hi all,

I'll not bother you on the usage of s3fs and other pieces of software. 
For having tested this, keeping s3 + s3fs + ec2 + storage daemon is not 
reliable as it should be for backup (!) purpose.

Proposed scenario: transfer the volume after write is complete, and 
delete after successful move.

- set the volume to not recycle nor reuse
- move the volume once completed to s3 bucket
- let the job expiry and set up a policy on the S3 objects for delete or 
move to glacier after specified time.
- copy back (only the needed ones) to storage daemon for restore from S3

After some experience, S3 is not reliable as a filesystem, and is 
virtually infinite.
So it would be better to think in a different way.

Does the above make sense or should I be kicked off from the list at my 
first post?

thanks
Ant





------------------------------------------------------------------------------
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users