When we try assign a new transaction to a TXG we must know beforehand
if there is sufficient free space on disk. This is to decide,
in dmu_tx_assign(), if we should reject the TX with ENOSPC.
We rely on spa_get_worst_case_asize() to inflate the size of our
logical writes by a factor of spa_asize_inflation which is
calculated as:
(VDEV_RAIDZ_MAXPARITY + 1) * SPA_DVAS_PER_BP * 2 == 24
The problem with the current implementation is that we don't take
into account what happens with very small writes on VDEVs with large
physical block sizes.
Consider the case of writes to a dataset with recordsize=512,
copies=3 on a VDEV with ashift=13 (usually SSD with 8K block size):
every logical IO will end up allocating 3 * 8K = 24K on disk, so 512
bytes multiplied by 48, which is double the size we account for.
If we allow this kind of writes to be assigned a TX it is possible,
when the pool is almost full, to trigger an allocation failure
(ENOSPC) in the ZIO pipeline, which will in turn result in the whole
pool being suspended.
The bug is fixed by using, in spa_get_worst_case_asize(), the MAX()
value chosen between the logical io size from zfs_write() and the
maximum physical block size used among our VDEVs.
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: loli10K <ezomori.nozomu@gmail.com>
Closes#5941
This patch implement the hardware accelerator method in GZIP compression
in ZFS. When the ZFS pool is enabled GZIP compression, the compression
API will be automatically transferred to the hardware accelerator to
free up CPU resource and speed up the compression time.
* To enable Intel QAT hardware acceleration in ZOL you need to have QAT
hardware and the driver installed:
* QAT hardware DH8950:
http://ark.intel.com/products/79483/Intel-QuickAssist-Adapter-8950
* QAT driver:
https://01.org/intel-quickassist-technology
* Start QAT driver in your system:
service qat_service start
* Enable QAT in ZFS, e.g.:
./configure --with-qat=<qat-driver-path>/QAT1.6
make
* Set GZIP compression in ZFS dataset:
zfs set compression = gzip <dataset>
* Get QAT hardware statistics by:
cat /proc/spl/kstat/zfs/qat
* To disable QAT in ZFS:
insmod zfs.ko zfs_qat_disable=1
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: Weigang Li <weigang.li@intel.com>
Closes#5846
Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com>
Reviewed by: Brad Lewis <brad.lewis@delphix.com>
Reviewed by: Saso Kiselkov <saso.kiselkov@nexenta.com>
Reviewed by: Brian Behlendorf <behlendorf1@llnl.gov>
Ported-by: Matthew Ahrens <mahrens@delphix.com>
spa_sync() iterates over all the dirty dnodes and processes each of them
by calling dnode_sync(). If there are many dirty dnodes (e.g. because we
created or removed a lot of files), the single thread of spa_sync()
calling dnode_sync() can become a bottleneck. Additionally, if many
dnodes are dirtied concurrently in open context (e.g. due to concurrent
file creation), the os_lock will experience lock contention via
dnode_setdirty().
The solution is to track dirty dnodes on a multilist_t, and for
spa_sync() to use separate threads to process each of the sublists in
the multilist.
OpenZFS-issue: https://www.illumos.org/issues/7968
OpenZFS-commit: https://github.com/openzfs/openzfs/commit/4a2a54cCloses#5752
Reviewed by: Steve Gonczi <steve.gonczi@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com>
Ported-by: Brian Behlendorf <behlendorf1@llnl.gov>
Background information: This assertion about tx_space_* verifies that we
are not dirtying more stuff than we thought we would. We “need” to know
how much we will dirty so that we can check if we should fail this
transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the
transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() —
typically less than a millisecond), we call dbuf_dirty() on the exact
blocks that will be modified. Once this happens, the temporary
accounting in tx_space_* is unnecessary, because we know exactly what
blocks are newly dirtied; we call dnode_willuse_space() to track this
more exact accounting.
The fundamental problem causing this bug is that dmu_tx_hold_*() relies
on the current state in the DMU (e.g. dn_nlevels) to predict how much
will be dirtied by this transaction, but this state can change before we
actually perform the transaction (i.e. call dbuf_dirty()).
This bug will be fixed by removing the assertion that the tx_space_*
accounting is perfectly accurate (i.e. we never dirty more than was
predicted by dmu_tx_hold_*()). By removing the requirement that this
accounting be perfectly accurate, we can also vastly simplify it, e.g.
removing most of the logic in dmu_tx_count_*().
The new tx space accounting will be very approximate, and may be more or
less than what is actually dirtied. It will still be used to determine
if this transaction will put us over quota. Transactions that are marked
by dmu_tx_mark_netfree() will be excepted from this check. We won’t make
an attempt to determine how much space will be freed by the transaction
— this was rarely accurate enough to determine if a transaction should
be permitted when we are over quota, which is why dmu_tx_mark_netfree()
was introduced in 2014.
We also won’t attempt to give “credit” when overwriting existing blocks,
if those blocks may be freed. This allows us to remove the
do_free_accounting logic in dbuf_dirty(), and associated routines. This
logic attempted to predict what will be on disk when this txg syncs, to
know if the overwritten block will be freed (i.e. exists, and has no
snapshots).
OpenZFS-issue: https://www.illumos.org/issues/7793
OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a
Upstream bugs: DLPX-32883a
Closes#5804
Porting notes:
- DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(),
Using the default dnode size would be slightly better.
- DEBUG_DMU_TX wrappers and configure option removed.
- Resolved _by_dnode() conflicts these changes have not yet been
applied to OpenZFS.
The deadman in ZoL didn't behave quite as it did in upstream
OpenZFS. In addition to the 2 purposes for which OpenZFS used the
zfs_deadman_synctime_ms parameter, ZoL also used it to determine how
frequently the deadman would fire once it has been triggered.
This patch adds the zfs_deadman_checktime_ms parameter to control how
frequently the subsequent checks are performed.
The deadman is now disabled for suspended pools.
As had been the case, unlike upstream OpenZFS, ZoL will not panic when
a hung IO is detected.
The module parameter documentation has been upated to include the new
parameter and to better describe the operation of the deadmen.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Signed-off-by: Tim Chase <tim@chase2k.com>
Closes#5695
The introduction of parallel zvol prefetch causes deadlock when using
vdev_file.
spa_async->(spa_namespace_lock)->txg_wait_synced->(wait for txg_sync)
txg_sync->zio_wait->(wait for vdev_file_io_fsync on system_taskq)
zvol_prefetch_minors_impl (on system_taskq)->spa_open_common->(wait for spa_namespace_lock)
We fix this by using dedicated taskq for vdev_file. This same change
was originally made in commit bc25c93 but reverted in commit aa9af22
when dynamic taskqs were added.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Chunwei Chen <tuxoko@gmail.com>
Closes#5506Closes#5495
Enable picky cstyle checks and resolve the new warnings. The vast
majority of the changes needed were to handle minor issues with
whitespace formatting. This patch contains no functional changes.
Non-whitespace changes are as follows:
* 8 times ; to { } in for/while loop
* fix missing ; in cmd/zed/agents/zfs_diagnosis.c
* comment (confim -> confirm)
* change endline , to ; in cmd/zpool/zpool_main.c
* a number of /* BEGIN CSTYLED */ /* END CSTYLED */ blocks
* /* CSTYLED */ markers
* change == 0 to !
* ulong to unsigned long in module/zfs/dsl_scan.c
* rearrangement of module_param lines in module/zfs/metaslab.c
* add { } block around statement after for_each_online_node
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Håkan Johansson <f96hajo@chalmers.se>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#5465
Speed up import and export speed by:
* Add system delay taskq
* Parallel prefetch zvol dnodes during zvol_create_minors
* Parallel zvol_free during zvol_remove_minors
* Reduce list linear search using ida and hash
Reviewed-by: Boris Protopopov <boris.protopopov@actifio.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Chunwei Chen <david.chen@osnexus.com>
Closes#5433
Save and reuse ddt dspace calculation when there have been no ddt changes.
This avoids unnecessary traversal of 168KiB of ddt histograms.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Gvozden Neskovic <neskovic@gmail.com>
Closes#5425
Use it for spa_deadman, zpl_posix_acl_free, snapentry_expire.
This free system_taskq from the above long delay tasks, and allow us to do
taskq_wait_outstanding on system_taskq without being blocked forever, making
system_taskq more generic and useful.
Signed-off-by: Chunwei Chen <david.chen@osnexus.com>
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: https://github.com/openzfs/openzfs/commit/4756c3d7Closes#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
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Saso Kiselkov <saso.kiselkov@nexenta.com>
Reviewed by: Richard Lowe <richlowe@richlowe.net>
Approved by: Garrett D'Amore <garrett@damore.org>
Ported by: Tony Hutter <hutter2@llnl.gov>
OpenZFS-issue: https://www.illumos.org/issues/4185
OpenZFS-commit: https://github.com/openzfs/openzfs/commit/45818ee
Porting Notes:
This code is ported on top of the Illumos Crypto Framework code:
https://github.com/zfsonlinux/zfs/pull/4329/commits/b5e030c8dbb9cd393d313571dee4756fbba8c22d
The list of porting changes includes:
- Copied module/icp/include/sha2/sha2.h directly from illumos
- Removed from module/icp/algs/sha2/sha2.c:
#pragma inline(SHA256Init, SHA384Init, SHA512Init)
- Added 'ctx' to lib/libzfs/libzfs_sendrecv.c:zio_checksum_SHA256() since
it now takes in an extra parameter.
- Added CTASSERT() to assert.h from for module/zfs/edonr_zfs.c
- Added skein & edonr to libicp/Makefile.am
- Added sha512.S. It was generated from sha512-x86_64.pl in Illumos.
- Updated ztest.c with new fletcher_4_*() args; used NULL for new CTX argument.
- In icp/algs/edonr/edonr_byteorder.h, Removed the #if defined(__linux) section
to not #include the non-existant endian.h.
- In skein_test.c, renane NULL to 0 in "no test vector" array entries to get
around a compiler warning.
- Fixup test files:
- Rename <sys/varargs.h> -> <varargs.h>, <strings.h> -> <string.h>,
- Remove <note.h> and define NOTE() as NOP.
- Define u_longlong_t
- Rename "#!/usr/bin/ksh" -> "#!/bin/ksh -p"
- Rename NULL to 0 in "no test vector" array entries to get around a
compiler warning.
- Remove "for isa in $($ISAINFO); do" stuff
- Add/update Makefiles
- Add some userspace headers like stdio.h/stdlib.h in places of
sys/types.h.
- EXPORT_SYMBOL *_Init/*_Update/*_Final... routines in ICP modules.
- Update scripts/zfs2zol-patch.sed
- include <sys/sha2.h> in sha2_impl.h
- Add sha2.h to include/sys/Makefile.am
- Add skein and edonr dirs to icp Makefile
- Add new checksums to zpool_get.cfg
- Move checksum switch block from zfs_secpolicy_setprop() to
zfs_check_settable()
- Fix -Wuninitialized error in edonr_byteorder.h on PPC
- Fix stack frame size errors on ARM32
- Don't unroll loops in Skein on 32-bit to save stack space
- Add memory barriers in sha2.c on 32-bit to save stack space
- Add filetest_001_pos.ksh checksum sanity test
- Add option to write psudorandom data in file_write utility
perf: 2.75x faster ddt_entry_compare()
First 256bits of ddt_key_t is a block checksum, which are expected
to be close to random data. Hence, on average, comparison only needs to
look at first few bytes of the keys. To reduce number of conditional
jump instructions, the result is computed as: sign(memcmp(k1, k2)).
Sign of an integer 'a' can be obtained as: `(0 < a) - (a < 0)` := {-1, 0, 1} ,
which is computed efficiently. Synthetic performance evaluation of
original and new algorithm over 1G random keys on 2.6GHz Intel(R) Xeon(R)
CPU E5-2660 v3:
old 6.85789 s
new 2.49089 s
perf: 2.8x faster vdev_queue_offset_compare() and vdev_queue_timestamp_compare()
Compute the result directly instead of using conditionals
perf: zfs_range_compare()
Speedup between 1.1x - 2.5x, depending on compiler version and
optimization level.
perf: spa_error_entry_compare()
`bcmp()` is not suitable for comparator use. Use `memcmp()` instead.
perf: 2.8x faster metaslab_compare() and metaslab_rangesize_compare()
perf: 2.8x faster zil_bp_compare()
perf: 2.8x faster mze_compare()
perf: faster dbuf_compare()
perf: faster compares in spa_misc
perf: 2.8x faster layout_hash_compare()
perf: 2.8x faster space_reftree_compare()
perf: libzfs: faster avl tree comparators
perf: guid_compare()
perf: dsl_deadlist_compare()
perf: perm_set_compare()
perf: 2x faster range_tree_seg_compare()
perf: faster unique_compare()
perf: faster vdev_cache _compare()
perf: faster vdev_uberblock_compare()
perf: faster fuid _compare()
perf: faster zfs_znode_hold_compare()
Signed-off-by: Gvozden Neskovic <neskovic@gmail.com>
Signed-off-by: Richard Elling <richard.elling@gmail.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#5033
Justification
-------------
This feature adds support for variable length dnodes. Our motivation is
to eliminate the overhead associated with using spill blocks. Spill
blocks are used to store system attribute data (i.e. file metadata) that
does not fit in the dnode's bonus buffer. By allowing a larger bonus
buffer area the use of a spill block can be avoided. Spill blocks
potentially incur an additional read I/O for every dnode in a dnode
block. As a worst case example, reading 32 dnodes from a 16k dnode block
and all of the spill blocks could issue 33 separate reads. Now suppose
those dnodes have size 1024 and therefore don't need spill blocks. Then
the worst case number of blocks read is reduced to from 33 to two--one
per dnode block. In practice spill blocks may tend to be co-located on
disk with the dnode blocks so the reduction in I/O would not be this
drastic. In a badly fragmented pool, however, the improvement could be
significant.
ZFS-on-Linux systems that make heavy use of extended attributes would
benefit from this feature. In particular, ZFS-on-Linux supports the
xattr=sa dataset property which allows file extended attribute data
to be stored in the dnode bonus buffer as an alternative to the
traditional directory-based format. Workloads such as SELinux and the
Lustre distributed filesystem often store enough xattr data to force
spill bocks when xattr=sa is in effect. Large dnodes may therefore
provide a performance benefit to such systems.
Other use cases that may benefit from this feature include files with
large ACLs and symbolic links with long target names. Furthermore,
this feature may be desirable on other platforms in case future
applications or features are developed that could make use of a
larger bonus buffer area.
Implementation
--------------
The size of a dnode may be a multiple of 512 bytes up to the size of
a dnode block (currently 16384 bytes). A dn_extra_slots field was
added to the current on-disk dnode_phys_t structure to describe the
size of the physical dnode on disk. The 8 bits for this field were
taken from the zero filled dn_pad2 field. The field represents how
many "extra" dnode_phys_t slots a dnode consumes in its dnode block.
This convention results in a value of 0 for 512 byte dnodes which
preserves on-disk format compatibility with older software.
Similarly, the in-memory dnode_t structure has a new dn_num_slots field
to represent the total number of dnode_phys_t slots consumed on disk.
Thus dn->dn_num_slots is 1 greater than the corresponding
dnp->dn_extra_slots. This difference in convention was adopted
because, unlike on-disk structures, backward compatibility is not a
concern for in-memory objects, so we used a more natural way to
represent size for a dnode_t.
The default size for newly created dnodes is determined by the value of
a new "dnodesize" dataset property. By default the property is set to
"legacy" which is compatible with older software. Setting the property
to "auto" will allow the filesystem to choose the most suitable dnode
size. Currently this just sets the default dnode size to 1k, but future
code improvements could dynamically choose a size based on observed
workload patterns. Dnodes of varying sizes can coexist within the same
dataset and even within the same dnode block. For example, to enable
automatically-sized dnodes, run
# zfs set dnodesize=auto tank/fish
The user can also specify literal values for the dnodesize property.
These are currently limited to powers of two from 1k to 16k. The
power-of-2 limitation is only for simplicity of the user interface.
Internally the implementation can handle any multiple of 512 up to 16k,
and consumers of the DMU API can specify any legal dnode value.
The size of a new dnode is determined at object allocation time and
stored as a new field in the znode in-memory structure. New DMU
interfaces are added to allow the consumer to specify the dnode size
that a newly allocated object should use. Existing interfaces are
unchanged to avoid having to update every call site and to preserve
compatibility with external consumers such as Lustre. The new
interfaces names are given below. The versions of these functions that
don't take a dnodesize parameter now just call the _dnsize() versions
with a dnodesize of 0, which means use the legacy dnode size.
New DMU interfaces:
dmu_object_alloc_dnsize()
dmu_object_claim_dnsize()
dmu_object_reclaim_dnsize()
New ZAP interfaces:
zap_create_dnsize()
zap_create_norm_dnsize()
zap_create_flags_dnsize()
zap_create_claim_norm_dnsize()
zap_create_link_dnsize()
The constant DN_MAX_BONUSLEN is renamed to DN_OLD_MAX_BONUSLEN. The
spa_maxdnodesize() function should be used to determine the maximum
bonus length for a pool.
These are a few noteworthy changes to key functions:
* The prototype for dnode_hold_impl() now takes a "slots" parameter.
When the DNODE_MUST_BE_FREE flag is set, this parameter is used to
ensure the hole at the specified object offset is large enough to
hold the dnode being created. The slots parameter is also used
to ensure a dnode does not span multiple dnode blocks. In both of
these cases, if a failure occurs, ENOSPC is returned. Keep in mind,
these failure cases are only possible when using DNODE_MUST_BE_FREE.
If the DNODE_MUST_BE_ALLOCATED flag is set, "slots" must be 0.
dnode_hold_impl() will check if the requested dnode is already
consumed as an extra dnode slot by an large dnode, in which case
it returns ENOENT.
* The function dmu_object_alloc() advances to the next dnode block
if dnode_hold_impl() returns an error for a requested object.
This is because the beginning of the next dnode block is the only
location it can safely assume to either be a hole or a valid
starting point for a dnode.
* dnode_next_offset_level() and other functions that iterate
through dnode blocks may no longer use a simple array indexing
scheme. These now use the current dnode's dn_num_slots field to
advance to the next dnode in the block. This is to ensure we
properly skip the current dnode's bonus area and don't interpret it
as a valid dnode.
zdb
---
The zdb command was updated to display a dnode's size under the
"dnsize" column when the object is dumped.
For ZIL create log records, zdb will now display the slot count for
the object.
ztest
-----
Ztest chooses a random dnodesize for every newly created object. The
random distribution is more heavily weighted toward small dnodes to
better simulate real-world datasets.
Unused bonus buffer space is filled with non-zero values computed from
the object number, dataset id, offset, and generation number. This
helps ensure that the dnode traversal code properly skips the interior
regions of large dnodes, and that these interior regions are not
overwritten by data belonging to other dnodes. A new test visits each
object in a dataset. It verifies that the actual dnode size matches what
was stored in the ztest block tag when it was created. It also verifies
that the unused bonus buffer space is filled with the expected data
patterns.
ZFS Test Suite
--------------
Added six new large dnode-specific tests, and integrated the dnodesize
property into existing tests for zfs allow and send/recv.
Send/Receive
------------
ZFS send streams for datasets containing large dnodes cannot be received
on pools that don't support the large_dnode feature. A send stream with
large dnodes sets a DMU_BACKUP_FEATURE_LARGE_DNODE flag which will be
unrecognized by an incompatible receiving pool so that the zfs receive
will fail gracefully.
While not implemented here, it may be possible to generate a
backward-compatible send stream from a dataset containing large
dnodes. The implementation may be tricky, however, because the send
object record for a large dnode would need to be resized to a 512
byte dnode, possibly kicking in a spill block in the process. This
means we would need to construct a new SA layout and possibly
register it in the SA layout object. The SA layout is normally just
sent as an ordinary object record. But if we are constructing new
layouts while generating the send stream we'd have to build the SA
layout object dynamically and send it at the end of the stream.
For sending and receiving between pools that do support large dnodes,
the drr_object send record type is extended with a new field to store
the dnode slot count. This field was repurposed from unused padding
in the structure.
ZIL Replay
----------
The dnode slot count is stored in the uppermost 8 bits of the lr_foid
field. The bits were unused as the object id is currently capped at
48 bits.
Resizing Dnodes
---------------
It should be possible to resize a dnode when it is dirtied if the
current dnodesize dataset property differs from the dnode's size, but
this functionality is not currently implemented. Clearly a dnode can
only grow if there are sufficient contiguous unused slots in the
dnode block, but it should always be possible to shrink a dnode.
Growing dnodes may be useful to reduce fragmentation in a pool with
many spill blocks in use. Shrinking dnodes may be useful to allow
sending a dataset to a pool that doesn't support the large_dnode
feature.
Feature Reference Counting
--------------------------
The reference count for the large_dnode pool feature tracks the
number of datasets that have ever contained a dnode of size larger
than 512 bytes. The first time a large dnode is created in a dataset
the dataset is converted to an extensible dataset. This is a one-way
operation and the only way to decrement the feature count is to
destroy the dataset, even if the dataset no longer contains any large
dnodes. The complexity of reference counting on a per-dnode basis was
too high, so we chose to track it on a per-dataset basis similarly to
the large_block feature.
Signed-off-by: Ned Bass <bass6@llnl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#3542
This is a new implementation of RAIDZ1/2/3 routines using x86_64
scalar, SSE, and AVX2 instruction sets. Included are 3 parity
generation routines (P, PQ, and PQR) and 7 reconstruction routines,
for all RAIDZ level. On module load, a quick benchmark of supported
routines will select the fastest for each operation and they will
be used at runtime. Original implementation is still present and
can be selected via module parameter.
Patch contains:
- specialized gen/rec routines for all RAIDZ levels,
- new scalar raidz implementation (unrolled),
- two x86_64 SIMD implementations (SSE and AVX2 instructions sets),
- fastest routines selected on module load (benchmark).
- cmd/raidz_test - verify and benchmark all implementations
- added raidz_test to the ZFS Test Suite
New zfs module parameters:
- zfs_vdev_raidz_impl (str): selects the implementation to use. On
module load, the parameter will only accept first 3 options, and
the other implementations can be set once module is finished
loading. Possible values for this option are:
"fastest" - use the fastest math available
"original" - use the original raidz code
"scalar" - new scalar impl
"sse" - new SSE impl if available
"avx2" - new AVX2 impl if available
See contents of `/sys/module/zfs/parameters/zfs_vdev_raidz_impl` to
get the list of supported values. If an implementation is not supported
on the system, it will not be shown. Currently selected option is
enclosed in `[]`.
Signed-off-by: Gvozden Neskovic <neskovic@gmail.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#4328
Since they both evaluate to zero, this is a semi-cosmetic change
but the latter is the proper value to use as an argument to
taskq_dispatch_delay().
Signed-off-by: Tim Chase <tim@chase2k.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#4393
Allow for easy turning of a pools reserved free space. Previous
versions of ZFS (v0.6.4 and earlier) held 1/64 of the pools capacity
in reserve. Commits 3d45fdd and 0c60cc3 increased this to 1/32.
Setting spa_slop_shift=6 will restore the previous default setting.
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#3724
Over the years the default values for the taskqs used on Linux have
differed slightly from illumos. In the vast majority of cases this
was done to avoid creating an obnoxious number of idle threads which
would pollute the process listing.
With the addition of support for dynamic taskqs all multi-threaded
queues should be created as dynamic taskqs. This allows us to get
the best of both worlds.
* The illumos default values for the I/O pipeline can be restored.
These values are known to work well for most workloads. The only
exception is the zio write interrupt taskq which is changed to
ZTI_P(12, 8). At least under Linux more threads has been shown
to improve performance, see commit 7e55f4e.
* Reduces the number of idle threads on the system when it's not
under heavy load. The maximum number of threads will only be
created when they are required.
* Remove the vdev_file_taskq and rely on the system_taskq instead
which is now dynamic and may have up to 64-threads. Again this
brings us back inline with upstream.
* Tasks dispatched with taskq_dispatch_ent() are allowed to use
dynamic taskqs. The Linux taskq implementation supports this.
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tim Chase <tim@chase2k.com>
Closes#3507
5027 zfs large block support
Reviewed by: Alek Pinchuk <pinchuk.alek@gmail.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Josef 'Jeff' Sipek <josef.sipek@nexenta.com>
Reviewed by: Richard Elling <richard.elling@richardelling.com>
Reviewed by: Saso Kiselkov <skiselkov.ml@gmail.com>
Reviewed by: Brian Behlendorf <behlendorf1@llnl.gov>
Approved by: Dan McDonald <danmcd@omniti.com>
References:
https://www.illumos.org/issues/5027https://github.com/illumos/illumos-gate/commit/b515258
Porting Notes:
* Included in this patch is a tiny ISP2() cleanup in zio_init() from
Illumos 5255.
* Unlike the upstream Illumos commit this patch does not impose an
arbitrary 128K block size limit on volumes. Volumes, like filesystems,
are limited by the zfs_max_recordsize=1M module option.
* By default the maximum record size is limited to 1M by the module
option zfs_max_recordsize. This value may be safely increased up to
16M which is the largest block size supported by the on-disk format.
At the moment, 1M blocks clearly offer a significant performance
improvement but the benefits of going beyond this for the majority
of workloads are less clear.
* The illumos version of this patch increased DMU_MAX_ACCESS to 32M.
This was determined not to be large enough when using 16M blocks
because the zfs_make_xattrdir() function will fail (EFBIG) when
assigning a TX. This was immediately observed under Linux because
all newly created files must have a security xattr created and
that was failing. Therefore, we've set DMU_MAX_ACCESS to 64M.
* On 32-bit platforms a hard limit of 1M is set for blocks due
to the limited virtual address space. We should be able to relax
this one the ABD patches are merged.
Ported-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#354
5056 ZFS deadlock on db_mtx and dn_holds
Author: Justin Gibbs <justing@spectralogic.com>
Reviewed by: Will Andrews <willa@spectralogic.com>
Reviewed by: Matt Ahrens <mahrens@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Approved by: Dan McDonald <danmcd@omniti.com>
References:
https://www.illumos.org/issues/5056https://github.com/illumos/illumos-gate/commit/bc9014e
Porting Notes:
sa_handle_get_from_db():
- the original patch includes an otherwise unmentioned fix for a
possible usage of an uninitialised variable
dmu_objset_open_impl():
- Under Illumos list_link_init() is the same as filling a list_node_t
with NULLs, so they don't notice if they miss doing list_link_init()
on a zero'd containing structure (e.g. allocated with kmem_zalloc as
here). Under Linux, not so much: an uninitialised list_node_t goes
"Boom!" some time later when it's used or destroyed.
dmu_objset_evict_dbufs():
- reduce stack usage using kmem_alloc()
Ported-by: Chris Dunlop <chris@onthe.net.au>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
This isn't required for the Linux port because the kernel tracks
if a module is busy. The prototype for spa_busy() is also removed
since its definition was already removed.
Signed-off-by: Isaac Huang <he.huang@intel.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#3262
The function spa_add_feature_stats() manipulates the shared nvlist
spa->spa_feat_stats in an unsafe concurrent manner. Add a mutex to
protect the list.
Signed-off-by: Ned Bass <bass6@llnl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#3335
Avoid issuing I/O to the pool when retrieving feature flags information.
Trying to read the ZAPs from disk means that zpool clear would hang if
the pool is suspended and recovery would require a reboot. To keep the
feature stats resident in memory, we hang a cached nvlist off of the
spa. It is built up from disk the first time spa_add_feature_stats() is
called, and refreshed thereafter using the cached feature reference
counts. spa_add_feature_stats() gets called at pool import time so we
can be sure the cached nvlist will be available if the pool is later
suspended.
Signed-off-by: Ned Bass <bass6@llnl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#3082
By marking DMU transaction processing contexts with PF_FSTRANS
we can revert the KM_PUSHPAGE -> KM_SLEEP changes. This brings
us back in line with upstream. In some cases this means simply
swapping the flags back. For others fnvlist_alloc() was replaced
by nvlist_alloc(..., KM_PUSHPAGE) and must be reverted back to
fnvlist_alloc() which assumes KM_SLEEP.
The one place KM_PUSHPAGE is kept is when allocating ARC buffers
which allows us to dip in to reserved memory. This is again the
same as upstream.
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Callers of kmem_alloc() which passed the KM_NODEBUG flag to suppress
the large allocation warning have been replaced by vmem_alloc() as
appropriate. The updated vmem_alloc() call will not print a warning
regardless of the size of the allocation.
A careful reader will notice that not all callers have been changed
to vmem_alloc(). Some have only had the KM_NODEBUG flag removed.
This was possible because the default warning threshold has been
increased to 32k. This is desirable because it minimizes the need
for Linux specific code changes.
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Add a table describing the debugging flags that can be set in the zfs_flags
module parameter. Also change the module_param type to 'uint' so users aren't
shown a negative value. The updated man page text is reproduced below for
convenience.
zfs_flags (int)
Set additional debugging flags. The following flags may be
bitwise-or'd together.
+-------------------------------------------------------+
|Value Symbolic Name |
| Description |
+-------------------------------------------------------+
| 1 ZFS_DEBUG_DPRINTF |
| Enable dprintf entries in the debug log. |
+-------------------------------------------------------+
| 2 ZFS_DEBUG_DBUF_VERIFY * |
| Enable extra dbuf verifications. |
+-------------------------------------------------------+
| 4 ZFS_DEBUG_DNODE_VERIFY * |
| Enable extra dnode verifications. |
+-------------------------------------------------------+
| 8 ZFS_DEBUG_SNAPNAMES |
| Enable snapshot name verification. |
+-------------------------------------------------------+
| 16 ZFS_DEBUG_MODIFY |
| Check for illegally modified ARC buffers. |
+-------------------------------------------------------+
| 32 ZFS_DEBUG_SPA |
| Enable spa_dbgmsg entries in the debug log. |
+-------------------------------------------------------+
| 64 ZFS_DEBUG_ZIO_FREE |
| Enable verification of block frees. |
+-------------------------------------------------------+
| 128 ZFS_DEBUG_HISTOGRAM_VERIFY |
| Enable extra spacemap histogram verifications. |
+-------------------------------------------------------+
* Requires debug build.
Default value: 0.
Signed-off-by: Ned Bass <bass6@llnl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#2988
This patch leverages Linux tracepoints from within the ZFS on Linux
code base. It also refactors the debug code to bring it back in sync
with Illumos.
The information exported via tracepoints can be used for a variety of
reasons (e.g. debugging, tuning, general exploration/understanding,
etc). It is advantageous to use Linux tracepoints as the mechanism to
export this kind of information (as opposed to something else) for a
number of reasons:
* A number of external tools can make use of our tracepoints
"automatically" (e.g. perf, systemtap)
* Tracepoints are designed to be extremely cheap when disabled
* It's one of the "accepted" ways to export this kind of
information; many other kernel subsystems use tracepoints too.
Unfortunately, though, there are a few caveats as well:
* Linux tracepoints appear to only be available to GPL licensed
modules due to the way certain kernel functions are exported.
Thus, to actually make use of the tracepoints introduced by this
patch, one might have to patch and re-compile the kernel;
exporting the necessary functions to non-GPL modules.
* Prior to upstream kernel version v3.14-rc6-30-g66cc69e, Linux
tracepoints are not available for unsigned kernel modules
(tracepoints will get disabled due to the module's 'F' taint).
Thus, one either has to sign the zfs kernel module prior to
loading it, or use a kernel versioned v3.14-rc6-30-g66cc69e or
newer.
Assuming the above two requirements are satisfied, lets look at an
example of how this patch can be used and what information it exposes
(all commands run as 'root'):
# list all zfs tracepoints available
$ ls /sys/kernel/debug/tracing/events/zfs
enable filter zfs_arc__delete
zfs_arc__evict zfs_arc__hit zfs_arc__miss
zfs_l2arc__evict zfs_l2arc__hit zfs_l2arc__iodone
zfs_l2arc__miss zfs_l2arc__read zfs_l2arc__write
zfs_new_state__mfu zfs_new_state__mru
# enable all zfs tracepoints, clear the tracepoint ring buffer
$ echo 1 > /sys/kernel/debug/tracing/events/zfs/enable
$ echo 0 > /sys/kernel/debug/tracing/trace
# import zpool called 'tank', inspect tracepoint data (each line was
# truncated, they're too long for a commit message otherwise)
$ zpool import tank
$ cat /sys/kernel/debug/tracing/trace | head -n35
# tracer: nop
#
# entries-in-buffer/entries-written: 1219/1219 #P:8
#
# _-----=> irqs-off
# / _----=> need-resched
# | / _---=> hardirq/softirq
# || / _--=> preempt-depth
# ||| / delay
# TASK-PID CPU# |||| TIMESTAMP FUNCTION
# | | | |||| | |
lt-zpool-30132 [003] .... 91344.200050: zfs_arc__miss: hdr...
z_rd_int/0-30156 [003] .... 91344.200611: zfs_new_state__mru...
lt-zpool-30132 [003] .... 91344.201173: zfs_arc__miss: hdr...
z_rd_int/1-30157 [003] .... 91344.201756: zfs_new_state__mru...
lt-zpool-30132 [003] .... 91344.201795: zfs_arc__miss: hdr...
z_rd_int/2-30158 [003] .... 91344.202099: zfs_new_state__mru...
lt-zpool-30132 [003] .... 91344.202126: zfs_arc__hit: hdr ...
lt-zpool-30132 [003] .... 91344.202130: zfs_arc__hit: hdr ...
lt-zpool-30132 [003] .... 91344.202134: zfs_arc__hit: hdr ...
lt-zpool-30132 [003] .... 91344.202146: zfs_arc__miss: hdr...
z_rd_int/3-30159 [003] .... 91344.202457: zfs_new_state__mru...
lt-zpool-30132 [003] .... 91344.202484: zfs_arc__miss: hdr...
z_rd_int/4-30160 [003] .... 91344.202866: zfs_new_state__mru...
lt-zpool-30132 [003] .... 91344.202891: zfs_arc__hit: hdr ...
lt-zpool-30132 [001] .... 91344.203034: zfs_arc__miss: hdr...
z_rd_iss/1-30149 [001] .... 91344.203749: zfs_new_state__mru...
lt-zpool-30132 [001] .... 91344.203789: zfs_arc__hit: hdr ...
lt-zpool-30132 [001] .... 91344.203878: zfs_arc__miss: hdr...
z_rd_iss/3-30151 [001] .... 91344.204315: zfs_new_state__mru...
lt-zpool-30132 [001] .... 91344.204332: zfs_arc__hit: hdr ...
lt-zpool-30132 [001] .... 91344.204337: zfs_arc__hit: hdr ...
lt-zpool-30132 [001] .... 91344.204352: zfs_arc__hit: hdr ...
lt-zpool-30132 [001] .... 91344.204356: zfs_arc__hit: hdr ...
lt-zpool-30132 [001] .... 91344.204360: zfs_arc__hit: hdr ...
To highlight the kind of detailed information that is being exported
using this infrastructure, I've taken the first tracepoint line from the
output above and reformatted it such that it fits in 80 columns:
lt-zpool-30132 [003] .... 91344.200050: zfs_arc__miss:
hdr {
dva 0x1:0x40082
birth 15491
cksum0 0x163edbff3a
flags 0x640
datacnt 1
type 1
size 2048
spa 3133524293419867460
state_type 0
access 0
mru_hits 0
mru_ghost_hits 0
mfu_hits 0
mfu_ghost_hits 0
l2_hits 0
refcount 1
} bp {
dva0 0x1:0x40082
dva1 0x1:0x3000e5
dva2 0x1:0x5a006e
cksum 0x163edbff3a:0x75af30b3dd6:0x1499263ff5f2b:0x288bd118815e00
lsize 2048
} zb {
objset 0
object 0
level -1
blkid 0
}
For the specific tracepoint shown here, 'zfs_arc__miss', data is
exported detailing the arc_buf_hdr_t (hdr), blkptr_t (bp), and
zbookmark_t (zb) that caused the ARC miss (down to the exact DVA!).
This kind of precise and detailed information can be extremely valuable
when trying to answer certain kinds of questions.
For anybody unfamiliar but looking to build on this, I found the XFS
source code along with the following three web links to be extremely
helpful:
* http://lwn.net/Articles/379903/
* http://lwn.net/Articles/381064/
* http://lwn.net/Articles/383362/
I should also node the more "boring" aspects of this patch:
* The ZFS_LINUX_COMPILE_IFELSE autoconf macro was modified to
support a sixth paramter. This parameter is used to populate the
contents of the new conftest.h file. If no sixth parameter is
provided, conftest.h will be empty.
* The ZFS_LINUX_TRY_COMPILE_HEADER autoconf macro was introduced.
This macro is nearly identical to the ZFS_LINUX_TRY_COMPILE macro,
except it has support for a fifth option that is then passed as
the sixth parameter to ZFS_LINUX_COMPILE_IFELSE.
These autoconf changes were needed to test the availability of the Linux
tracepoint macros. Due to the odd nature of the Linux tracepoint macro
API, a separate ".h" must be created (the path and filename is used
internally by the kernel's define_trace.h file).
* The HAVE_DECLARE_EVENT_CLASS autoconf macro was introduced. This
is to determine if we can safely enable the Linux tracepoint
functionality. We need to selectively disable the tracepoint code
due to the kernel exporting certain functions as GPL only. Without
this check, the build process will fail at link time.
In addition, the SET_ERROR macro was modified into a tracepoint as well.
To do this, the 'sdt.h' file was moved into the 'include/sys' directory
and now contains a userspace portion and a kernel space portion. The
dprintf and zfs_dbgmsg* interfaces are now implemented as tracepoint as
well.
Signed-off-by: Prakash Surya <surya1@llnl.gov>
Signed-off-by: Ned Bass <bass6@llnl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
4101 metaslab_debug should allow for fine-grained control
4102 space_maps should store more information about themselves
4103 space map object blocksize should be increased
4105 removing a mirrored log device results in a leaked object
4106 asynchronously load metaslab
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Adam Leventhal <ahl@delphix.com>
Reviewed by: Sebastien Roy <seb@delphix.com>
Approved by: Garrett D'Amore <garrett@damore.org>
Prior to this patch, space_maps were preferred solely based on the
amount of free space left in each. Unfortunately, this heuristic didn't
contain any information about the make-up of that free space, which
meant we could keep preferring and loading a highly fragmented space map
that wouldn't actually have enough contiguous space to satisfy the
allocation; then unloading that space_map and repeating the process.
This change modifies the space_map's to store additional information
about the contiguous space in the space_map, so that we can use this
information to make a better decision about which space_map to load.
This requires reallocating all space_map objects to increase their
bonus buffer size sizes enough to fit the new metadata.
The above feature can be enabled via a new feature flag introduced by
this change: com.delphix:spacemap_histogram
In addition to the above, this patch allows the space_map block size to
be increase. Currently the block size is set to be 4K in size, which has
certain implications including the following:
* 4K sector devices will not see any compression benefit
* large space_maps require more metadata on-disk
* large space_maps require more time to load (typically random reads)
Now the space_map block size can adjust as needed up to the maximum size
set via the space_map_max_blksz variable.
A bug was fixed which resulted in potentially leaking an object when
removing a mirrored log device. The previous logic for vdev_remove() did
not deal with removing top-level vdevs that are interior vdevs (i.e.
mirror) correctly. The problem would occur when removing a mirrored log
device, and result in the DTL space map object being leaked; because
top-level vdevs don't have DTL space map objects associated with them.
References:
https://www.illumos.org/issues/4101https://www.illumos.org/issues/4102https://www.illumos.org/issues/4103https://www.illumos.org/issues/4105https://www.illumos.org/issues/4106https://github.com/illumos/illumos-gate/commit/0713e23
Porting notes:
A handful of kmem_alloc() calls were converted to kmem_zalloc(). Also,
the KM_PUSHPAGE and TQ_PUSHPAGE flags were used as necessary.
Ported-by: Tim Chase <tim@chase2k.com>
Signed-off-by: Prakash Surya <surya1@llnl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#2488
Originally, vdev_file used system_taskq. This would cause a deadlock,
especially on system with few CPUs. The reason is that the prefetcher
threads, which are on system_taskq, will sometimes be blocked waiting
for I/O to finish. If the prefetcher threads consume all the tasks in
system_taskq, the I/O cannot be served and thus results in a deadlock.
We fix this by creating a dedicated vdev_file_taskq for vdev_file I/O.
Signed-off-by: Chunwei Chen <tuxoko@gmail.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#2270
The dva_get_dsize_sync() function incorrectly assumes that the call
to vdev_lookup_top() cannot fail. However, the NULL dereference at
clearly shows that under certain circumstances it is possible. Note
that offset 0x570 (1376) maps as expected to vd->vdev_deflate_ratio.
BUG: unable to handle kernel NULL pointer dereference at 00000570
crash> struct -o vdev
struct vdev {
[0] uint64_t vdev_id;
... ...
[1376] uint64_t vdev_deflate_ratio;
Given that this can happen this patch add the required error handling.
In the case where vdev_lookup_top() fails assume that no deflation
will occur for the DVA and use the asize.
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ned Bass <bass6@llnl.gov>
Signed-off-by: Alexey Zhuravlev <alexey.zhuravlev@intel.com>
Closes#1707Closes#1987Closes#1891
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
The spa_label_features nvlist is used in the sync context during pool
version upgrade.
Signed-off-by: Tim Chase <tim@chase2k.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#2168
Back the allocations for ddt tables+entries and l2arc headers with
kmem caches. This will reduce the cost of allocating these commonly
used structures and allow for greater visibility of them through the
/proc/spl/kmem/slab interface.
Signed-off-by: John Layman <jlayman@sagecloud.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#1893
The vast majority of these changes are in Linux specific code.
They are the result of not having an automated style checker to
validate the code when it was originally written. Others were
caused when the common code was slightly adjusted for Linux.
This patch contains no functional changes. It only refreshes
the code to conform to style guide.
Everyone submitting patches for inclusion upstream should now
run 'make checkstyle' and resolve any warning prior to opening
a pull request. The automated builders have been updated to
fail a build if when 'make checkstyle' detects an issue.
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#1821
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/4045illumos/illumos-gate@69962b5647
Ported-by: Ned Bass <bass6@llnl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#1913
LOLi