proxmox-mirrors/mirror_zfs
1
0
Fork 0
mirror of https://git.proxmox.com/git/mirror_zfs synced 2025-04-28 09:46:17 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
94a3fabcb0
mirror_zfs/include/sys/dsl_pool.h
Alexander Motin 94a3fabcb0
Unified allocation throttling (#17020) 
Existing allocation throttling had a goal to improve write speed
by allocating more data to vdevs that are able to write it faster.
But in the process it completely broken the original mechanism,
designed to balance vdev space usage.  With severe vdev space use
imbalance it is possible that some with higher use start growing
fragmentation sooner than others and after getting full will stop
any writes at all.  Also after vdev addition it might take a very
long time for pool to restore the balance, since the new vdev does
not have any real preference, unless the old one is already much
slower due to fragmentation.  Also the old throttling was request-
based, which was unpredictable with block sizes varying from 512B
to 16MB, neither it made much sense in case of I/O aggregation,
when its 32-100 requests could be aggregated into few, leaving
device underutilized, submitting fewer and/or shorter requests,
or in opposite try to queue up to 1.6GB of writes per device.

This change presents a completely new throttling algorithm. Unlike
the request-based old one, this one measures allocation queue in
bytes.  It makes possible to integrate with the reworked allocation
quota (aliquot) mechanism, which is also byte-based.  Unlike the
original code, balancing the vdevs amounts of free space, this one
balances their free/used space fractions.  It should result in a
lower and more uniform fragmentation in a long run.

This algorithm still allows to improve write speed by allocating
more data to faster vdevs, but does it in more controllable way.
On top of space-based allocation quota, it also calculates minimum
queue depth that vdev is allowed to maintain, and respectively the
amount of extra allocations it can receive if it appear faster.
That amount is based on vdev's capacity and space usage, but also
applied only when the pool is busy.  This way the code can choose
between faster writes when needed and better vdev balance when not,
with the choice gradually reducing together with the free space.

This change also makes allocation queues per-class, allowing them
to throttle independently and in parallel.  Allocations that are
bounced between classes due to allocation errors will be able to
properly throttle in the new class.  Allocations that should not
be throttled (ZIL, gang, copies) are not, but may still follow
the rotor and allocation quota mechanism of the class without
disrupting it.

Signed-off-by:	Alexander Motin <mav@FreeBSD.org>
Sponsored by:	iXsystems, Inc.
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Paul Dagnelie <pcd@delphix.com>
2025-03-24 09:25:01 -07:00

209 lines
6.6 KiB
C
Raw Blame History

// SPDX-License-Identifier: CDDL-1.0
/*
* 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 https://opensource.org/licenses/CDDL-1.0.
* 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, 2018 by Delphix. All rights reserved.
* Copyright 2016 Nexenta Systems, Inc. All rights reserved.
*/
#ifndef _SYS_DSL_POOL_H
#define _SYS_DSL_POOL_H
#include <sys/spa.h>
#include <sys/txg.h>
#include <sys/txg_impl.h>
#include <sys/zfs_context.h>
#include <sys/zio.h>
#include <sys/dnode.h>
#include <sys/ddt.h>
#include <sys/arc.h>
#include <sys/bpobj.h>
#include <sys/bptree.h>
#include <sys/rrwlock.h>
#include <sys/dsl_synctask.h>
#include <sys/mmp.h>
#include <sys/aggsum.h>
#ifdef __cplusplus
extern "C" {
#endif
extern int zfs_txg_synctime_ms;
struct objset;
struct dsl_dir;
struct dsl_dataset;
struct dsl_pool;
struct dmu_tx;
struct dsl_scan;
struct dsl_crypto_params;
struct dsl_deadlist;
extern uint64_t zfs_dirty_data_max;
extern uint64_t zfs_dirty_data_max_max;
extern uint64_t zfs_wrlog_data_max;
extern uint_t zfs_dirty_data_max_percent;
extern uint_t zfs_dirty_data_max_max_percent;
extern uint_t zfs_delay_min_dirty_percent;
extern uint_t zfs_vdev_async_write_active_min_dirty_percent;
extern uint_t zfs_vdev_async_write_active_max_dirty_percent;
extern uint64_t zfs_delay_scale;
/* These macros are for indexing into the zfs_all_blkstats_t. */
#define DMU_OT_DEFERRED DMU_OT_NONE
#define DMU_OT_OTHER DMU_OT_NUMTYPES /* place holder for DMU_OT() types */
#define DMU_OT_TOTAL (DMU_OT_NUMTYPES + 1)
typedef struct zfs_blkstat {
uint64_t zb_count;
uint64_t zb_asize;
uint64_t zb_lsize;
uint64_t zb_psize;
uint64_t zb_gangs;
uint64_t zb_ditto_2_of_2_samevdev;
uint64_t zb_ditto_2_of_3_samevdev;
uint64_t zb_ditto_3_of_3_samevdev;
} zfs_blkstat_t;
typedef struct zfs_all_blkstats {
zfs_blkstat_t zab_type[DN_MAX_LEVELS + 1][DMU_OT_TOTAL + 1];
} zfs_all_blkstats_t;
typedef struct dsl_pool {
/* Immutable */
spa_t *dp_spa;
struct objset *dp_meta_objset;
struct dsl_dir *dp_root_dir;
struct dsl_dir *dp_mos_dir;
struct dsl_dir *dp_free_dir;
struct dsl_dir *dp_leak_dir;
struct dsl_dataset *dp_origin_snap;
uint64_t dp_root_dir_obj;
struct taskq *dp_zrele_taskq;
struct taskq *dp_unlinked_drain_taskq;
/* No lock needed - sync context only */
blkptr_t dp_meta_rootbp;
uint64_t dp_tmp_userrefs_obj;
bpobj_t dp_free_bpobj;
uint64_t dp_bptree_obj;
uint64_t dp_empty_bpobj;
bpobj_t dp_obsolete_bpobj;
struct dsl_scan *dp_scan;
/* Uses dp_lock */
kmutex_t dp_lock;
kcondvar_t dp_spaceavail_cv;
uint64_t dp_dirty_pertxg[TXG_SIZE];
uint64_t dp_dirty_total;
uint64_t dp_long_free_dirty_pertxg[TXG_SIZE];
uint64_t dp_mos_used_delta;
uint64_t dp_mos_compressed_delta;
uint64_t dp_mos_uncompressed_delta;
aggsum_t dp_wrlog_pertxg[TXG_SIZE];
aggsum_t dp_wrlog_total;
/*
* Time of most recently scheduled (furthest in the future)
* wakeup for delayed transactions.
*/
hrtime_t dp_last_wakeup;
/* Has its own locking */
tx_state_t dp_tx;
txg_list_t dp_dirty_datasets;
txg_list_t dp_dirty_zilogs;
txg_list_t dp_dirty_dirs;
txg_list_t dp_sync_tasks;
txg_list_t dp_early_sync_tasks;
taskq_t *dp_sync_taskq;
taskq_t *dp_zil_clean_taskq;
/*
* Protects administrative changes (properties, namespace)
*
* It is only held for write in syncing context. Therefore
* syncing context does not need to ever have it for read, since
* nobody else could possibly have it for write.
*/
rrwlock_t dp_config_rwlock;
zfs_all_blkstats_t *dp_blkstats;
} dsl_pool_t;
int dsl_pool_init(spa_t *spa, uint64_t txg, dsl_pool_t **dpp);
int dsl_pool_open(dsl_pool_t *dp);
void dsl_pool_close(dsl_pool_t *dp);
dsl_pool_t *dsl_pool_create(spa_t *spa, nvlist_t *zplprops,
struct dsl_crypto_params *dcp, uint64_t txg);
void dsl_pool_sync(dsl_pool_t *dp, uint64_t txg);
void dsl_pool_sync_done(dsl_pool_t *dp, uint64_t txg);
int dsl_pool_sync_context(dsl_pool_t *dp);
uint64_t dsl_pool_adjustedsize(dsl_pool_t *dp, zfs_space_check_t slop_policy);
uint64_t dsl_pool_unreserved_space(dsl_pool_t *dp,
zfs_space_check_t slop_policy);
uint64_t dsl_pool_deferred_space(dsl_pool_t *dp);
void dsl_pool_wrlog_count(dsl_pool_t *dp, int64_t size, uint64_t txg);
boolean_t dsl_pool_need_wrlog_delay(dsl_pool_t *dp);
void dsl_pool_dirty_space(dsl_pool_t *dp, int64_t space, dmu_tx_t *tx);
void dsl_pool_undirty_space(dsl_pool_t *dp, int64_t space, uint64_t txg);
void dsl_free(dsl_pool_t *dp, uint64_t txg, const blkptr_t *bpp);
void dsl_free_sync(zio_t *pio, dsl_pool_t *dp, uint64_t txg,
const blkptr_t *bpp);
void dsl_pool_create_origin(dsl_pool_t *dp, dmu_tx_t *tx);
void dsl_pool_upgrade_clones(dsl_pool_t *dp, dmu_tx_t *tx);
void dsl_pool_upgrade_dir_clones(dsl_pool_t *dp, dmu_tx_t *tx);
void dsl_pool_mos_diduse_space(dsl_pool_t *dp,
int64_t used, int64_t comp, int64_t uncomp);
void dsl_pool_ckpoint_diduse_space(dsl_pool_t *dp,
int64_t used, int64_t comp, int64_t uncomp);
boolean_t dsl_pool_need_dirty_delay(dsl_pool_t *dp);
void dsl_pool_config_enter(dsl_pool_t *dp, const void *tag);
void dsl_pool_config_enter_prio(dsl_pool_t *dp, const void *tag);
void dsl_pool_config_exit(dsl_pool_t *dp, const void *tag);
boolean_t dsl_pool_config_held(dsl_pool_t *dp);
boolean_t dsl_pool_config_held_writer(dsl_pool_t *dp);
taskq_t *dsl_pool_zrele_taskq(dsl_pool_t *dp);
taskq_t *dsl_pool_unlinked_drain_taskq(dsl_pool_t *dp);
int dsl_pool_user_hold(dsl_pool_t *dp, uint64_t dsobj,
const char *tag, uint64_t now, dmu_tx_t *tx);
int dsl_pool_user_release(dsl_pool_t *dp, uint64_t dsobj,
const char *tag, dmu_tx_t *tx);
void dsl_pool_clean_tmp_userrefs(dsl_pool_t *dp);
int dsl_pool_open_special_dir(dsl_pool_t *dp, const char *name, dsl_dir_t **);
int dsl_pool_hold(const char *name, const void *tag, dsl_pool_t **dp);
void dsl_pool_rele(dsl_pool_t *dp, const void *tag);
void dsl_pool_create_obsolete_bpobj(dsl_pool_t *dp, dmu_tx_t *tx);
void dsl_pool_destroy_obsolete_bpobj(dsl_pool_t *dp, dmu_tx_t *tx);
#ifdef __cplusplus
}
#endif
#endif /* _SYS_DSL_POOL_H */
Reference in New Issue View Git Blame Copy Permalink