The SQLite access is the limited factor if many small files must be backuped.
Here's a collection of ideas to accelerate the DB access...
Current PRAGMA statements are:
"PRAGMA journal_mode = MEMORY;",
"PRAGMA temp_store = MEMORY;",
"PRAGMA synchronous = OFF;"
see:
|
def setup_sqlite(sender, connection, **kwargs): |
bedup makes this:
# Uncripple the SQL implementation
cur.execute('PRAGMA foreign_keys = ON')
# So that writers do not block readers
# https://www.sqlite.org/wal.html
cur.execute('PRAGMA journal_mode = WAL')
see: https://github.com/g2p/bedup/blob/7283da47c3a2ffc01336389dae84f510caa034d9/bedup/__main__.py#L78
Maybe journal_mode = WAL will be faster then journal_mode = MEMORY ?
The SQLite access is the limited factor if many small files must be backuped.
Here's a collection of ideas to accelerate the DB access...
Current PRAGMA statements are:
see:
PyHardLinkBackup/PyHardLinkBackup/backup_app/models.py
Line 18 in 35d2d3b
bedup makes this:
see: https://github.com/g2p/bedup/blob/7283da47c3a2ffc01336389dae84f510caa034d9/bedup/__main__.py#L78
Maybe
journal_mode = WALwill be faster thenjournal_mode = MEMORY?