- turn off the Galaxy instance and make sure nothing is trying to read/write to the DB
- shut down postgresql using pg_ctl
pg_ctl stop -m smart
- add following parameters to postgresql.conf
test ! -f [path]
returns true if [path] does not exist
%f is placeholder for filename, %p is placeholder for path name of file to archive
wal_level = replica
archive_mode = on
archive_command = ' test ! -f /lustre/fs5/vgl/scratch/vgl_galaxy/pg_wal_backups/%f && cp %p /lustre/fs5/vgl/scratch/vgl_galaxy/pg_wal_backups/%f'
- create the base backup using pg_basebackup (backs up the entire database cluster, not just indiv databases like pg_dump would)
## need to double-check the name of the galaxy database........
pg_basebackup -h postgres -D /lustre/fs5/vgl/scratch/vgl_galaxy/pg_srvr_backups
- actually back up the data directory
???