I am running clearos for sometime at a school for a DC, 4 days ago something happened on this server, the boot partition is unknown, if I try to rescue the system , tried vgdisplay etc and no success, the UUID name doesn't come up in vgscan for the boot drive, Please anyone , HELP\
Share this post:
Responses (8)
-
Accepted Answer
I have a feeling you're going to be into disk recovery. I also suspect /dev/sda1 is not part of a logical volume. In the past, a lot of searches have pointed to TestDisk as the free tool of choice. As a general principle you should obtain a spare disk of equal or greater size than you problem disk then dd all the old disk onto the new disk. (I think there is a better command than dd but I don't remember it and dd is not a good search term). Then you do all your recovery operations on the new disk.
If you have a third disk, if the worst comes to the worst you may have to reformat sda1. In that case do another installation of ClearOS onto a third disk and you may then be able to copy /boot (/dev/sda1) from your third disk onto the copy of the disk you are salvaging. You may then need to boot a ClearOS recovery disk and reinstall grub (I don't know). Lastly you'll need to edit /etc/fstab to point to the new UUID of /dev/sda1 as formatting it may well scrub the original UUID ot you could change the entru "UUID=....." to /dev/sda1 temporarily.
This assumes it is just /dev/sda1 which is corrupt. -
Accepted Answer
I just subscribed my News-server (which runs leafnode) to the Newsgroup "gmane.linux.lvm.general" and the content quality seems good. This is a copy of the mailing list "linux-lvm@redhat.com". see https://www.redhat.com/mailman/listinfo/linux-lvm The LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/ might also be helpful. -
Accepted Answer
No sorry, I don't know the experts as I do not use LVM on my production machines. One tip though, put yourself in the position of the reader...
/dev/sda1 starts at 1 and ends at 64
1 to 64 what? - I am guessing cylinders rather than blocks (that depends if you use the -u option of fdisk or not), as 64 cylinders is a more reasonable value. But, what size are YOUR cylinders - can be different depending on the disk mapping!
I assume "system linux" means it has a partition id of 83? If so, then you should be able to both run fsck against it with the -n option (for safety - no changes - just report) as well as mounting it and looking inside the partition to see what files are there. I assume this is your /boot partition? -
Accepted Answer
Tony Ellis wrote:
I second what Nick has just said. My ClearOS 7.2 /boot partition is 135 MB and can only contain two kernels. I found out the hard way when I tried to install a third. It failed during the rebuilding of the initramfs for extras drivers - disk full. This necessitated rebooting the oldest kernel which fortunately hadn't been corrupted, cleaning out the remnants of the latest kernel plus the other kernel using yum remove kernel 3.10.0-xxx.yy.z.v7.x86_64 commands. A self-induced problem as I did a custom install and choose the size.
I haven't used LVM since the hey-days of IBM's OS/2, therefore much of my knowledge is either forgotten or out-of-date. A quick check with google revealed a plethora of recovery hits for LVM, some rather hairy, others reasonable. There are also mailing lists specifically dedicated to LVM. No idea of quality of content though. If you do not get a solution here, I suggest you find out where the LVM experts hang-out, and ask there.
Thank you Tony, Much appreciated, If you know where these experts hang out , Please let me know asap, Please -
Accepted Answer
I second what Nick has just said. My ClearOS 7.2 /boot partition is 135 MB and can only contain two kernels. I found out the hard way when I tried to install a third. It failed during the rebuilding of the initramfs for extras drivers - disk full. This necessitated rebooting the oldest kernel which fortunately hadn't been corrupted, cleaning out the remnants of the latest kernel plus the other kernel using yum remove kernel 3.10.0-xxx.yy.z.v7.x86_64 commands. A self-induced problem as I did a custom install and choose the size.
I haven't used LVM since the hey-days of IBM's OS/2, therefore much of my knowledge is either forgotten or out-of-date. A quick check with google revealed a plethora of recovery hits for LVM, some rather hairy, others reasonable. There are also mailing lists specifically dedicated to LVM. No idea of quality of content though. If you do not get a solution here, I suggest you find out where the LVM experts hang-out, and ask there. -
Accepted Answer
-
Accepted Answer
-
Accepted Answer
I don't know much about this, but I don't think the /boot partition is covered by the LVM (vgdisplay?). You could try:
This should list all your (formatted?) partitions.ls -l /dev/disk/by-uuid
One thing I get a bit concerned about is kernel updates. The /boot partition is not very big by default (100MB?) and I would have thought too many kernel updates will fill the partition. I always do my kernel updates manually for various reasons including this.
Please login to post a reply
You will need to be logged in to be able to post a reply. Login using the form on the right or register an account if you are new here.
Register Here »