How can you add a custom file attribute based on some logic at the time a file is originally written or while it's being replicated?
I have a use case where I need to replicate files from one data center to others. Files land in one arbitrary data center via sftp and ftps servers which write the files to Maprfs over NFS. The files are then replicated to other data centers.
I want to tag the files with Linux custom file attributes when they first land; or by hooking into the replication process. This way, when looking at a folder, processes can discern the original data center each file landed in.
Not sure whether my file tagging approach is the best or even doable (not finding replication options in online documentation) at this point.
Note that mapr is a commercial version of Hadoop.