Sequential scrub and resilvers

Currently, scrubs and resilvers can take an extremely
long time to complete. This is largely due to the fact
that zfs scans process pools in logical order, as
determined by each block's bookmark. This makes sense
from a simplicity perspective, but blocks in zfs are
often scattered randomly across disks, particularly
due to zfs's copy-on-write mechanisms.

This patch improves performance by splitting scrubs
and resilvers into a metadata scanning phase and an IO
issuing phase. The metadata scan reads through the
structure of the pool and gathers an in-memory queue
of I/Os, sorted by size and offset on disk. The issuing
phase will then issue the scrub I/Os as sequentially as
possible, greatly improving performance.

This patch also updates and cleans up some of the scan
code which has not been updated in several years.

Reviewed-by: Brian Behlendorf <[email protected]>
Authored-by: Saso Kiselkov <[email protected]>
Authored-by: Alek Pinchuk <[email protected]>
Authored-by: Tom Caputi <[email protected]>
Signed-off-by: Tom Caputi <[email protected]>
Closes #3625 
Closes #6256 

1 files changed, 42 insertions, 5 deletions

diff --git a/include/sys/dsl_scan.h b/include/sys/dsl_scan.h
index 5303d9a69..7a29d9788 100644
--- a/include/sys/dsl_scan.h
+++ b/include/sys/dsl_scan.h
@@ -108,22 +108,56 @@ typedef enum dsl_scan_flags {
  */
 typedef struct dsl_scan {
 	struct dsl_pool *scn_dp;
-
-	boolean_t scn_suspending;
 	uint64_t scn_restart_txg;
 	uint64_t scn_done_txg;
 	uint64_t scn_sync_start_time;
-	zio_t *scn_zio_root;
+	uint64_t scn_issued_before_pass;
 
 	/* for freeing blocks */
 	boolean_t scn_is_bptree;
 	boolean_t scn_async_destroying;
 	boolean_t scn_async_stalled;
-	uint64_t scn_visited_this_txg;
 
-	dsl_scan_phys_t scn_phys;
+	/* flags and stats for controlling scan state */
+	boolean_t scn_is_sorted;	/* doing sequential scan */
+	boolean_t scn_clearing;		/* scan is issuing sequential extents */
+	boolean_t scn_checkpointing;	/* scan is issuing all queued extents */
+	boolean_t scn_suspending;	/* scan is suspending until next txg */
+	uint64_t scn_last_checkpoint;	/* time of last checkpoint */
+
+	/* members for thread synchronization */
+	zio_t *scn_zio_root;		/* root zio for waiting on IO */
+	taskq_t *scn_taskq;		/* task queue for issuing extents */
+
+	/* for controlling scan prefetch, protected by spa_scrub_lock */
+	boolean_t scn_prefetch_stop;	/* prefetch should stop */
+	zbookmark_phys_t scn_prefetch_bookmark;	/* prefetch start bookmark */
+	avl_tree_t scn_prefetch_queue;	/* priority queue of prefetch IOs */
+	uint64_t scn_maxinflight_bytes; /* max bytes in flight for pool */
+
+	/* per txg statistics */
+	uint64_t scn_visited_this_txg;	/* total bps visited this txg */
+	uint64_t scn_holes_this_txg;
+	uint64_t scn_lt_min_this_txg;
+	uint64_t scn_gt_max_this_txg;
+	uint64_t scn_ddt_contained_this_txg;
+	uint64_t scn_objsets_visited_this_txg;
+	uint64_t scn_avg_seg_size_this_txg;
+	uint64_t scn_segs_this_txg;
+	uint64_t scn_avg_zio_size_this_txg;
+	uint64_t scn_zios_this_txg;
+
+	/* members needed for syncing scan status to disk */
+	dsl_scan_phys_t scn_phys;	/* on disk representation of scan */
+	dsl_scan_phys_t scn_phys_cached;
+	avl_tree_t scn_queue;		/* queue of datasets to scan */
+	uint64_t scn_bytes_pending;	/* outstanding data to issue */
 } dsl_scan_t;
 
+typedef struct dsl_scan_io_queue dsl_scan_io_queue_t;
+
+void scan_init(void);
+void scan_fini(void);
 int dsl_scan_init(struct dsl_pool *dp, uint64_t txg);
 void dsl_scan_fini(struct dsl_pool *dp);
 void dsl_scan_sync(struct dsl_pool *, dmu_tx_t *);
@@ -142,6 +176,9 @@ void dsl_scan_ds_clone_swapped(struct dsl_dataset *ds1, struct dsl_dataset *ds2,
     struct dmu_tx *tx);
 boolean_t dsl_scan_active(dsl_scan_t *scn);
 boolean_t dsl_scan_is_paused_scrub(const dsl_scan_t *scn);
+void dsl_scan_freed(spa_t *spa, const blkptr_t *bp);
+void dsl_scan_io_queue_destroy(dsl_scan_io_queue_t *queue);
+void dsl_scan_io_queue_vdev_xfer(vdev_t *svd, vdev_t *tvd);
 
 #ifdef	__cplusplus
 }