e8b96c6007
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
125 lines
4.7 KiB
C
125 lines
4.7 KiB
C
/*
|
|
* 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 2009 Sun Microsystems, Inc. All rights reserved.
|
|
* Use is subject to license terms.
|
|
*/
|
|
|
|
/*
|
|
* Copyright (c) 2013 by Delphix. All rights reserved.
|
|
*/
|
|
|
|
#ifndef _SYS_TXG_IMPL_H
|
|
#define _SYS_TXG_IMPL_H
|
|
|
|
#include <sys/spa.h>
|
|
#include <sys/txg.h>
|
|
|
|
#ifdef __cplusplus
|
|
extern "C" {
|
|
#endif
|
|
|
|
/*
|
|
* The tx_cpu structure is a per-cpu structure that is used to track
|
|
* the number of active transaction holds (tc_count). As transactions
|
|
* are assigned into a transaction group the appropriate tc_count is
|
|
* incremented to indicate that there are pending changes that have yet
|
|
* to quiesce. Consumers evenutally call txg_rele_to_sync() to decrement
|
|
* the tc_count. A transaction group is not considered quiesced until all
|
|
* tx_cpu structures have reached a tc_count of zero.
|
|
*
|
|
* This structure is a per-cpu structure by design. Updates to this structure
|
|
* are frequent and concurrent. Having a single structure would result in
|
|
* heavy lock contention so a per-cpu design was implemented. With the fanned
|
|
* out mutex design, consumers only need to lock the mutex associated with
|
|
* thread's cpu.
|
|
*
|
|
* The tx_cpu contains two locks, the tc_lock and tc_open_lock.
|
|
* The tc_lock is used to protect all members of the tx_cpu structure with
|
|
* the exception of the tc_open_lock. This lock should only be held for a
|
|
* short period of time, typically when updating the value of tc_count.
|
|
*
|
|
* The tc_open_lock protects the tx_open_txg member of the tx_state structure.
|
|
* This lock is used to ensure that transactions are only assigned into
|
|
* the current open transaction group. In order to move the current open
|
|
* transaction group to the quiesce phase, the txg_quiesce thread must
|
|
* grab all tc_open_locks, increment the tx_open_txg, and drop the locks.
|
|
* The tc_open_lock is held until the transaction is assigned into the
|
|
* transaction group. Typically, this is a short operation but if throttling
|
|
* is occuring it may be held for longer periods of time.
|
|
*/
|
|
struct tx_cpu {
|
|
kmutex_t tc_open_lock; /* protects tx_open_txg */
|
|
kmutex_t tc_lock; /* protects the rest of this struct */
|
|
kcondvar_t tc_cv[TXG_SIZE];
|
|
uint64_t tc_count[TXG_SIZE]; /* tx hold count on each txg */
|
|
list_t tc_callbacks[TXG_SIZE]; /* commit cb list */
|
|
char tc_pad[8]; /* pad to fill 3 cache lines */
|
|
};
|
|
|
|
/*
|
|
* The tx_state structure maintains the state information about the different
|
|
* stages of the pool's transcation groups. A per pool tx_state structure
|
|
* is used to track this information. The tx_state structure also points to
|
|
* an array of tx_cpu structures (described above). Although the tx_sync_lock
|
|
* is used to protect the members of this structure, it is not used to
|
|
* protect the tx_open_txg. Instead a special lock in the tx_cpu structure
|
|
* is used. Readers of tx_open_txg must grab the per-cpu tc_open_lock.
|
|
* Any thread wishing to update tx_open_txg must grab the tc_open_lock on
|
|
* every cpu (see txg_quiesce()).
|
|
*/
|
|
typedef struct tx_state {
|
|
tx_cpu_t *tx_cpu; /* protects access to tx_open_txg */
|
|
kmutex_t tx_sync_lock; /* protects the rest of this struct */
|
|
|
|
uint64_t tx_open_txg; /* currently open txg id */
|
|
uint64_t tx_quiesced_txg; /* quiesced txg waiting for sync */
|
|
uint64_t tx_syncing_txg; /* currently syncing txg id */
|
|
uint64_t tx_synced_txg; /* last synced txg id */
|
|
|
|
hrtime_t tx_open_time; /* start time of tx_open_txg */
|
|
|
|
uint64_t tx_sync_txg_waiting; /* txg we're waiting to sync */
|
|
uint64_t tx_quiesce_txg_waiting; /* txg we're waiting to open */
|
|
|
|
kcondvar_t tx_sync_more_cv;
|
|
kcondvar_t tx_sync_done_cv;
|
|
kcondvar_t tx_quiesce_more_cv;
|
|
kcondvar_t tx_quiesce_done_cv;
|
|
kcondvar_t tx_timeout_cv;
|
|
kcondvar_t tx_exit_cv; /* wait for all threads to exit */
|
|
|
|
uint8_t tx_threads; /* number of threads */
|
|
uint8_t tx_exiting; /* set when we're exiting */
|
|
|
|
kthread_t *tx_sync_thread;
|
|
kthread_t *tx_quiesce_thread;
|
|
|
|
taskq_t *tx_commit_cb_taskq; /* commit callback taskq */
|
|
} tx_state_t;
|
|
|
|
#ifdef __cplusplus
|
|
}
|
|
#endif
|
|
|
|
#endif /* _SYS_TXG_IMPL_H */
|