Resolution:

1).  Can an AS-persistence file created in one server be used for recovery in another server using same member name?

 - Yes . The same member name can be used.

2).  Can an AS-persistence file created in one server be used for recovery if using a different member name?

 - No, it cannot.

3).  Can we edit a persistence file to change SeederList?

 - No. AS does not provided any tools to edit SN files.

4).  Can we recreate a DR setup, replicating the present setup using a persistence file? This means that there is no initial data load, just recovery from a persistence file of running setup?

 - Yes you can but you have to use the same members names.  

5).  How do we ensure a persistence file is not corrupted?

- If you run the as-dump utility on an SN file, it will provide information such as EntryCount and InvalidBLockCount.

6).  Why do we see multiple persistence files?

- If the seeder has fully joined the space either after recovery or during a normal startup without any files, there should be only one current file present per seeder. However, if you do start, kill, start of agents....where AS does not see that agent/node as fully joined, then there can be multiple files which are created and initialized with a zero view number.

7).  On restart, a new persistence file is created and on recovery, the data is copied from an older persistence file. Once the copy is done, the older one gets deleted. During the copy process, whatever data is copied to the newer file, the same gets deleted from the older file. Is this understanding correct?

- It will create a new file and keep the old file, but the <ViewNumber> will be incremented. AS will always check for a higher <ViewNumber> and load data from that SN file. With the latest AS 2.1.6 version, AS will delete the old files and the .invalid file. 

8).  What are the conditions for an invalid file?

-------------------------
- If the seeder has fully joined the space either after recovery or during a normal startup without any files, then there should be only one current file present per seeder. However, if you do start, kill, start of agents....where AS does not see that agent/node as fully joined, then there can be multiple files which are created and initialized with a zero view number. When you start up again, then seeders with such files will rename them with the .INVALID extension. INVALID files could be either due to this repeated start/kill or they could also be created if some failure is encountered when trying to parse old files during recovery.

Such files are not deleted as they will be left in the event some sort of offline check could be done via as-dump/as-convert to get as much data out of invalid files as possible. A mechanism may be provided in the future to delete INVALID files, as well as upon startup. For now you can follow the manual procedure below to avoid creation of files with the .INVALID extension. (An enhancement request has been created for "option to clean up all invalid and old files after a successful recovery". )

        When you start / restart the as-agent, wait until the as-agent fully joins before you kill the same as-agent.
        Before you start the as-agent, make a backup of all invalid files and place them into a different folder.
        Start all as-agents and wait until the agent fully joins as "seeder".
        Check data consistency from as-admin for all spaces.
        If data is consistent then delete the old invalid files.

---------------------------   

9).  Can we take the persistent files of one metaspace and use them for loading some other metaspace? We have created two metaspaces called "MS1" and another metaspace called "MS2". We would like to use the persistent files of MS1 and load into MS2. This is  a different metaspace running different a set of servers.

- SN files have information about metaspace names on the <FilePath> tag. AS skips checking metaspace names while loading data from SN files. 

Frequently asked questions (FAQ) : Shared-nothing Persistence.