A ghetto file clustering package (and a way to learn Python).

Uses ssh + rsync to communicate; on startup, will read config.txt (or specified file), attempt to copy a clean version from the configured source config (if not local), then will basically execute "rsync source replica", and later build a JSON list of all sync'd files and stat(3) info. Compares the stat lists between source and replicae to determing what's complete, current, etc.

Most of this functionality could be reimplemented with a ~ dozen lines of shell.

source config: srchost:~/gc/config-test.txt ignore suffix: .DS_Store, .fseventsd, .Spotlight-V100, .DocumentRevisions-V100, .sync lazy write: 10 cycle: 21600 logfile: ~/gc/gc.log rsync options: timeout=180 testing: True

source: bucko:/home/austin/gc/test-data replica: mini-pi:/tmp/test-data

source: srchost:/path/to/files replica: dsthost1:/mnt/files replica: dsthost2:/mnt/files ignore suffix: mobile, index.html, streams rsync options: bwlimit=10m, delete-after

source: srchost:/path/to/morefiles replica: dsthost1:/mnt/morefiles rsync options: delete

source: srchost:/path/to/morefiles replica: dsthost2:/mnt/morefiles ignore suffix: doc