Once before, alicloud’s ECS disk was full and couldn’t connect to the host computer. Later, it couldn’t be started up by clicking restart. The engineer who issued the work order to alicloud got the following reply:
Your server system will be stuck in the process of startup and cannot be started normally. When the system starts up, you can enter the system normally by setting the single user mode. When you enter the single user mode, you can check the system and find no exception. You try to cancel all the startup services under / etc / RC3. D / and restart the server. At present, the system startup problem has not been repaired.
Currently, it is in single user mode. You can also log in to the server from the console to check the startup service in the system.
You can back up the data in the server, consider reinitializing the system to the server, and restore to the original normal system startup.
I later recreated an ECS, using the same system version as the failed machine, adding a data disk, which was created from the previous backup snapshot.
After startup, fdisk – L looks at the partition and mounts it successfully with mount / dev / vdb1 / MNT（ It should be noted that the partition name of each system may be different. You can actually see it before mounting it.)
At this point, I have a basic idea. I backup the faulty machine, reload the system, and mount it with mount / dev / vdb1 / MNT.
Summarize the recovery process:
1. Backup snapshot
2. Refitting the system
3. Fdisk – L view partition
4. Mount data disk
mount /dev/vdb1 /mnt
(it should be noted that the partition name of each system may be different. You can actually see it before mounting it.)
5. Pay attention to setting the disk to mount automatically after power on
You can also open a new machine to mount the cloud disk created by snapshot.
In particular, my data is all on RDS, so I can retrieve the files on the data disk. The system doesn’t matter. It is necessary to separate data from web server, system from website file disk.