CentOS Grub, BASH fault, solution, centosbash

CentOS Grub, BASH fault, solution, centosbash

Introduction:

Two common faults of Grub: Grub. conf file loss and MBR damage (no matter what the recovery is, back up the file first)

1. Grub. conf file loss

Shell> rm-rf/boot/grub. conf # delete files, simulate faults, shell> reboot

# Faults:

GNU GRUB verstion 0.97 (634k lower / 1046400k upper memory )[ Minimal BASH-like line editing is supported. For the first word, TAB lists possible command completions. Anywhere else TAB lists the possible completions of a device/filename. ]grub>

# In this way, you cannot log on to the system. The solution is as follows:

Grub> root (hd0,) # Enter root (hd0,). Then press the TAB key to print all the partitions in the system (note that TAB is not supported in emergency mode, which is very harmful ~ ) Possible partitions are: Partition num: 0, Filesystem type is ext2fs, partition type 0x83Partition num: 1, Filesystem type unknown, partition type 0x82Partition num: 2, Filesystem type is ext2fs, partition type 0x83

#0, 1, 2 indicates the partition number. The above shows the file system and partition type. For example, 83 is a common Linux partition, and 82 is a virtual partition SWAP (hard disk partition knowledge)
# Hd0 indicates the first hard disk, hd0 indicates the first partition of the first hard disk, regardless of IDE/SATA, in the system, sda1 is used to represent the first partition of the first hard disk.

Grub> root (hd0, 0)/# Enter root (hd0, 0)/click the TAB key here to search for the kernel not in this partition (boot partition ), if the following information is output (try 83 partitions and 82 is a swap partition, it is unnecessary to try) Possible files are: lost + found grub efi. VMlinuz-2.6.32-504.e16.x86_64.hmac System. map-2.6.32-504.e16.x86_64 config-2.6.32-504.x86_64 symvers-2.6.32-504.x86_64.gz VMlinuz-2.6.32-504.e16.x86_64

# If you can see the vmlinuz-And initramfs-words, this is what we want

Grub> root (hd0, 2)/# The following cannot be found. Take a closer look (it is found that this is the root partition Ah ~~ Hd0 and 2 are root partitions. We use SATA, so the root partition is/dev/sda3 !) Possible files are: lost + found boot dev proc sys var tmp etc root selinux lib64 usr bin home lib media mnt opt sbin srv. autorelabelgrub> root (hd0, 0) # bind the boot partition Filesystem type is ext2fs, partition type 0x83grub> kernel/VMlinuz-2.6.32-504.e16.x86_64 ro root =/dev/sda3 # specify the boot kernel, ro read-only mode, root = specifies the root partition, which must be a pair (it is possible if the boot does not have a separate partition not specified) grub> initrd/initramfs-2.6.32-504.e16.x86_64.img # This is the kernel image (kernel, initrd are found above, enter a little press TAB can complete) grub> boot # Start the system, manually write grub. conf or restore the backup shell> vim/boot/grub. confdefault = 0 # Start which system timeout = 5 # Wait time title CentOS 6.6 # description information kernel/VMlinuz-2.6.32-504.e16.x86_64 ro root =/dev/sda3 # kernel initrd/initramfs-2.6.32-504.e16.x86_64.img # Start Image

# Here is a vim usage. On the host, vim cannot copy the command line content with the mouse. the kernel and initrd are so long and hard to remember.
# Method: Input in last line mode .! Ls/boot/vmlinuz-press TAB to complete the information. Press Enter to put the information in the current file.

shell > reboot

# In this way, the recovery is successful! Remember to back up

# If the kernel does not refer to the root partition, this will happen ~ Restart the system and try again.

dracut Warning: Signal caught!?dracut Warning: Boot has faild. To debug this issue add "rdshell" to the kernel command line.?Kernel panic - not syncing: Attempted to kill init!?Pid: 1, comm: init Not tainted 2.6.32-504.e16.x86_64 #1?Call Trace:[<ffffffff815292bc] ? panic+0xa7/0x16f? [<ffffffff815292bc] ? do_exit+0xa7/0x16f? [<ffffffff815292bc] ? fput+0xa7/0x16f? [<ffffffff815292bc] ? system_call_fastpath+0xa7/0x16f??

Ii. MBR damage

Shell> dd if =/dev/zero of =/dev/sda count = 1 bs = 200 # overwrite MBR, simulate a fault (MBR is the first 512 bytes in/dev/sda) shell> reboot

# Faults:

Operating System not found # basically, the System cannot be found on the local hard disk, obtained from PXE, DHCP xxxxxxxxxxxxx

# Solution: Use emergency recovery

> Insert the system disc and enter the emergency mode (this will be the first aid mode. Choose CentOS directly, RedHat boot: linux rescue)

Shell> grub # Go to grub mode grub> root (hd0, 0) # directly specify the partition where the kernel is located and cannot TAB (you must find a way to find it, for example: first, chroot/mnt/sysimage, and then df-HT can see all the partitions. If sda1 is the start partition, then the corresponding partition is hd0, 0.) grub> setup (hd0) # re-install MBR to the disk, which is actually/dev/sdagrub> quitshell> reboot

# Now you can successfully access the system. The MBR has been repaired.

Additional: BASH fault solution

Shell> rm-rf/bin/bash # Delete bash, simulate fault shell> reboot

# Failure: the system cannot be initialized, And the init command does not exist. In fact, no command can be executed.
# Enter Emergency mode and reinstall bash

Shell> mkdir/media # directly operate on the disc system, because you cannot chroot to/mnt/sysimageshell> mount/dev/sr0/media # mount the disc image, /dev/sr0 indicates the optical drive, if the IDE interface should be/dev/hdcshell> cd/media/Packagesshell> rpm-ivh -- replacepkgs -- root/mnt/sysimage bash-4.1.2-29.e16.x86_64.rpm # Install bash to -- root to specify the root directory -- replacepkgs replaces the original package shell> reboot

# OK, the fault has been fixed!