let Windows do the partitioning and formatting for Windows, since I'd read that it is recommended to let Windows do that for Windows and Linux for Linux. However, I may have done the Windows partitionnig with QTParted. Or, the Windows XP installation may have given me the option to not use all of the HD for Windows, and if so, I would have elected to do that and leave space for CentOS5. I believe that I created and formatted the 4 Windows partitions within the Win XP installation. Not sure what went on that confused day. TIA! Lanny
on 5-8-2008 7:39 AM lannyma@gmail.com spake the following:
On 5/7/08, Kai Schaetzl <maillists AT conactive.com> wrote:
Lanny Marcus wrote on Wed, 07 May 2008 15:10:58 -0500:
Kai: I am not using Windows Boot Manager. Grub comes up, as on the 2 boxes, where things are working properly.
Just to be sure, it's really grub? You get a somewhat blueish screen that says "booting centos in x seconds, press any key to see options" or so? I think there's also a CentOS symbol on it, but am not sure. We have to be absolutely sure about that.
Kai: On my box (dell2400) it is GNU GRUB v.0.97, with 3 choices (2 for Linux, 1 for WInXP). On my wife's box (compaq1300) it is also GNU GRUB v.0.97, but with 2 choices (1 for the original Linux kernel, 1 for WinXP).
And if you select Windows from that boot screen, does that boot right into Windows or do you get another boot menu that lists only Windows?
Ah. This is where the problem on her box probably comes from! Last November, when I installed Windows XP, in English, my native language, I did not understand something in Microsoft's English. There were four (4) partitions for Windows (C,D,E,F) and apparently when it was finished, I clicked incorrectly and it installed Windows again, So, it installed MS Windows XP onto the four (4) Windows partitions on her box! If Windows is selected on the GNU GRUB menu, then there is a 2nd boot menu, of Windows. The one to select is the 2nd from the top. I would redo her box completely, if she goes out of town, but then she will need to reinstall a lot of stuff. The installation of MS Windows on her box is a disaster! Lots of wasted space on the HD too.
Questions: (a) Can I copy /boot/grub/grub.conf on my box and replace that file on my wife's box, with my version? Would that work OK? Worth a try?
No, this wouldn't help, because the grub.conf that *we know of* is fine. It's just not getting used, because you are booting from another one. AFAIK, grub cannot embed a boot menu in the MBR (Master Boot Record), so that information must be coming from somewhere else. You have *two* grub.conf's (and two /boot partitions) on the machine AFAIS. You would have to *merge* the two: you need the options for booting Windows from the first one and all the other options from the second one. AFAIK, the MBR on your disk does not boot from hd(0,2), but from another partition. You have to find out which one that is and change the grub.conf on that partition accordingly. The caveat of this is that you would have to do this each time the kernel changes or you would need to change a bit more, so that this becomes the new boot partition. Another option would be to grub-install again and overwrite the current information in the MBR, so that it then boots from hd(0,2). I'm not confident enough about both options to talk you thru. Maybe I'm missing other possibilities why that happens, but the basic problem is that your machine does not boot from that hd(0,2), but with information from elsewhere.
There was confusion on my part, when I installed Windows XP on my wife's box. Hers was the first one I installed Win XP on, which I'd never installed before and it ended up getting installed more than once.
Did you install it after CentOS or before it?
When I began using Linux, I read that for dual boot boxes, MS Windows should always be installed first. So, I always install MS Windows first and then Linux.
You will need to make a list of all partitions. Not sure what the best way to do this would be. Probably fdisk. Run fdisk, then type "p" (for printing the partition table), then leave it with "q". Be careful, as printing the table is only the least dangerous action in fdisk!
Per Scott Silva's suggestion: [root@compaq1300 ~]# fdisk -l
Disk /dev/hda: 81.9 GB, 81964302336 bytes 240 heads, 63 sectors/track, 10587 cylinders Units = cylinders of 15120 * 512 = 7741440 bytes
Device Boot Start End Blocks Id System /dev/hda1 * 1 783 5919448+ 7 HPFS/NTFS /dev/hda2 784 5347 34503840 f W95 Ext'd (LBA) /dev/hda3 5348 10586 39606840 8e Linux LVM /dev/hda5 784 1566 5919448+ 7 HPFS/NTFS /dev/hda6 1567 2920 10236208+ 7 HPFS/NTFS /dev/hda7 2921 3733 6146248+ 7 HPFS/NTFS /dev/hda8 3734 3747 105808+ 83 Linux /dev/hda9 3748 5347 12095968+ 8e Linux LVM [root@compaq1300 ~]#
Trying to remember what I did on approximately 24 November 2007 is difficult. Probably, after shrinking the Windows partition with QTParted, to allow space for the CentOS installation, I let Windows do the partitioning and formatting for Windows, since I'd read that it is recommended to let Windows do that for Windows and Linux for Linux. However, I may have done the Windows partitionnig with QTParted. Or, the Windows XP installation may have given me the option to not use all of the HD for Windows, and if so, I would have elected to do that and leave space for CentOS5. I believe that I created and formatted the 4 Windows partitions within the Win XP installation. Not sure what went on that confused day. TIA! Lanny
Looking at the above partition table, it does look like you installed Windows first, which is usually much easier since Windows throws fits if it is not the first partition. Linux is on the last 2 partitions, with what looks like a boot partition and the rest on LVM.
On Thu, 2008-05-08 at 10:30 -0700, Scott Silva wrote:
on 5-8-2008 7:39 AM lannyma@gmail.com spake the
<snip>
[root@compaq1300 ~]# fdisk -l
Disk /dev/hda: 81.9 GB, 81964302336 bytes 240 heads, 63 sectors/track, 10587 cylinders Units = cylinders of 15120 * 512 = 7741440 bytes
Device Boot Start End Blocks Id System /dev/hda1 * 1 783 5919448+ 7 HPFS/NTFS /dev/hda2 784 5347 34503840 f W95 Ext'd (LBA) /dev/hda3 5348 10586 39606840 8e Linux LVM /dev/hda5 784 1566 5919448+ 7 HPFS/NTFS /dev/hda6 1567 2920 10236208+ 7 HPFS/NTFS /dev/hda7 2921 3733 6146248+ 7 HPFS/NTFS /dev/hda8 3734 3747 105808+ 83 Linux /dev/hda9 3748 5347 12095968+ 8e Linux LVM [root@compaq1300 ~]#
<snip>
Looking at the above partition table, it does look like you installed Windows first, which is usually much easier since Windows throws fits if it is not the first partition. Linux is on the last 2 partitions, with what looks like a boot partition and the rest on LVM.
Scott: I am positive that I installed MS Windows XP first, after wiping the HD. As previously explained, my first installation of WinXP did *NOT* go well and there are 4 iterations on the HD..... I have never installed Windows after Linux.
Yes, it has a very small /boot for CentOS and then the LVM. Take care, Lanny
wrote on Thu, 8 May 2008 09:39:15 -0500:
Ah. This is where the problem on her box probably comes from! Last November, when I installed Windows XP, in English, my native language, I did not understand something in Microsoft's English. There were four (4) partitions for Windows (C,D,E,F) and apparently when it was finished, I clicked incorrectly and it installed Windows again, So, it installed MS Windows XP onto the four (4) Windows partitions on her box!
No, surely not. Windows installs only to one partition, the one you selected for installation. The other three are only NTFS partitions and can be used by Windows, but unless your wife installs something on them or puts data on them they are empty. Anyway, it's not a problem.
If Windows is selected on the GNU GRUB menu, then there is a 2nd
boot menu, of Windows.
Ah, okay. Grub probably copied the MBR to a file and when you select to boot from Windows it runs that MBR which then runs the boot.ini menu from the Windows system partition. I think that's the standard way.
Did you install it after CentOS or before it?
When I began using Linux, I read that for dual boot boxes, MS Windows should always be installed first. So, I always install MS Windows first and then Linux.
Yeah, that's fine. Although with Windows systems since 2000 you can just install them second, they integrate grub in their boot.ini.
Device Boot Start End Blocks Id System /dev/hda1 * 1 783 5919448+ 7 HPFS/NTFS
that's the active primary partition
/dev/hda2 784 5347 34503840 f W95 Ext'd (LBA)
that's the extended partition that holds hda5-9
/dev/hda3 5348 10586 39606840 8e Linux LVM
this is the third primary partition with LVM on it
/dev/hda5 784 1566 5919448+ 7 HPFS/NTFS /dev/hda6 1567 2920 10236208+ 7 HPFS/NTFS /dev/hda7 2921 3733 6146248+ 7 HPFS/NTFS /dev/hda8 3734 3747 105808+ 83 Linux
this looks like the boot partition
/dev/hda9 3748 5347 12095968+ 8e Linux LVM
another LVM partition
hda3 and hda9 are your Linux LVM partitions, maybe they belong to one volume group, I don't know (your fstab would tell more, there's also a graphical frontend for LVM in your desktop).
From your grub.conf we know that it thinks it's installed on (hd0,2), but
hd0,2 is hda3 (if I understand that correctly) and that is LVM, and grub can't boot from LVM because grub boots the kernel and only that knows about LVM. So, you are probably booting from hda8, but it's not in your fstab as the /boot partition. What does a "df" say? Does it list hda8 among the partitions? Probably not? Mount it and have a look at that partition, does it contain the same stuff as your /boot partition? If not mounted, do: mkdir /mnt/hda8 mount /dev/hda8 /mnt/hda8 cat /mnt/hda8/boot/grub/grub.conf Does this look like the grub.conf that is the *real* one booting your system?
[root@compaq1300 ~]#
Trying to remember what I did on approximately 24 November 2007 is difficult. Probably, after shrinking the Windows partition with QTParted, to allow space for the CentOS installation, I let Windows do the partitioning and formatting for Windows, since I'd read that it is recommended to let Windows do that for Windows and Linux for Linux. However, I may have done the Windows partitionnig with QTParted. Or, the Windows XP installation may have given me the option to not use all of the HD for Windows, and if so, I would have elected to do that and leave space for CentOS5. I believe that I created and formatted the 4 Windows partitions within the Win XP installation. Not sure what went on that confused day.
As you installed CentOS after Windows the whole problem probably hasn't anything to do with the Windows installation, but it's a mystery how and why you got two /boot partitions, one (non-accessable) on LVM and the real one (that doesn't seem to be known to the system).
Kai
On Fri, 2008-05-09 at 01:31 +0200, Kai Schaetzl wrote: <snip>
No, surely not. Windows installs only to one partition, the one you selected for installation. The other three are only NTFS partitions and can be used by Windows, but unless your wife installs something on them or puts data on them they are empty. Anyway, it's not a problem.
Kai: What you describe above is how it works on my box, my daughters box and on previous installations of MS Windows I"ve done. I always have four (4) partitions for Windows: C (Windows), D (Programs), E (Data) and F (Swap/Temp). Something went terribly wrong, when I installed on my wife's box and when it boots, at the Grub screen, if one selects "Windows", then one gets a Microsoft screen with four (4) options and one of those is selected to run Windows XP. On our other Desktops, selecting "Windows" at the Grub menu starts Windows XP booting. It appears that those four (4) options are all installations of Windows XP. The second one is the one that works properly.
I will reply to the things you asked about, in your last post to me, ASAP, when her box is available to me and give you whatever additional data I can find. TIA, Lanny
On 5/8/08, Kai Schaetzl <maillists AT conactive DOT com> wrote: <snip>
hda3 and hda9 are your Linux LVM partitions, maybe they belong to one volume group, I don't know (your fstab would tell more, there's also a graphical frontend for LVM in your desktop).
From your grub.conf we know that it thinks it's installed on (hd0,2), but
hd0,2 is hda3 (if I understa
nd that correctly) and that is LVM, and grub
can't > boot from LVM because grub boots the kernel and only that knows about >LVM. > So, > you are probably booting from hda8, but it's not in your fstab as >the /boot > partition.
What does a "df" say? Does it list hda8 among the partitions? Probably not?
[root@compaq1300 ~]# df Filesystem 1K-blocks Used Available Use% Mounted on /dev/mapper/VolGroup00-LogVol00 10696956 4597688 5547128 46% / /dev/hda3 102486 22174 75020 23% /boot tmpfs 257260 0 257260 0% /dev/shm [root@compaq1300 ~]#
Mount it and have a look at that partition, does it contain the same stuff as your /boot partition? If not mounted, do: mkdir /mnt/hda8 mount /dev/hda8 /mnt/hda8 cat /mnt/hda8/boot/grub/grub.conf Does this look like the grub.conf that is the *real* one booting your system?
[root@compaq1300 ~]# mkdir /mnt/hda8 [root@compaq1300 ~]# mount /dev/hda8 /mnt/hda8 [root@compaq1300 ~]# cat /mnt/hda8/boot/grub/grub.conf cat: /mnt/hda8/boot/grub/grub.conf: No such file or directory [root@compaq1300 ~]#
Kai: Before I got the above data this morning, I let PUP download/install the latest kernel (2.6.18-53.1.19.el5.i686) but after rebooting, it comes up with the original kernel that is on the CentOS 5 Install DVD I used last November. Not surprising that it does not boot this newest kernel. The download/install seemed to go perfectly, so the "Subject" changed from yum not updating the kernel to where is the proper boot file.... When it boots Linux, CentOS gives a message something like booting root (hd 0, 7). TIA, Lanny
lannyma@gmail.com wrote:
On 5/8/08, Kai Schaetzl <maillists AT conactive DOT com> wrote:
<snip> > hda3 and hda9 are your Linux LVM partitions, maybe they belong to one volume > group, I don't know (your fstab would tell more, there's also a graphical > frontend for LVM in your desktop). > > From your grub.conf we know that it thinks it's installed on (hd0,2), but > hd0,2 is hda3 (if I understand that correctly) and that is LVM, and grub > can't boot from LVM because grub boots the kernel and only that knows about > LVM. So, you are probably booting from hda8, but it's not in your fstab as > the /boot partition. > > What does a "df" say? Does it list hda8 among the partitions? Probably not?
[root@compaq1300 ~]# df Filesystem 1K-blocks Used Available Use% Mounted on /dev/mapper/VolGroup00-LogVol00 10696956 4597688 5547128 46% / /dev/hda3 102486 22174 75020 23% /boot tmpfs 257260 0 257260 0% /dev/shm [root@compaq1300 ~]#
Mount it and have a look at that partition, does it contain the same stuff as your /boot partition? If not mounted, do: mkdir /mnt/hda8 mount /dev/hda8 /mnt/hda8 cat /mnt/hda8/boot/grub/grub.conf Does this look like the grub.conf that is the *real* one booting your system?
[root@compaq1300 ~]# mkdir /mnt/hda8 [root@compaq1300 ~]# mount /dev/hda8 /mnt/hda8 [root@compaq1300 ~]# cat /mnt/hda8/boot/grub/grub.conf cat: /mnt/hda8/boot/grub/grub.conf: No such file or directory [root@compaq1300 ~]#
The proper location of the grub.conf is:
/mnt/hda8/grub/grub.conf
'boot' was the name of the mount point which isn't part of the 'boot' file system.
Kai: Before I got the above data this morning, I let PUP download/install the latest kernel (2.6.18-53.1.19.el5.i686) but after rebooting, it comes up with the original kernel that is on the CentOS 5 Install DVD I used last November. Not surprising that it does not boot this newest kernel. The download/install seemed to go perfectly, so the "Subject" changed from yum not updating the kernel to where is the proper boot file.... When it boots Linux, CentOS gives a message something like booting root (hd 0, 7). TIA, Lanny
Kai, is right though, chances are grub from the MBR is looking into a different partition for it's config and shows one of the problems with grub. I think there is a version of grub that will keep it's configs in the remaining sectors (sectors 2-62) of the first track and boot the kernels directly from another partition, but that's non-standard.
You could use a single 'boot' partition for all your Linux distros though, but make it bigger, say 256MB (or 512MB if you have a lot of distros installed).
I would typically have /dev/hda1 setup as a 256MB 'boot' and reuse it for other distros, just make sure not to format it on install or you'll bork the first distro's kernels!
-Ross
______________________________________________________________________ This e-mail, and any attachments thereto, is intended only for use by the addressee(s) named herein and may contain legally privileged and/or confidential information. If you are not the intended recipient of this e-mail, you are hereby notified that any dissemination, distribution or copying of this e-mail, and any attachments thereto, is strictly prohibited. If you have received this e-mail in error, please immediately notify the sender and permanently delete the original and any copy or printout thereof.
2984D980D4BD78EF9022A71CD@MFG-NYC-EXCH2.mfg.prv> X-Rcpt-To: centos@centos.org
Ross S. W. Walker wrote on Mon, 12 May 2008 10:52:16 -0400:
The proper location of the grub.conf is:
/mnt/hda8/grub/grub.conf
right, if that is the boot partition, there won't be a boot directory.
Kai
On Mon, 2008-05-12 at 10:52 -0400, Ross S. W. Walker wrote:
lannyma@gmail.com wrote:
On 5/8/08, Kai Schaetzl <maillists AT conactive DOT com> wrote:
<snip> > hda3 and hda9 are your Linux LVM partitions, maybe they belong to one volume > group, I don't know (your fstab would tell more, there's also a graphical > frontend for LVM in your desktop). > > From your grub.conf we know that it thinks it's installed on (hd0,2), but > hd0,2 is hda3 (if I understand that correctly) and that is LVM, and grub > can't boot from LVM because grub boots the kernel and only that knows about > LVM. So, you are probably booting from hda8, but it's not in your fstab as > the /boot partition. > > What does a "df" say? Does it list hda8 among the partitions? Probably not?
[root@compaq1300 ~]# df Filesystem 1K-blocks Used Available Use% Mounted on /dev/mapper/VolGroup00-LogVol00 10696956 4597688 5547128 46% / /dev/hda3 102486 22174 75020 23% /boot tmpfs 257260 0 257260 0% /dev/shm [root@compaq1300 ~]#
Mount it and have a look at that partition, does it contain the same stuff as your /boot partition? If not mounted, do: mkdir /mnt/hda8 mount /dev/hda8 /mnt/hda8 cat /mnt/hda8/boot/grub/grub.conf Does this look like the grub.conf that is the *real* one booting your system?
[root@compaq1300 ~]# mkdir /mnt/hda8 [root@compaq1300 ~]# mount /dev/hda8 /mnt/hda8 [root@compaq1300 ~]# cat /mnt/hda8/boot/grub/grub.conf cat: /mnt/hda8/boot/grub/grub.conf: No such file or directory [root@compaq1300 ~]#
The proper location of the grub.conf is:
/mnt/hda8/grub/grub.conf
'boot' was the name of the mount point which isn't part of the 'boot' file system.
<snip>
Kai, is right though, chances are grub from the MBR is looking into a different partition for it's config and shows one of the problems with grub. I think there is a version of grub that will keep it's configs in the remaining sectors (sectors 2-62) of the first track and boot the kernels directly from another partition, but that's non-standard.
You could use a single 'boot' partition for all your Linux distros though, but make it bigger, say 256MB (or 512MB if you have a lot of distros installed).
Ross: You suspect that I have more than one Linux distro installed, but that is not true. There are 2 OS installed: (a) MS Windows XP Home Edition (the installation of that did not go well on the box with this problem) and (b) CentOS 5. After I wiped the HDs in the 3 boxes, last Thanksgiving weekend, each of them got a /boot partition of approximately 100 MB. If you have any ideas that are non destructive, please let me know what they are. If this problem was on my box or my daughters box, worst case is I would "learn by destroying" and need to wipe the HD and start over. However, this is on my wife's box and if I screw it up, I have problems with her. :-) TIA, Lanny
I would typically have /dev/hda1 setup as a 256MB 'boot' and reuse it for other distros, just make sure not to format it on install or you'll bork the first distro's kernels!
-Ross
651k51947373y3956817f807f49d@mail.gmail.com> X-Rcpt-To: centos@centos.org
wrote on Mon, 12 May 2008 08:51:46 -0500:
Mount it and have a look at that partition, does it contain the same stuff as your /boot partition?
And this question?
Kai
651k51947373y3956817f807f49d@mail.gmail.com> X-Rcpt-To: centos@centos.org
wrote on Mon, 12 May 2008 08:51:46 -0500:
[root@compaq1300 ~]# df Filesystem 1K-blocks Used Available Use% Mounted on /dev/mapper/VolGroup00-LogVol00 10696956 4597688 5547128 46% / /dev/hda3 102486 22174 75020 23% /boot
I don't understand how this can go together with this partition table: /dev/hda3 5348 10586 39606840 8e Linux LVM /dev/hda8 3734 3747 105808+ 83 Linux /dev/hda9 3748 5347 12095968+ 8e Linux LVM
Your /dev/mapper/VolGroup00-LogVol00 might be on /dev/hda9 and the correct boot partition for it is probably /dev/hda8. Then you have /dev/hda3 which is another LVM partition, but which is not used by your installation at all (at least the small size of VolGroup00- LogVol00 suggests this). And at the same time your installation has mounted /dev/hda3 as a normal partition (although it is LVM) and uses it to install the kernel updates and thinks it's the boot partition. However, the /dev/hda3 that your system uses is about 100 MB while the /dev/hda3 of the partition table is roughly half the size of your whole disk and LVM managed. This all doesn't fit together. Ross thinks you have more than one distribution on that disk. That could indeed be an explanation. Did you do a repair or so? The twofold installation of Windows somehow hosed the booting and you tried to repair the system and somehow the boot partitions got mixed up or so? Do an "lvdisplay" and post some lines from it here, the LV Name and LV Size lines should be sufficient. And the output of pvdisplay.
Kai
On Mon, 2008-05-12 at 18:37 +0200, Kai Schaetzl wrote:
651k51947373y3956817f807f49d@mail.gmail.com> X-Rcpt-To: centos@centos.org
wrote on Mon, 12 May 2008 08:51:46 -0500:
[root@compaq1300 ~]# df Filesystem 1K-blocks Used Available Use% Mounted on /dev/mapper/VolGroup00-LogVol00 10696956 4597688 5547128 46% / /dev/hda3 102486 22174 75020 23% /boot
I don't understand how this can go together with this partition table: /dev/hda3 5348 10586 39606840 8e Linux LVM /dev/hda8 3734 3747 105808+ 83 Linux /dev/hda9 3748 5347 12095968+ 8e Linux LVM
Your /dev/mapper/VolGroup00-LogVol00 might be on /dev/hda9 and the correct boot partition for it is probably /dev/hda8. Then you have /dev/hda3 which is another LVM partition, but which is not used by your installation at all (at least the small size of VolGroup00- LogVol00 suggests this). And at the same time your installation has mounted /dev/hda3 as a normal partition (although it is LVM) and uses it to install the kernel updates and thinks it's the boot partition. However, the /dev/hda3 that your system uses is about 100 MB while the /dev/hda3 of the partition table is roughly half the size of your whole disk and LVM managed. This all doesn't fit together. Ross thinks you have more than one distribution on that disk. That could indeed be an explanation.
Kai: As I just replied to Ross, no, the only Linux distro on our boxes is CentOS 5.
Did you do a repair or so? The twofold installation of Windows somehow hosed the booting and you tried to repair the system and somehow the boot partitions got mixed up or so?
No repair was attempted.
Do an "lvdisplay" and post some lines from it here, the LV Name and LV Size lines should be sufficient. And the output of pvdisplay.
I will run those commands and give you the output, after she stops using the box. ASAP. TIA, Lanny
On 5/12/08, Kai Schaetzl <maillists AT conactive DOT com> wrote:
[root@compaq1300 ~]# df Filesystem 1K-blocks Used Available Use% Mounted on /dev/mapper/VolGroup00-LogVol00 10696956 4597688 5547128 46% / /dev/hda3 102486 22174 75020 23% /boot
I don't understand how this can go together with this partition table: /dev/hda3 5348 10586 39606840 8e Linux LVM /dev/hda8 3734 3747 105808+ 83 Linux /dev/hda9 3748 5347 12095968+ 8e Linux LVM
Your /dev/mapper/VolGroup00-LogVol00 might be on /dev/hda9 and the correct boot partition for it is probably /dev/hda8. Then you have /dev/hda3 which is another LVM partition, but which is not used by your installation at all (at least the small size of VolGroup00- LogVol00 suggests this). And at the same time your installation has mounted /dev/hda3 as a normal partition (although it is LVM) and uses it to install the kernel updates and thinks it's the boot partition. However, the /dev/hda3 that your system uses is about 100 MB while the /dev/hda3 of the partition table is roughly half the size of your whole disk and LVM managed. This all doesn't fit together. Ross thinks you have more than one distribution on that disk. That could indeed be an explanation. Did you do a repair or so? The twofold installation of Windows somehow hosed the booting and you tried to repair the system and somehow the boot partitions got mixed up or so? Do an "lvdisplay" and post some lines from it here, the LV Name and LV Size lines should be sufficient.
[root@compaq1300 ~]# lvdisplay --- Logical volume --- LV Name /dev/VolGroup00/LogVol00
LV Size 10.53 GB
And the output of pvdisplay.
[root@compaq1300 ~]# pvdisplay --- Physical volume --- PV Name /dev/hda9 VG Name VolGroup00 PV Size 11.54 GB / not usable 4.47 MB Allocatable yes (but full) PE Size (KByte) 32768 Total PE 369 Free PE 0 Allocated PE 369 PV UUID VT1z1b-Mjeu-Yaes-9jjv-FLz6-DYYl-6XbLOu
Thank you Kai! Lanny
53x73f2288el11aefa215d45923@mail.gmail.com> X-Rcpt-To: centos@centos.org
wrote on Tue, 13 May 2008 12:53:21 -0500:
LV Name /dev/VolGroup00/LogVol00
LV Size 10.53 GB
And the output of pvdisplay.
[root@compaq1300 ~]# pvdisplay --- Physical volume --- PV Name /dev/hda9 VG Name VolGroup00
Ok, that clarifies that the VG on /dev/hda3 is not in use at all and your CentOS is indeed installed on the LV on hda9. I still wonder how this mess was created. Did you have an earlier Linux installation on it and forgot to wipe that completely before you installed CentOS? I think the easiest way to get you back on track is to edit your fstab. There is a line about /boot in it that points to /dev/hda3. Change that to /dev/hda8. This is all. But before you do that, please check that there is indeed a grub.conf on it that contains the "old" information. You know the path I gave you was slightly wrong. Once you have confirmed that you can make the change to fstab (/etc/fstab). Be really careful when you do that as the wrong changes can make your system unbootable. Once the change is done the *next* kernel that gets installed will go to the correct boot partition and the correct grub.conf will be updated with the correct information to boot with the new kernel.
Kai
On Tue, 2008-05-13 at 21:11 +0200, Kai Schaetzl wrote:
53x73f2288el11aefa215d45923@mail.gmail.com> X-Rcpt-To: centos@centos.org
wrote on Tue, 13 May 2008 12:53:21 -0500:
LV Name /dev/VolGroup00/LogVol00
LV Size 10.53 GB
And the output of pvdisplay.
[root@compaq1300 ~]# pvdisplay --- Physical volume --- PV Name /dev/hda9 VG Name VolGroup00
Ok, that clarifies that the VG on /dev/hda3 is not in use at all and your CentOS is indeed installed on the LV on hda9. I still wonder how this mess was created. Did you have an earlier Linux installation on it and forgot to wipe that completely before you installed CentOS?
Kai: There was an earlier installation with Fedora Core or CentOS, and Windows 98 on it.. I think that I wiped the drive, before installing Win XP and CentOS 5, but that was almost 6 months ago and I am not 100% positive that I did in fact wipe the HD. For the earlier Linux installations, I did the partitioning manually, with Disk Druid.
I think the easiest way to get you back on track is to edit your fstab. There is a line about /boot in it that points to /dev/hda3. Change that to /dev/hda8. This is all. But before you do that, please check that there is indeed a grub.conf on it that contains the "old" information. You know the path I gave you was slightly wrong. Once you have confirmed that you can make the change to fstab (/etc/fstab). Be really careful when you do that as the wrong changes can make your system unbootable.
I will reread this entire thread and then keep my fingers crossed and do it. If I make her system unbootable, my wife will be unhappy......
Once the change is done the *next* kernel that gets installed will go to the correct boot partition and the correct grub.conf will be updated with the correct information to boot with the new kernel.
Thank you, for all of the time and expertise you have shared with me! Lanny
On Wed, May 7, 2008 at 3:27 AM, lannyma@gmail.com wrote:
Yes. I woke up about 430 this morning and I realized that rpm can locate the file by itself. :-) However, I want to learn how to use find that Mark (mhr) mentioned!
Lanny (offlist):
Thanks - I really did mean that in a good way.
The best way to learn to use find is to play with it. I have aliases and functions that use it for many, many purposes, especially for finding things in source code files.
I strongly suggest that you take the time to wade through the man page. Find is a very powerful command, and even if the man page is poorly laid out, there is a lot you can do with it. For this purpose, what you probably would need to do is this (asoot):
find / -name "kernel*.rpm"
It may take a while, but it will absolutely find every kernel-related rpm on your system.
HTH.
mhr
-
Kai Schaetzl -
Lanny Marcus -
lannyma＠gmail.com
-
MHR -
Michael Simpson -
Robert Nichols -
Ross S. W. Walker -
Scott Silva -
William L. Maltby