c/mirror_zfs
1
0
Fork 0
mirror of https://git.proxmox.com/git/mirror_zfs.git synced 2024-12-27 03:19:35 +03:00
Code Issues Packages Projects Releases Wiki Activity
69c7bdb6e7
mirror_zfs/include/sys/dbuf.h
Matthew Ahrens e8b96c6007 Illumos #4045 write throttle & i/o scheduler performance work 
4045 zfs write throttle & i/o scheduler performance work

1. The ZFS i/o scheduler (vdev_queue.c) now divides i/os into 5 classes: sync
read, sync write, async read, async write, and scrub/resilver.  The scheduler
issues a number of concurrent i/os from each class to the device.  Once a class
has been selected, an i/o is selected from this class using either an elevator
algorithem (async, scrub classes) or FIFO (sync classes).  The number of
concurrent async write i/os is tuned dynamically based on i/o load, to achieve
good sync i/o latency when there is not a high load of writes, and good write
throughput when there is.  See the block comment in vdev_queue.c (reproduced
below) for more details.

2. The write throttle (dsl_pool_tempreserve_space() and
txg_constrain_throughput()) is rewritten to produce much more consistent delays
when under constant load.  The new write throttle is based on the amount of
dirty data, rather than guesses about future performance of the system.  When
there is a lot of dirty data, each transaction (e.g. write() syscall) will be
delayed by the same small amount.  This eliminates the "brick wall of wait"
that the old write throttle could hit, causing all transactions to wait several
seconds until the next txg opens.  One of the keys to the new write throttle is
decrementing the amount of dirty data as i/o completes, rather than at the end
of spa_sync().  Note that the write throttle is only applied once the i/o
scheduler is issuing the maximum number of outstanding async writes.  See the
block comments in dsl_pool.c and above dmu_tx_delay() (reproduced below) for
more details.

This diff has several other effects, including:

 * the commonly-tuned global variable zfs_vdev_max_pending has been removed;
use per-class zfs_vdev_*_max_active values or zfs_vdev_max_active instead.

 * the size of each txg (meaning the amount of dirty data written, and thus the
time it takes to write out) is now controlled differently.  There is no longer
an explicit time goal; the primary determinant is amount of dirty data.
Systems that are under light or medium load will now often see that a txg is
always syncing, but the impact to performance (e.g. read latency) is minimal.
Tune zfs_dirty_data_max and zfs_dirty_data_sync to control this.

 * zio_taskq_batch_pct = 75 -- Only use 75% of all CPUs for compression,
checksum, etc.  This improves latency by not allowing these CPU-intensive tasks
to consume all CPU (on machines with at least 4 CPU's; the percentage is
rounded up).

--matt

APPENDIX: problems with the current i/o scheduler

The current ZFS i/o scheduler (vdev_queue.c) is deadline based.  The problem
with this is that if there are always i/os pending, then certain classes of
i/os can see very long delays.

For example, if there are always synchronous reads outstanding, then no async
writes will be serviced until they become "past due".  One symptom of this
situation is that each pass of the txg sync takes at least several seconds
(typically 3 seconds).

If many i/os become "past due" (their deadline is in the past), then we must
service all of these overdue i/os before any new i/os.  This happens when we
enqueue a batch of async writes for the txg sync, with deadlines 2.5 seconds in
the future.  If we can't complete all the i/os in 2.5 seconds (e.g. because
there were always reads pending), then these i/os will become past due.  Now we
must service all the "async" writes (which could be hundreds of megabytes)
before we service any reads, introducing considerable latency to synchronous
i/os (reads or ZIL writes).

Notes on porting to ZFS on Linux:

- zio_t gained new members io_physdone and io_phys_children.  Because
  object caches in the Linux port call the constructor only once at
  allocation time, objects may contain residual data when retrieved
  from the cache. Therefore zio_create() was updated to zero out the two
  new fields.

- vdev_mirror_pending() relied on the depth of the per-vdev pending queue
  (vq->vq_pending_tree) to select the least-busy leaf vdev to read from.
  This tree has been replaced by vq->vq_active_tree which is now used
  for the same purpose.

- vdev_queue_init() used the value of zfs_vdev_max_pending to determine
  the number of vdev I/O buffers to pre-allocate.  That global no longer
  exists, so we instead use the sum of the *_max_active values for each of
  the five I/O classes described above.

- The Illumos implementation of dmu_tx_delay() delays a transaction by
  sleeping in condition variable embedded in the thread
  (curthread->t_delay_cv).  We do not have an equivalent CV to use in
  Linux, so this change replaced the delay logic with a wrapper called
  zfs_sleep_until(). This wrapper could be adopted upstream and in other
  downstream ports to abstract away operating system-specific delay logic.

- These tunables are added as module parameters, and descriptions added
  to the zfs-module-parameters.5 man page.

  spa_asize_inflation
  zfs_deadman_synctime_ms
  zfs_vdev_max_active
  zfs_vdev_async_write_active_min_dirty_percent
  zfs_vdev_async_write_active_max_dirty_percent
  zfs_vdev_async_read_max_active
  zfs_vdev_async_read_min_active
  zfs_vdev_async_write_max_active
  zfs_vdev_async_write_min_active
  zfs_vdev_scrub_max_active
  zfs_vdev_scrub_min_active
  zfs_vdev_sync_read_max_active
  zfs_vdev_sync_read_min_active
  zfs_vdev_sync_write_max_active
  zfs_vdev_sync_write_min_active
  zfs_dirty_data_max_percent
  zfs_delay_min_dirty_percent
  zfs_dirty_data_max_max_percent
  zfs_dirty_data_max
  zfs_dirty_data_max_max
  zfs_dirty_data_sync
  zfs_delay_scale

  The latter four have type unsigned long, whereas they are uint64_t in
  Illumos.  This accommodates Linux's module_param() supported types, but
  means they may overflow on 32-bit architectures.

  The values zfs_dirty_data_max and zfs_dirty_data_max_max are the most
  likely to overflow on 32-bit systems, since they express physical RAM
  sizes in bytes.  In fact, Illumos initializes zfs_dirty_data_max_max to
  2^32 which does overflow. To resolve that, this port instead initializes
  it in arc_init() to 25% of physical RAM, and adds the tunable
  zfs_dirty_data_max_max_percent to override that percentage.  While this
  solution doesn't completely avoid the overflow issue, it should be a
  reasonable default for most systems, and the minority of affected
  systems can work around the issue by overriding the defaults.

- Fixed reversed logic in comment above zfs_delay_scale declaration.

- Clarified comments in vdev_queue.c regarding when per-queue minimums take
  effect.

- Replaced dmu_tx_write_limit in the dmu_tx kstat file
  with dmu_tx_dirty_delay and dmu_tx_dirty_over_max.  The first counts
  how many times a transaction has been delayed because the pool dirty
  data has exceeded zfs_delay_min_dirty_percent.  The latter counts how
  many times the pool dirty data has exceeded zfs_dirty_data_max (which
  we expect to never happen).

- The original patch would have regressed the bug fixed in
  zfsonlinux/zfs@c418410, which prevented users from setting the
  zfs_vdev_aggregation_limit tuning larger than SPA_MAXBLOCKSIZE.
  A similar fix is added to vdev_queue_aggregate().

- In vdev_queue_io_to_issue(), dynamically allocate 'zio_t search' on the
  heap instead of the stack.  In Linux we can't afford such large
  structures on the stack.

Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Adam Leventhal <ahl@delphix.com>
Reviewed by: Christopher Siden <christopher.siden@delphix.com>
Reviewed by: Ned Bass <bass6@llnl.gov>
Reviewed by: Brendan Gregg <brendan.gregg@joyent.com>
Approved by: Robert Mustacchi <rm@joyent.com>

References:
  http://www.illumos.org/issues/4045
  illumos/illumos-gate@69962b5647

Ported-by: Ned Bass <bass6@llnl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes #1913
2013-12-06 09:32:43 -08:00

383 lines
10 KiB
C
Raw Blame History

/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or http://www.opensolaris.org/os/licensing.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2013 by Delphix. All rights reserved.
* Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
*/
#ifndef _SYS_DBUF_H
#define _SYS_DBUF_H
#include <sys/dmu.h>
#include <sys/spa.h>
#include <sys/txg.h>
#include <sys/zio.h>
#include <sys/arc.h>
#include <sys/zfs_context.h>
#include <sys/refcount.h>
#include <sys/zrlock.h>
#ifdef __cplusplus
extern "C" {
#endif
#define IN_DMU_SYNC 2
/*
* define flags for dbuf_read
*/
#define DB_RF_MUST_SUCCEED (1 << 0)
#define DB_RF_CANFAIL (1 << 1)
#define DB_RF_HAVESTRUCT (1 << 2)
#define DB_RF_NOPREFETCH (1 << 3)
#define DB_RF_NEVERWAIT (1 << 4)
#define DB_RF_CACHED (1 << 5)
/*
* The simplified state transition diagram for dbufs looks like:
*
* +----> READ ----+
* | |
* | V
* (alloc)-->UNCACHED CACHED-->EVICTING-->(free)
* | ^ ^
* | | |
* +----> FILL ----+ |
* | |
* | |
* +--------> NOFILL -------+
*/
typedef enum dbuf_states {
DB_UNCACHED,
DB_FILL,
DB_NOFILL,
DB_READ,
DB_CACHED,
DB_EVICTING
} dbuf_states_t;
struct dnode;
struct dmu_tx;
/*
* level = 0 means the user data
* level = 1 means the single indirect block
* etc.
*/
struct dmu_buf_impl;
typedef enum override_states {
DR_NOT_OVERRIDDEN,
DR_IN_DMU_SYNC,
DR_OVERRIDDEN
} override_states_t;
typedef struct dbuf_dirty_record {
/* link on our parents dirty list */
list_node_t dr_dirty_node;
/* transaction group this data will sync in */
uint64_t dr_txg;
/* zio of outstanding write IO */
zio_t *dr_zio;
/* pointer back to our dbuf */
struct dmu_buf_impl *dr_dbuf;
/* pointer to next dirty record */
struct dbuf_dirty_record *dr_next;
/* pointer to parent dirty record */
struct dbuf_dirty_record *dr_parent;
/* How much space was changed to dsl_pool_dirty_space() for this? */
unsigned int dr_accounted;
union dirty_types {
struct dirty_indirect {
/* protect access to list */
kmutex_t dr_mtx;
/* Our list of dirty children */
list_t dr_children;
} di;
struct dirty_leaf {
/*
* dr_data is set when we dirty the buffer
* so that we can retain the pointer even if it
* gets COW'd in a subsequent transaction group.
*/
arc_buf_t *dr_data;
blkptr_t dr_overridden_by;
override_states_t dr_override_state;
uint8_t dr_copies;
boolean_t dr_nopwrite;
} dl;
} dt;
} dbuf_dirty_record_t;
typedef struct dmu_buf_impl {
/*
* The following members are immutable, with the exception of
* db.db_data, which is protected by db_mtx.
*/
/* the publicly visible structure */
dmu_buf_t db;
/* the objset we belong to */
struct objset *db_objset;
/*
* handle to safely access the dnode we belong to (NULL when evicted)
*/
struct dnode_handle *db_dnode_handle;
/*
* our parent buffer; if the dnode points to us directly,
* db_parent == db_dnode_handle->dnh_dnode->dn_dbuf
* only accessed by sync thread ???
* (NULL when evicted)
* May change from NULL to non-NULL under the protection of db_mtx
* (see dbuf_check_blkptr())
*/
struct dmu_buf_impl *db_parent;
/*
* link for hash table of all dmu_buf_impl_t's
*/
struct dmu_buf_impl *db_hash_next;
/* our block number */
uint64_t db_blkid;
/*
* Pointer to the blkptr_t which points to us. May be NULL if we
* don't have one yet. (NULL when evicted)
*/
blkptr_t *db_blkptr;
/*
* Our indirection level. Data buffers have db_level==0.
* Indirect buffers which point to data buffers have
* db_level==1. etc. Buffers which contain dnodes have
* db_level==0, since the dnodes are stored in a file.
*/
uint8_t db_level;
/* db_mtx protects the members below */
kmutex_t db_mtx;
/*
* Current state of the buffer
*/
dbuf_states_t db_state;
/*
* Refcount accessed by dmu_buf_{hold,rele}.
* If nonzero, the buffer can't be destroyed.
* Protected by db_mtx.
*/
refcount_t db_holds;
/* buffer holding our data */
arc_buf_t *db_buf;
kcondvar_t db_changed;
dbuf_dirty_record_t *db_data_pending;
/* pointer to most recent dirty record for this buffer */
dbuf_dirty_record_t *db_last_dirty;
/*
* Our link on the owner dnodes's dn_dbufs list.
* Protected by its dn_dbufs_mtx.
*/
list_node_t db_link;
/* Data which is unique to data (leaf) blocks: */
/* stuff we store for the user (see dmu_buf_set_user) */
void *db_user_ptr;
void **db_user_data_ptr_ptr;
dmu_buf_evict_func_t *db_evict_func;
uint8_t db_immediate_evict;
uint8_t db_freed_in_flight;
uint8_t db_dirtycnt;
} dmu_buf_impl_t;
/* Note: the dbuf hash table is exposed only for the mdb module */
#define DBUF_MUTEXES 256
#define DBUF_HASH_MUTEX(h, idx) (&(h)->hash_mutexes[(idx) & (DBUF_MUTEXES-1)])
typedef struct dbuf_hash_table {
uint64_t hash_table_mask;
dmu_buf_impl_t **hash_table;
kmutex_t hash_mutexes[DBUF_MUTEXES];
} dbuf_hash_table_t;
uint64_t dbuf_whichblock(struct dnode *di, uint64_t offset);
void dbuf_create_bonus(struct dnode *dn);
int dbuf_spill_set_blksz(dmu_buf_t *db, uint64_t blksz, dmu_tx_t *tx);
void dbuf_rm_spill(struct dnode *dn, dmu_tx_t *tx);
dmu_buf_impl_t *dbuf_hold(struct dnode *dn, uint64_t blkid, void *tag);
dmu_buf_impl_t *dbuf_hold_level(struct dnode *dn, int level, uint64_t blkid,
void *tag);
int dbuf_hold_impl(struct dnode *dn, uint8_t level, uint64_t blkid, int create,
void *tag, dmu_buf_impl_t **dbp);
void dbuf_prefetch(struct dnode *dn, uint64_t blkid, zio_priority_t prio);
void dbuf_add_ref(dmu_buf_impl_t *db, void *tag);
uint64_t dbuf_refcount(dmu_buf_impl_t *db);
void dbuf_rele(dmu_buf_impl_t *db, void *tag);
void dbuf_rele_and_unlock(dmu_buf_impl_t *db, void *tag);
dmu_buf_impl_t *dbuf_find(struct dnode *dn, uint8_t level, uint64_t blkid);
int dbuf_read(dmu_buf_impl_t *db, zio_t *zio, uint32_t flags);
void dbuf_will_dirty(dmu_buf_impl_t *db, dmu_tx_t *tx);
void dbuf_fill_done(dmu_buf_impl_t *db, dmu_tx_t *tx);
void dmu_buf_will_not_fill(dmu_buf_t *db, dmu_tx_t *tx);
void dmu_buf_will_fill(dmu_buf_t *db, dmu_tx_t *tx);
void dmu_buf_fill_done(dmu_buf_t *db, dmu_tx_t *tx);
void dbuf_assign_arcbuf(dmu_buf_impl_t *db, arc_buf_t *buf, dmu_tx_t *tx);
dbuf_dirty_record_t *dbuf_dirty(dmu_buf_impl_t *db, dmu_tx_t *tx);
arc_buf_t *dbuf_loan_arcbuf(dmu_buf_impl_t *db);
void dbuf_clear(dmu_buf_impl_t *db);
void dbuf_evict(dmu_buf_impl_t *db);
void dbuf_unoverride(dbuf_dirty_record_t *dr);
void dbuf_sync_list(list_t *list, dmu_tx_t *tx);
void dbuf_release_bp(dmu_buf_impl_t *db);
void dbuf_free_range(struct dnode *dn, uint64_t start, uint64_t end,
struct dmu_tx *);
void dbuf_new_size(dmu_buf_impl_t *db, int size, dmu_tx_t *tx);
void dbuf_stats_init(dbuf_hash_table_t *hash);
void dbuf_stats_destroy(void);
#define DB_DNODE(_db) ((_db)->db_dnode_handle->dnh_dnode)
#define DB_DNODE_LOCK(_db) ((_db)->db_dnode_handle->dnh_zrlock)
#define DB_DNODE_ENTER(_db) (zrl_add(&DB_DNODE_LOCK(_db)))
#define DB_DNODE_EXIT(_db) (zrl_remove(&DB_DNODE_LOCK(_db)))
#define DB_DNODE_HELD(_db) (!zrl_is_zero(&DB_DNODE_LOCK(_db)))
#define DB_GET_SPA(_spa_p, _db) { \
dnode_t *__dn; \
DB_DNODE_ENTER(_db); \
__dn = DB_DNODE(_db); \
*(_spa_p) = __dn->dn_objset->os_spa; \
DB_DNODE_EXIT(_db); \
}
#define DB_GET_OBJSET(_os_p, _db) { \
dnode_t *__dn; \
DB_DNODE_ENTER(_db); \
__dn = DB_DNODE(_db); \
*(_os_p) = __dn->dn_objset; \
DB_DNODE_EXIT(_db); \
}
void dbuf_init(void);
void dbuf_fini(void);
boolean_t dbuf_is_metadata(dmu_buf_impl_t *db);
#define DBUF_GET_BUFC_TYPE(_db) \
(dbuf_is_metadata(_db) ? ARC_BUFC_METADATA : ARC_BUFC_DATA)
#define DBUF_IS_CACHEABLE(_db) \
((_db)->db_objset->os_primary_cache == ZFS_CACHE_ALL || \
(dbuf_is_metadata(_db) && \
((_db)->db_objset->os_primary_cache == ZFS_CACHE_METADATA)))
#define DBUF_IS_L2CACHEABLE(_db) \
((_db)->db_objset->os_secondary_cache == ZFS_CACHE_ALL || \
(dbuf_is_metadata(_db) && \
((_db)->db_objset->os_secondary_cache == ZFS_CACHE_METADATA)))
#define DBUF_IS_L2COMPRESSIBLE(_db) \
((_db)->db_objset->os_compress != ZIO_COMPRESS_OFF || \
(dbuf_is_metadata(_db) && zfs_mdcomp_disable == B_FALSE))
#ifdef ZFS_DEBUG
/*
* There should be a ## between the string literal and fmt, to make it
* clear that we're joining two strings together, but gcc does not
* support that preprocessor token.
*/
#define dprintf_dbuf(dbuf, fmt, ...) do { \
if (zfs_flags & ZFS_DEBUG_DPRINTF) { \
char __db_buf[32]; \
uint64_t __db_obj = (dbuf)->db.db_object; \
if (__db_obj == DMU_META_DNODE_OBJECT) \
(void) strcpy(__db_buf, "mdn"); \
else \
(void) snprintf(__db_buf, sizeof (__db_buf), "%lld", \
(u_longlong_t)__db_obj); \
dprintf_ds((dbuf)->db_objset->os_dsl_dataset, \
"obj=%s lvl=%u blkid=%lld " fmt, \
__db_buf, (dbuf)->db_level, \
(u_longlong_t)(dbuf)->db_blkid, __VA_ARGS__); \
} \
_NOTE(CONSTCOND) } while (0)
#define dprintf_dbuf_bp(db, bp, fmt, ...) do { \
if (zfs_flags & ZFS_DEBUG_DPRINTF) { \
char *__blkbuf = kmem_alloc(BP_SPRINTF_LEN, KM_PUSHPAGE); \
sprintf_blkptr(__blkbuf, bp); \
dprintf_dbuf(db, fmt " %s\n", __VA_ARGS__, __blkbuf); \
kmem_free(__blkbuf, BP_SPRINTF_LEN); \
} \
_NOTE(CONSTCOND) } while (0)
#define DBUF_VERIFY(db) dbuf_verify(db)
#else
#define dprintf_dbuf(db, fmt, ...)
#define dprintf_dbuf_bp(db, bp, fmt, ...)
#define DBUF_VERIFY(db)
#endif
#ifdef __cplusplus
}
#endif
#endif /* _SYS_DBUF_H */
Reference in New Issue View Git Blame Copy Permalink