OpenZFS 7090 - zfs should throttle allocations

OpenZFS 7090 - zfs should throttle allocations

Authored by: George Wilson <george.wilson@delphix.com>
Reviewed by: Alex Reece <alex@delphix.com>
Reviewed by: Christopher Siden <christopher.siden@delphix.com>
Reviewed by: Dan Kimmel <dan.kimmel@delphix.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Paul Dagnelie <paul.dagnelie@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Sebastien Roy <sebastien.roy@delphix.com>
Approved by: Matthew Ahrens <mahrens@delphix.com>
Ported-by: Don Brady <don.brady@intel.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>

When write I/Os are issued, they are issued in block order but the ZIO
pipeline will drive them asynchronously through the allocation stage
which can result in blocks being allocated out-of-order. It would be
nice to preserve as much of the logical order as possible.

In addition, the allocations are equally scattered across all top-level
VDEVs but not all top-level VDEVs are created equally. The pipeline
should be able to detect devices that are more capable of handling
allocations and should allocate more blocks to those devices. This
allows for dynamic allocation distribution when devices are imbalanced
as fuller devices will tend to be slower than empty devices.

The change includes a new pool-wide allocation queue which would
throttle and order allocations in the ZIO pipeline. The queue would be
ordered by issued time and offset and would provide an initial amount of
allocation of work to each top-level vdev. The allocation logic utilizes
a reservation system to reserve allocations that will be performed by
the allocator. Once an allocation is successfully completed it's
scheduled on a given top-level vdev. Each top-level vdev maintains a
maximum number of allocations that it can handle (mg_alloc_queue_depth).
The pool-wide reserved allocations (top-levels * mg_alloc_queue_depth)
are distributed across the top-level vdevs metaslab groups and round
robin across all eligible metaslab groups to distribute the work. As
top-levels complete their work, they receive additional work from the
pool-wide allocation queue until the allocation queue is emptied.

OpenZFS-issue: https://www.illumos.org/issues/7090
OpenZFS-commit: openzfs/openzfs@4756c3d7
Closes openzfs#5258 

Porting Notes:
- Maintained minimal stack in zio_done
- Preserve linux-specific io sizes in zio_write_compress
- Added module params and documentation
- Updated to use optimize AVL cmp macros

include/sys/fs/zfs.h

@@ -1038,7 +1038,8 @@ typedef enum {
 	SPA_LOAD_IMPORT,	/* import in progress	*/
 	SPA_LOAD_TRYIMPORT,	/* tryimport in progress */
 	SPA_LOAD_RECOVER,	/* recovery requested	*/
-	SPA_LOAD_ERROR		/* load failed		*/
+	SPA_LOAD_ERROR,		/* load failed		*/
+	SPA_LOAD_CREATE		/* creation in progress */
 } spa_load_state_t;
 
 /*

include/sys/metaslab.h

@@ -20,7 +20,7 @@
  */
 /*
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2011, 2014 by Delphix. All rights reserved.
+ * Copyright (c) 2011, 2015 by Delphix. All rights reserved.
  */
 
 #ifndef _SYS_METASLAB_H
@@ -55,15 +55,16 @@ void metaslab_sync_done(metaslab_t *, uint64_t);
 void metaslab_sync_reassess(metaslab_group_t *);
 uint64_t metaslab_block_maxsize(metaslab_t *);
 
-#define	METASLAB_HINTBP_FAVOR	0x0
-#define	METASLAB_HINTBP_AVOID	0x1
-#define	METASLAB_GANG_HEADER	0x2
-#define	METASLAB_GANG_CHILD	0x4
-#define	METASLAB_GANG_AVOID	0x8
-#define	METASLAB_FASTWRITE	0x10
+#define	METASLAB_HINTBP_FAVOR		0x0
+#define	METASLAB_HINTBP_AVOID		0x1
+#define	METASLAB_GANG_HEADER		0x2
+#define	METASLAB_GANG_CHILD		0x4
+#define	METASLAB_ASYNC_ALLOC		0x8
+#define	METASLAB_DONT_THROTTLE		0x10
+#define	METASLAB_FASTWRITE		0x20
 
 int metaslab_alloc(spa_t *, metaslab_class_t *, uint64_t,
-    blkptr_t *, int, uint64_t, blkptr_t *, int);
+    blkptr_t *, int, uint64_t, blkptr_t *, int, zio_t *);
 void metaslab_free(spa_t *, const blkptr_t *, uint64_t, boolean_t);
 int metaslab_claim(spa_t *, const blkptr_t *, uint64_t);
 void metaslab_check_free(spa_t *, const blkptr_t *);
@@ -76,6 +77,9 @@ int metaslab_class_validate(metaslab_class_t *);
 void metaslab_class_histogram_verify(metaslab_class_t *);
 uint64_t metaslab_class_fragmentation(metaslab_class_t *);
 uint64_t metaslab_class_expandable_space(metaslab_class_t *);
+boolean_t metaslab_class_throttle_reserve(metaslab_class_t *, int,
+    zio_t *, int);
+void metaslab_class_throttle_unreserve(metaslab_class_t *, int, zio_t *);
 
 void metaslab_class_space_update(metaslab_class_t *, int64_t, int64_t,
     int64_t, int64_t);
@@ -88,10 +92,13 @@ metaslab_group_t *metaslab_group_create(metaslab_class_t *, vdev_t *);
 void metaslab_group_destroy(metaslab_group_t *);
 void metaslab_group_activate(metaslab_group_t *);
 void metaslab_group_passivate(metaslab_group_t *);
+boolean_t metaslab_group_initialized(metaslab_group_t *);
 uint64_t metaslab_group_get_space(metaslab_group_t *);
 void metaslab_group_histogram_verify(metaslab_group_t *);
 uint64_t metaslab_group_fragmentation(metaslab_group_t *);
 void metaslab_group_histogram_remove(metaslab_group_t *, metaslab_t *);
+void metaslab_group_alloc_decrement(spa_t *, uint64_t, void *, int);
+void metaslab_group_alloc_verify(spa_t *, const blkptr_t *, void *);
 
 #ifdef	__cplusplus
 }

include/sys/metaslab_impl.h

@@ -24,7 +24,7 @@
  */
 
 /*
- * Copyright (c) 2011, 2014 by Delphix. All rights reserved.
+ * Copyright (c) 2011, 2015 by Delphix. All rights reserved.
  */
 
 #ifndef _SYS_METASLAB_IMPL_H
@@ -59,11 +59,42 @@ extern "C" {
  * to use a block allocator that best suits that class.
  */
 struct metaslab_class {
+	kmutex_t		mc_lock;
 	spa_t			*mc_spa;
 	metaslab_group_t	*mc_rotor;
 	metaslab_ops_t		*mc_ops;
 	uint64_t		mc_aliquot;
+
+	/*
+	 * Track the number of metaslab groups that have been initialized
+	 * and can accept allocations. An initialized metaslab group is
+	 * one has been completely added to the config (i.e. we have
+	 * updated the MOS config and the space has been added to the pool).
+	 */
+	uint64_t		mc_groups;
+
+	/*
+	 * Toggle to enable/disable the allocation throttle.
+	 */
+	boolean_t		mc_alloc_throttle_enabled;
+
+	/*
+	 * The allocation throttle works on a reservation system. Whenever
+	 * an asynchronous zio wants to perform an allocation it must
+	 * first reserve the number of blocks that it wants to allocate.
+	 * If there aren't sufficient slots available for the pending zio
+	 * then that I/O is throttled until more slots free up. The current
+	 * number of reserved allocations is maintained by the mc_alloc_slots
+	 * refcount. The mc_alloc_max_slots value determines the maximum
+	 * number of allocations that the system allows. Gang blocks are
+	 * allowed to reserve slots even if we've reached the maximum
+	 * number of allocations allowed.
+	 */
+	uint64_t		mc_alloc_max_slots;
+	refcount_t		mc_alloc_slots;
+
 	uint64_t		mc_alloc_groups; /* # of allocatable groups */
+
 	uint64_t		mc_alloc;	/* total allocated space */
 	uint64_t		mc_deferred;	/* total deferred frees */
 	uint64_t		mc_space;	/* total space (alloc + free) */
@@ -85,6 +116,15 @@ struct metaslab_group {
 	avl_tree_t		mg_metaslab_tree;
 	uint64_t		mg_aliquot;
 	boolean_t		mg_allocatable;		/* can we allocate? */
+
+	/*
+	 * A metaslab group is considered to be initialized only after
+	 * we have updated the MOS config and added the space to the pool.
+	 * We only allow allocation attempts to a metaslab group if it
+	 * has been initialized.
+	 */
+	boolean_t		mg_initialized;
+
 	uint64_t		mg_free_capacity;	/* percentage free */
 	int64_t			mg_bias;
 	int64_t			mg_activation_count;
@@ -93,6 +133,27 @@ struct metaslab_group {
 	taskq_t			*mg_taskq;
 	metaslab_group_t	*mg_prev;
 	metaslab_group_t	*mg_next;
+
+	/*
+	 * Each metaslab group can handle mg_max_alloc_queue_depth allocations
+	 * which are tracked by mg_alloc_queue_depth. It's possible for a
+	 * metaslab group to handle more allocations than its max. This
+	 * can occur when gang blocks are required or when other groups
+	 * are unable to handle their share of allocations.
+	 */
+	uint64_t		mg_max_alloc_queue_depth;
+	refcount_t		mg_alloc_queue_depth;
+
+	/*
+	 * A metalab group that can no longer allocate the minimum block
+	 * size will set mg_no_free_space. Once a metaslab group is out
+	 * of space then its share of work must be distributed to other
+	 * groups.
+	 */
+	boolean_t		mg_no_free_space;
+
+	uint64_t		mg_allocations;
+	uint64_t		mg_failed_allocations;
 	uint64_t		mg_fragmentation;
 	uint64_t		mg_histogram[RANGE_TREE_HISTOGRAM_SIZE];
 };

include/sys/refcount.h

@@ -20,6 +20,7 @@
  */
 /*
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2012, 2015 by Delphix. All rights reserved.
  */
 
 #ifndef	_SYS_REFCOUNT_H
@@ -61,6 +62,7 @@ typedef struct refcount {
 
 void refcount_create(refcount_t *rc);
 void refcount_create_untracked(refcount_t *rc);
+void refcount_create_tracked(refcount_t *rc);
 void refcount_destroy(refcount_t *rc);
 void refcount_destroy_many(refcount_t *rc, uint64_t number);
 int refcount_is_zero(refcount_t *rc);
@@ -71,6 +73,8 @@ int64_t refcount_add_many(refcount_t *rc, uint64_t number, void *holder_tag);
 int64_t refcount_remove_many(refcount_t *rc, uint64_t number, void *holder_tag);
 void refcount_transfer(refcount_t *dst, refcount_t *src);
 void refcount_transfer_ownership(refcount_t *, void *, void *);
+boolean_t refcount_held(refcount_t *, void *);
+boolean_t refcount_not_held(refcount_t *, void *);
 
 void refcount_init(void);
 void refcount_fini(void);
@@ -83,6 +87,7 @@ typedef struct refcount {
 
 #define	refcount_create(rc) ((rc)->rc_count = 0)
 #define	refcount_create_untracked(rc) ((rc)->rc_count = 0)
+#define	refcount_create_tracked(rc) ((rc)->rc_count = 0)
 #define	refcount_destroy(rc) ((rc)->rc_count = 0)
 #define	refcount_destroy_many(rc, number) ((rc)->rc_count = 0)
 #define	refcount_is_zero(rc) ((rc)->rc_count == 0)
@@ -99,6 +104,8 @@ typedef struct refcount {
 	atomic_add_64(&(dst)->rc_count, __tmp); \
 }
 #define	refcount_transfer_ownership(rc, current_holder, new_holder)	(void)0
+#define	refcount_held(rc, holder)		((rc)->rc_count > 0)
+#define	refcount_not_held(rc, holder)		(B_TRUE)
 
 #define	refcount_init()
 #define	refcount_fini()

include/sys/spa_impl.h

@@ -165,6 +165,8 @@ struct spa {
 	uint64_t	spa_last_synced_guid;	/* last synced guid */
 	list_t		spa_config_dirty_list;	/* vdevs with dirty config */
 	list_t		spa_state_dirty_list;	/* vdevs with dirty state */
+	kmutex_t	spa_alloc_lock;
+	avl_tree_t	spa_alloc_tree;
 	spa_aux_vdev_t	spa_spares;		/* hot spares */
 	spa_aux_vdev_t	spa_l2cache;		/* L2ARC cache devices */
 	nvlist_t	*spa_label_features;	/* Features for reading MOS */

include/sys/vdev_impl.h

@@ -53,6 +53,9 @@ typedef struct vdev_queue vdev_queue_t;
 typedef struct vdev_cache vdev_cache_t;
 typedef struct vdev_cache_entry vdev_cache_entry_t;
 
+extern int zfs_vdev_queue_depth_pct;
+extern uint32_t zfs_vdev_async_write_max_active;
+
 /*
  * Virtual device operations
  */
@@ -177,9 +180,20 @@ struct vdev {
 	uint64_t	vdev_deflate_ratio; /* deflation ratio (x512)	*/
 	uint64_t	vdev_islog;	/* is an intent log device	*/
 	uint64_t	vdev_removing;	/* device is being removed?	*/
-	boolean_t	vdev_ishole;	/* is a hole in the namespace 	*/
+	boolean_t	vdev_ishole;	/* is a hole in the namespace	*/
+	kmutex_t	vdev_queue_lock; /* protects vdev_queue_depth	*/
 	uint64_t	vdev_top_zap;
 
+	/*
+	 * The queue depth parameters determine how many async writes are
+	 * still pending (i.e. allocated by net yet issued to disk) per
+	 * top-level (vdev_async_write_queue_depth) and the maximum allowed
+	 * (vdev_max_async_write_queue_depth). These values only apply to
+	 * top-level vdevs.
+	 */
+	uint64_t	vdev_async_write_queue_depth;
+	uint64_t	vdev_max_async_write_queue_depth;
+
 	/*
 	 * Leaf vdev state.
 	 */

include/sys/zio.h

@@ -157,35 +157,36 @@ enum zio_flag {
 	ZIO_FLAG_DONT_CACHE	= 1 << 11,
 	ZIO_FLAG_NODATA		= 1 << 12,
 	ZIO_FLAG_INDUCE_DAMAGE	= 1 << 13,
+	ZIO_FLAG_IO_ALLOCATING  = 1 << 14,
 
 #define	ZIO_FLAG_DDT_INHERIT	(ZIO_FLAG_IO_RETRY - 1)
 #define	ZIO_FLAG_GANG_INHERIT	(ZIO_FLAG_IO_RETRY - 1)
 
 	/*
 	 * Flags inherited by vdev children.
 	 */
-	ZIO_FLAG_IO_RETRY	= 1 << 14,	/* must be first for INHERIT */
-	ZIO_FLAG_PROBE		= 1 << 15,
-	ZIO_FLAG_TRYHARD	= 1 << 16,
-	ZIO_FLAG_OPTIONAL	= 1 << 17,
+	ZIO_FLAG_IO_RETRY	= 1 << 15,	/* must be first for INHERIT */
+	ZIO_FLAG_PROBE		= 1 << 16,
+	ZIO_FLAG_TRYHARD	= 1 << 17,
+	ZIO_FLAG_OPTIONAL	= 1 << 18,
 
 #define	ZIO_FLAG_VDEV_INHERIT	(ZIO_FLAG_DONT_QUEUE - 1)
 
 	/*
 	 * Flags not inherited by any children.
 	 */
-	ZIO_FLAG_DONT_QUEUE	= 1 << 18,	/* must be first for INHERIT */
-	ZIO_FLAG_DONT_PROPAGATE	= 1 << 19,
-	ZIO_FLAG_IO_BYPASS	= 1 << 20,
-	ZIO_FLAG_IO_REWRITE	= 1 << 21,
-	ZIO_FLAG_RAW		= 1 << 22,
-	ZIO_FLAG_GANG_CHILD	= 1 << 23,
-	ZIO_FLAG_DDT_CHILD	= 1 << 24,
-	ZIO_FLAG_GODFATHER	= 1 << 25,
-	ZIO_FLAG_NOPWRITE	= 1 << 26,
-	ZIO_FLAG_REEXECUTED	= 1 << 27,
-	ZIO_FLAG_DELEGATED	= 1 << 28,
-	ZIO_FLAG_FASTWRITE	= 1 << 29,
+	ZIO_FLAG_DONT_QUEUE	= 1 << 19,	/* must be first for INHERIT */
+	ZIO_FLAG_DONT_PROPAGATE	= 1 << 20,
+	ZIO_FLAG_IO_BYPASS	= 1 << 21,
+	ZIO_FLAG_IO_REWRITE	= 1 << 22,
+	ZIO_FLAG_RAW		= 1 << 23,
+	ZIO_FLAG_GANG_CHILD	= 1 << 24,
+	ZIO_FLAG_DDT_CHILD	= 1 << 25,
+	ZIO_FLAG_GODFATHER	= 1 << 26,
+	ZIO_FLAG_NOPWRITE	= 1 << 27,
+	ZIO_FLAG_REEXECUTED	= 1 << 28,
+	ZIO_FLAG_DELEGATED	= 1 << 29,
+	ZIO_FLAG_FASTWRITE	= 1 << 30
 };
 
 #define	ZIO_FLAG_MUSTSUCCEED		0
@@ -225,6 +226,7 @@ enum zio_wait_type {
 
 typedef void zio_done_func_t(zio_t *zio);
 
+extern int zio_dva_throttle_enabled;
 extern const char *zio_type_name[ZIO_TYPES];
 
 /*
@@ -379,7 +381,6 @@ struct zio {
 	blkptr_t	io_bp_copy;
 	list_t		io_parent_list;
 	list_t		io_child_list;
-	zio_link_t	*io_walk_link;
 	zio_t		*io_logical;
 	zio_transform_t *io_transform_stack;
 
@@ -407,12 +408,14 @@ struct zio {
 
 	uint64_t	io_offset;
 	hrtime_t	io_timestamp;	/* submitted at */
+	hrtime_t	io_queued_timestamp;
 	hrtime_t	io_target_timestamp;
 	hrtime_t	io_delta;	/* vdev queue service delta */
 	hrtime_t	io_delay;	/* Device access time (disk or */
 					/* file). */
 	avl_node_t	io_queue_node;
 	avl_node_t	io_offset_node;
+	avl_node_t	io_alloc_node;
 
 	/* Internal pipeline state */
 	enum zio_flag	io_flags;
@@ -421,6 +424,7 @@ struct zio {
 	enum zio_flag	io_orig_flags;
 	enum zio_stage	io_orig_stage;
 	enum zio_stage	io_orig_pipeline;
+	enum zio_stage	io_pipeline_trace;
 	int		io_error;
 	int		io_child_error[ZIO_CHILD_TYPES];
 	uint64_t	io_children[ZIO_CHILD_TYPES][ZIO_WAIT_TYPES];
@@ -443,6 +447,8 @@ struct zio {
 	taskq_ent_t	io_tqent;
 };
 
+extern int zio_timestamp_compare(const void *, const void *);
+
 extern zio_t *zio_null(zio_t *pio, spa_t *spa, vdev_t *vd,
     zio_done_func_t *done, void *private, enum zio_flag flags);
 
@@ -502,8 +508,8 @@ extern void zio_interrupt(zio_t *zio);
 extern void zio_delay_init(zio_t *zio);
 extern void zio_delay_interrupt(zio_t *zio);
 
-extern zio_t *zio_walk_parents(zio_t *cio);
-extern zio_t *zio_walk_children(zio_t *pio);
+extern zio_t *zio_walk_parents(zio_t *cio, zio_link_t **);
+extern zio_t *zio_walk_children(zio_t *pio, zio_link_t **);
 extern zio_t *zio_unique_parent(zio_t *cio);
 extern void zio_add_child(zio_t *pio, zio_t *cio);