While standing up our own self-hosted version of WebKNOSSOS, we have been looking into data integrity/site reliability -- mostly, how to efficiently backup and restore our systems to a certain point in time.
While this is not a blocker (we intend to routinely export the data off the EC2 instance to S3, etc.), it could become a useful convenience for users, and to quickly get systems back online if multiple backups can live next door to running instances.
Let us know -- we are happy to attempt a PR if this is something that could be useful -- otherwise, if you have insight into how you rountinely backup and restore FossilDB instances, that would be much appreciated!
We typically run mv, scp or equivalent after calling backup. The entire backup is contained in the backup folder, so it is easy to move around and to keep multiple versions available.
Detailed Description
While standing up our own self-hosted version of WebKNOSSOS, we have been looking into data integrity/site reliability -- mostly, how to efficiently backup and restore our systems to a certain point in time.
Via the fossildb-client, the backup and restore commands work as intended -- as provided here; however, upon
backup
, we observe that the most recent backup is overwritten, and that no option is provided for where to point a newrestore
command (if multiple backups existed on the file system)Specifically for the
restore
command, further logic could be available upon theprotobuf
extension ofRestoreFromBackupRequest
While this is not a blocker (we intend to routinely export the data off the EC2 instance to S3, etc.), it could become a useful convenience for users, and to quickly get systems back online if multiple backups can live next door to running instances.
Let us know -- we are happy to attempt a PR if this is something that could be useful -- otherwise, if you have insight into how you rountinely backup and restore
FossilDB
instances, that would be much appreciated!Cc @kabilar