Hello list.
Since morning I have an emergency. After last kernel update one of my servers didn't start. On display I see only "GRUB" :( I have 2 more servers based on CentOS too, but after the same update this servers works.
All servers have 2 HDs and works as software mirror.
I have started damaged machine using Knoppix and I can see partitions and data on both disks. Of course I have made a backup.
Anybody has the same problem? It can be hardware problem - HD diode light all the time. Or maybe something went wrong during update?
Regards
On Mon, 2006-05-29 at 13:02 +0200, Dominik Składanowski wrote:
Hello list.
Since morning I have an emergency. After last kernel update one of my servers didn't start. On display I see only "GRUB" :( I have 2 more servers based on CentOS too, but after the same update this servers works.
All servers have 2 HDs and works as software mirror.
I have started damaged machine using Knoppix and I can see partitions and data on both disks. Of course I have made a backup.
Anybody has the same problem? It can be hardware problem - HD diode light all the time. Or maybe something went wrong during update?
It is very strange that grub (which was already installed and working) would go bad during an upgrade.
I would boot from a CentOS-4.3 CD-1 and do this command at the isolinux prompt:
linux rescue
If it mounts the install at /mnt/sysimage ... do this:
chroot /mnt/sysimage
Then I would follow this guide to make sure that grub is installed on both hard drives and try rebooting:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=114690
(comment #7 should make you work ... assuming you have a software mirror using sda/sdb or hda/hdb)
You may need to look at the file /boot/grub/Device.map to insure all the drives are listed properly.
Thanks, Johnny Hughes
2006/5/29, Johnny Hughes mailing-lists@hughesjr.com:
Since morning I have an emergency. After last kernel update one of my servers didn't start. On display I see only "GRUB" :( I have 2 more servers based on CentOS too, but after the same update this servers works.
All servers have 2 HDs and works as software mirror.
I have started damaged machine using Knoppix and I can see partitions and data on both disks. Of course I have made a backup.
Anybody has the same problem? It can be hardware problem - HD diode light all the time. Or maybe something went wrong during update?
It is very strange that grub (which was already installed and working) would go bad during an upgrade.
I would boot from a CentOS-4.3 CD-1 and do this command at the isolinux prompt:
linux rescue
If it mounts the install at /mnt/sysimage ... do this:
chroot /mnt/sysimage
Then I would follow this guide to make sure that grub is installed on both hard drives and try rebooting:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=114690
(comment #7 should make you work ... assuming you have a software mirror using sda/sdb or hda/hdb)
You may need to look at the file /boot/grub/Device.map to insure all the drives are listed properly.
Thanks for answer Johnny. It was exactly problem described above. I have reinstalled grub using rescue evironment.
Thanks.
Hi all,
i've just updated CentOS4.3 . After updating it i rebooted the machine (cause there was some kernel pkg's) and it seems that centos lost my LVM config .... I tried manually a vgscan and: (root@backupsrv) boot # vgscan connect() failed on local socket: Connection refused Locking type 2 initialisation failed.
and my /dev/VG0/LogVol00 dissapeared :-(
Any tips ?
Nicholas Anderson wrote:
Hi all,
i've just updated CentOS4.3 . After updating it i rebooted the machine (cause there was some kernel pkg's) and it seems that centos lost my LVM config .... I tried manually a vgscan and: (root@backupsrv) boot # vgscan connect() failed on local socket: Connection refused Locking type 2 initialisation failed.
and my /dev/VG0/LogVol00 dissapeared :-(
Any tips ?
I'm running cluster and got the same error. Go back to 2.6.9-34.
Or remove lvm2-cluster ?
Bob Marcan wrote:
Nicholas Anderson wrote:
Hi all,
i've just updated CentOS4.3 . After updating it i rebooted the machine (cause there was some kernel pkg's) and it seems that centos lost my LVM config .... I tried manually a vgscan and: (root@backupsrv) boot # vgscan connect() failed on local socket: Connection refused Locking type 2 initialisation failed.
and my /dev/VG0/LogVol00 dissapeared :-(
Any tips ?
I'm running cluster and got the same error. Go back to 2.6.9-34.
Or remove lvm2-cluster ?
Hi,
i've just solved the problem :-P up2date has changed my lvm.conf .... I'm new to Centos, and i didnt know that it saves a copy of the replaced file .... im my case it was the lvm.conf.
The new file (called "bad" :-) ) uses type 2 of locking_type ....
below is a short version (cutting the comments and line numbers) of a diff between both files ....
(root@backupsrv) lvm # diff -ru lvm.conf lvm.conf.bad
+ library_dir = "/usr/lib" + locking_library = "liblvm2clusterlock.so"
- locking_type = 1 + locking_type = 2
Thanks anyway,
Nick
-
Bob Marcan -
Dominik Składanowski -
Johnny Hughes -
Nicholas Anderson