Cluster server do big data, there are 10 units, then starts in the cluster, the original engineers remote guidance I check on the implementation of the cluster environment, found behind the management node of a device appeared abnormal SSSD service, with the help of the original implementation engineer failed to solve, then told me to call 400 for help, but because I am in the process of operation, the execution of the rm - rf/var/lib/SSS/db/cache_xxx LDB command, accidentally click fast enter and lose a rm - rf/var/lib/SSS/db this command, leading to SSS under directory are deleted, and then 400 a management node from another copy to come over, after copy to start the service again, still can't start, finally problem cannot be solved,
Then find the leadership to deal with this problem, several big data is mainly responsible for leadership, as well as the project manager is responsible for, or even project account manager to help solve, don't give a complete solution, big data, head of the mean call I go to look for Linux is the manufacturer to solve this problem, I called the red hat Linux customer service, they said the customer didn't buy services or products, technical support, in the end the problem still not solved, for help, now, does anyone encountered such a situation, how to solve ah!!!!!!
Project is currently in the acceptance stage, really can not be an accident ah,,,, every brother please,
CodePudding user response:
I understand what you want is to restore the deleted directory,To shut down, prevent covered by deleting files are the new data,
Next, does not guarantee that can restore,
1, disk pull down, mounted on the other computer
2, determine the directory partition, fdisk -l, df -h, LSBLK etc can all be
3, the use of debugfs, & gt; Said to a command prompt at the back of the debugfs
The open/dev/sda1. (if the sda1 partition)
Ls -d/path/to/deleted_directory (& lt;> The number is the inode information)
Logdump -i & lt; Inode> (remember printed information)
The quit
4, dd if=/dev/sda1 of=/path/to/restore bs=offset count=1 skip=block
After the restoration of is path, bs is logdump offset, skip is logdump block
But if there is a condition, can use third-party tools testdisk, according to its tip steps after installation,
CodePudding user response:
Remind, you is it the LVM partition, type of LVM partition the above methods can not,Everything, all want to good again decided to do not do, data recovery is not easy,