We have VM backups configured for GCP machines in our environment. Now we have the below scenario as we need to update custom hostname for the machines through terraform which needs deletion and recreation of VM. The approach we followed is as below.
1. Take FS backup of the machine
2. Delete and recreate the machine with custom hostname
3. Have the FS backup restored to the newly created machine so that we have the system back to the state it was before the deletion.
While trying this approach we are encountering many issues like losing connectivity to the machine, data which says restored on the console but not visible at the OS level and so on. We had raised a support case as well and got an update that we cannot do a full "/" restore as this will overwrite the kernel OS file which will crash the system during the restore phase and OS level files/directory you should not overwrite on a running machine (/usr /etc /lib /root /boot etc)
Could someone please let us know what would be the best approach in this scenario. Full VM restore and 1-touch backup restore are not feasible options.
The best way to perform a bare metal recovery is using a the 1 touch backup / restore method.
However as you mentioned 1 touch is not feasible, you may try the following manual method
Thank you for your response. However, over the link it says, do not do restore to “/” partition. However in our systems we have only “/” partition available to which we have to restore. That’s where we are looking for the solution
can you filter out the folders or files that make it crash when you do the restore?