Clarify Checkpointing Behavior in v0.5.2

[michaelmedellin]

I'm using v0.5.2 as a library within a go server side application I'm developing and I'm seeing unbounded WAL size growth. I was using 0.3.x previously, which handled checkpointing the WAL file automatically and assumed that would be the same behavior in 0.5.x, but based on my review of the code I'm not sure automatic checkpointing within Litestream is happening.

litestream/db.go

Line 586 in 9be9408

// TODO(ltx): Move checkpointing into its own goroutine?

/*
		// TODO(ltx): Move checkpointing into its own goroutine?

		// If WAL size is greater than max threshold, force checkpoint.
		// If WAL size is greater than min threshold, attempt checkpoint.
		var checkpoint bool
		checkpointMode := CheckpointModePassive
		if db.TruncatePageN > 0 && origWALSize >= calcWALSize(db.pageSize, db.TruncatePageN) {
			checkpoint, checkpointMode = true, CheckpointModeTruncate
		} else if db.MaxCheckpointPageN > 0 && newWALSize >= calcWALSize(db.pageSize, db.MaxCheckpointPageN) {
			checkpoint, checkpointMode = true, CheckpointModeRestart
		} else if newWALSize >= calcWALSize(db.pageSize, db.MinCheckpointPageN) {
			checkpoint = true
		} else if db.CheckpointInterval > 0 && !info.dbModTime.IsZero() && time.Since(info.dbModTime) > db.CheckpointInterval && newWALSize > calcWALSize(db.pageSize, 1) {
			checkpoint = true
		}

		// Issue the checkpoint.
		if checkpoint {
			changed = true

			if err := db.checkpoint(ctx, checkpointMode); err != nil {
				return fmt.Errorf("checkpoint: mode=%v err=%w", checkpointMode, err)
			}
		}
	*/

In version 0.3.x this code was uncommented and I believe the only execution call for checkpointing when using the replicate command.

Perhaps I'm missing something, but I'm looking for confirmation on whether or not I should expect to see checkpointing via Litestream in 0.5.x

[michaelmedellin]

Tracking this: #824

Thanks!