In the case of one pool being built on another pool, we want
to make sure we don't end up throttling the lower (backing)
pool when the upper pool is the majority contributor to dirty
data. To insure we make forward progress during throttling, we
also check the current pool's net dirty data and only throttle
if it exceeds zfs_arc_pool_dirty_percent of the anonymous dirty
data in the cache.
Authored by: Don Brady <don.brady@delphix.com>
Reviewed by: Sebastien Roy <sebastien.roy@delphix.com>
Reviewed by: Matt Ahrens <matt@delphix.com>
Reviewed by: Prashanth Sreenivasa <pks@delphix.com>
Approved by: Robert Mustacchi <rm@joyent.com>
Ported-by: Brian Behlendorf <behlendorf1@llnl.gov>
Porting Notes:
* The new global variables zfs_arc_dirty_limit_percent,
zfs_arc_anon_limit_percent, and zfs_arc_pool_dirty_percent
were intentially not added as tunable module parameters.
OpenZFS-issue: https://illumos.org/issues/9465
OpenZFS-commit: https://github.com/openzfs/openzfs/commit/d6a4c3efCloses#7749
Currently, the decryption and block authentication code in
the ZIO / ARC layers is a bit inconsistent with regards to
the ereports that are produces and the error codes that are
passed to calling functions. This patch ensures that all of
these errors (which begin as ECKSUM) are converted to EIO
before they leave the ZIO or ARC layer and that ereports
are correctly generated on each decryption / authentication
failure.
In addition, this patch fixes a bug in zio_decrypt() where
ECKSUM never gets written to zio->io_error.
Reviewed by: Matt Ahrens <matt@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes#7372
When the compressed ARC feature was added in commit d3c2ae1
the method of reference counting in the ARC was modified. As
part of this accounting change the arc_buf_add_ref() function
was removed entirely.
This would have be fine but the arc_buf_add_ref() function
served a second undocumented purpose of updating the ARC access
information when taking a hold on a dbuf. Without this logic
in place a cached dbuf would not migrate its associated
arc_buf_hdr_t to the MFU list. This would negatively impact
the ARC hit rate, particularly on systems with a small ARC.
This change reinstates the missing call to arc_access() from
dbuf_hold() by implementing a new arc_buf_access() function.
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Tim Chase <tim@chase2k.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#6171Closes#6852Closes#6989
Currently, scrubs and resilvers can take an extremely
long time to complete. This is largely due to the fact
that zfs scans process pools in logical order, as
determined by each block's bookmark. This makes sense
from a simplicity perspective, but blocks in zfs are
often scattered randomly across disks, particularly
due to zfs's copy-on-write mechanisms.
This patch improves performance by splitting scrubs
and resilvers into a metadata scanning phase and an IO
issuing phase. The metadata scan reads through the
structure of the pool and gathers an in-memory queue
of I/Os, sorted by size and offset on disk. The issuing
phase will then issue the scrub I/Os as sequentially as
possible, greatly improving performance.
This patch also updates and cleans up some of the scan
code which has not been updated in several years.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Authored-by: Saso Kiselkov <saso.kiselkov@nexenta.com>
Authored-by: Alek Pinchuk <apinchuk@datto.com>
Authored-by: Tom Caputi <tcaputi@datto.com>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes#3625Closes#6256
Commit d3c2ae1 introduced a dbuf cache with a default size of the
minimum of 100M or 1/32 maximum ARC size. (These figures may be adjusted
using dbuf_cache_max_bytes and dbuf_cache_max_shift.) The dbuf cache
is counted as metadata for the purposes of ARC size calculations.
On a 1GB box the ARC maximum size defaults to c_max 493M which gives a
dbuf cache default minimum size of 15.4M, and the ARC metadata defaults
to minimum 16M. I.e. the dbuf cache is an significant proportion of the
minimum metadata size. With other overheads involved this actually means
the ARC metadata doesn't get down to the minimum.
This patch dynamically scales the dbuf cache to the target ARC size
instead of statically scaling it to the maximum ARC size. (The scale is
still set by dbuf_cache_max_shift and the maximum size is still fixed by
dbuf_cache_max_bytes.) Using the target ARC size rather than the current
ARC size is done to help the ARC reach the target rather than simply
focusing on the current size.
Reviewed-by: Chunwei Chen <tuxoko@gmail.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Signed-off-by: Chris Dunlop <chris@onthe.net.au>
Issue #6506Closes#6561
This change incorporates three major pieces:
The first change is a keystore that manages wrapping
and encryption keys for encrypted datasets. These
commands mostly involve manipulating the new
DSL Crypto Key ZAP Objects that live in the MOS. Each
encrypted dataset has its own DSL Crypto Key that is
protected with a user's key. This level of indirection
allows users to change their keys without re-encrypting
their entire datasets. The change implements the new
subcommands "zfs load-key", "zfs unload-key" and
"zfs change-key" which allow the user to manage their
encryption keys and settings. In addition, several new
flags and properties have been added to allow dataset
creation and to make mounting and unmounting more
convenient.
The second piece of this patch provides the ability to
encrypt, decyrpt, and authenticate protected datasets.
Each object set maintains a Merkel tree of Message
Authentication Codes that protect the lower layers,
similarly to how checksums are maintained. This part
impacts the zio layer, which handles the actual
encryption and generation of MACs, as well as the ARC
and DMU, which need to be able to handle encrypted
buffers and protected data.
The last addition is the ability to do raw, encrypted
sends and receives. The idea here is to send raw
encrypted and compressed data and receive it exactly
as is on a backup system. This means that the dataset
on the receiving system is protected using the same
user key that is in use on the sending side. By doing
so, datasets can be efficiently backed up to an
untrusted system without fear of data being
compromised.
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Jorgen Lundman <lundman@lundman.net>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes#494Closes#5769
The global tunable zfs_arc_num_sublists_per_state is used by the ARC and
the dbuf cache, and other users are planned. We should change this
tunable to be common to all multilists. This tuning may be overridden
on a per-multilist basis.
Reviewed-by: Pavel Zakharov <pavel.zakharov@delphix.com>
Reviewed-by: Dan Kimmel <dan.kimmel@delphix.com>
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Matthew Ahrens <mahrens@delphix.com>
Closes#5764
Authored by: George Wilson <george.wilson@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Dan Kimmel <dan.kimmel@delphix.com>
Reviewed by: Matt Ahrens <mahrens@delphix.com>
Reviewed by: Paul Dagnelie <pcd@delphix.com>
Reviewed by: Tom Caputi <tcaputi@datto.com>
Reviewed by: Brian Behlendorf <behlendorf1@llnl.gov>
Ported by: David Quigley <david.quigley@intel.com>
This review covers the reading and writing of compressed arc headers, sharing
data between the arc_hdr_t and the arc_buf_t, and the implementation of a new
dbuf cache to keep frequently access data uncompressed.
I've added a new member to l1 arc hdr called b_pdata. The b_pdata always hangs
off the arc_buf_hdr_t (if an L1 hdr is in use) and points to the physical block
for that DVA. The physical block may or may not be compressed. If compressed
arc is enabled and the block on-disk is compressed, then the b_pdata will match
the block on-disk and remain compressed in memory. If the block on disk is not
compressed, then neither will the b_pdata. Lastly, if compressed arc is
disabled, then b_pdata will always be an uncompressed version of the on-disk
block.
Typically the arc will cache only the arc_buf_hdr_t and will aggressively evict
any arc_buf_t's that are no longer referenced. This means that the arc will
primarily have compressed blocks as the arc_buf_t's are considered overhead and
are always uncompressed. When a consumer reads a block we first look to see if
the arc_buf_hdr_t is cached. If the hdr is cached then we allocate a new
arc_buf_t and decompress the b_pdata contents into the arc_buf_t's b_data. If
the hdr already has a arc_buf_t, then we will allocate an additional arc_buf_t
and bcopy the uncompressed contents from the first arc_buf_t to the new one.
Writing to the compressed arc requires that we first discard the b_pdata since
the physical block is about to be rewritten. The new data contents will be
passed in via an arc_buf_t (uncompressed) and during the I/O pipeline stages we
will copy the physical block contents to a newly allocated b_pdata.
When an l2arc is inuse it will also take advantage of the b_pdata. Now the
l2arc will always write the contents of b_pdata to the l2arc. This means that
when compressed arc is enabled that the l2arc blocks are identical to those
stored in the main data pool. This provides a significant advantage since we
can leverage the bp's checksum when reading from the l2arc to determine if the
contents are valid. If the compressed arc is disabled, then we must first
transform the read block to look like the physical block in the main data pool
before comparing the checksum and determining it's valid.
OpenZFS-issue: https://www.illumos.org/issues/6950
OpenZFS-commit: https://github.com/openzfs/openzfs/commit/7fc10f0
Issue #5078
Metadata-intensive workloads can cause the ARC to become permanently
filled with dnode_t objects as they're pinned by the VFS layer.
Subsequent data-intensive workloads may only benefit from about
25% of the potential ARC (arc_c_max - arc_meta_limit).
In order to help track metadata usage more precisely, the other_size
metadata arcstat has replaced with dbuf_size, dnode_size and bonus_size.
The new zfs_arc_dnode_limit tunable, which defaults to 10% of
zfs_arc_meta_limit, defines the minimum number of bytes which is desirable
to be consumed by dnodes. Attempts to evict non-metadata will trigger
async prune tasks if the space used by dnodes exceeds this limit.
The new zfs_arc_dnode_reduce_percent tunable specifies the amount by
which the excess dnode space is attempted to be pruned as a percentage of
the amount by which zfs_arc_dnode_limit is being exceeded. By default,
it tries to unpin 10% of the dnodes.
The problem of dnode metadata pinning was observed with the following
testing procedure (in this example, zfs_arc_max is set to 4GiB):
- Create a large number of small files until arc_meta_used exceeds
arc_meta_limit (3GiB with default tuning) and arc_prune
starts increasing.
- Create a 3GiB file with dd. Observe arc_mata_used. It will still
be around 3GiB.
- Repeatedly read the 3GiB file and observe arc_meta_limit as before.
It will continue to stay around 3GiB.
With this modification, space for the 3GiB file is gradually made
available as subsequent demands on the ARC are made. The previous behavior
can be restored by setting zfs_arc_dnode_limit to the same value as the
zfs_arc_meta_limit.
Signed-off-by: Tim Chase <tim@chase2k.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Issue #4345
Issue #4512
Issue #4773Closes#4858
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Boris Protopopov <bprotopopov@hotmail.com>
Approved by: Richard Lowe <richlowe@richlowe.net>a
Ported by: Boris Protopopov <bprotopopov@actifio.com>
Signed-off-by: Boris Protopopov <bprotopopov@actifio.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
OpenZFS-issue: https://www.illumos.org/issues/6513
OpenZFS-commit: https://github.com/openzfs/openzfs/commit/8df0bcf0
If a ZFS object contains a hole at level one, and then a data block is
created at level 0 underneath that l1 block, l0 holes will be created.
However, these l0 holes do not have the birth time property set; as a
result, incremental sends will not send those holes.
Fix is to modify the dbuf_read code to fill in birth time data.
6214 zpools going south
Reviewed by: Igor Kozhukhov <ikozhukhov@gmail.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Dan McDonald <danmcd@omniti.com>
Reviewed by: Saso Kiselkov <skiselkov.ml@gmail.com>
References:
https://www.illumos.org/issues/6214http://cr.illumos.org/~webrev/sensille/6214_zpools_going_south/
Porting Notes:
Reintroduce b_compress to the l2arc_buf_hdr_t. In commit b9541d6
the compression flags were moved to the generic b_flags in the
arc_buf_hdr_t. This is a problem because l2arc_compress_buf()
may manipulate the compression flags and this can only be done
safely under the hash lock which is not held. See Illumos 6214
for a detailed analysis of the race.
HDR_GET_COMPRESS() macro was removed from arc_buf_info().
Ported-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#3757
Commit 49ddb31506 added the
zfs_arc_average_blocksize parameter to allow control over the size of
the arc hash table. The dbuf hash table's size should be determined
similarly.
Signed-off-by: Tim Chase <tim@chase2k.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#3721
As described in the comment above arc_adapt_thread() it is critical
that the arc_adapt_thread() function never sleep while holding a hash
lock. This behavior was possible in the Linux implementation because
the arc_prune() logic was implemented to be synchronous. Under
illumos the analogous dnlc_reduce_cache() function is asynchronous.
To address this the arc_do_user_prune() function is has been reworked
in to two new functions as follows:
* arc_prune_async() is an asynchronous implementation which dispatches
the prune callback to be run by the system taskq. This makes it
suitable to use in the context of the arc_adapt_thread().
* arc_prune() is a synchronous implementation which depends on the
arc_prune_async() implementation but blocks until the outstanding
callbacks complete. This is used in arc_kmem_reap_now() where it
is safe, and expected, that memory will be freed.
This patch additionally adds the zfs_arc_meta_strategy module option
while allows the meta reclaim strategy to be configured. It defaults
to a balanced strategy which has been proved to work well under Linux
but the illumos meta-only strategy can be enabled.
Signed-off-by: Tim Chase <tim@chase2k.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Richard Elling <richard.elling@richardelling.com>
Approved by: Dan McDonald <danmcd@omniti.com>
Porting notes and other significant code changes:
The illumos 5368 patch (ARC should cache more metadata), which
was never picked up by ZoL, is mostly reverted by this patch.
Since ZoL relies on the kernel asynchronously calling the shrinker to
actually reap memory, the shrinker wakes up arc_reclaim_waiters_cv every
time it runs.
The arc_adapt_thread() function no longer calls arc_do_user_evicts()
since the newly-added arc_user_evicts_thread() calls it periodically.
Notable conflicting ZoL commits which conflicted with this patch or
whose effects are either duplicated or un-done by this patch:
302f753 - Integrate ARC more tightly with Linux
39e055c - Adjust arc_p based on "bytes" in arc_shrink
f521ce1 - Allow "arc_p" to drop to zero or grow to "arc_c"
77765b5 - Remove "arc_meta_used" from arc_adjust calculation
94520ca - Prune metadata from ghost lists in arc_adjust_meta
Trace support for multilist_insert() and multilist_remove() has been
added and produces the following output:
fio-12498 [077] .... 112936.448324: zfs_multilist__insert: ml { offset 240 numsublists 80 sublistidx 63 }
fio-12498 [077] .... 112936.448347: zfs_multilist__remove: ml { offset 240 numsublists 80 sublistidx 29 }
The following arcstats have been removed:
recycle_miss - Used by arcstat.py and arc_summary.py, both of which
have been updated appropriately.
l2_writes_hdr_miss
The following arcstats have been added:
evict_not_enough - Number of times arc_evict_state() was unable to
evict enough buffers to reach its target amount.
evict_l2_skip - Number of times arc_evict_hdr() skipped eviction
because it was being written to the l2arc.
l2_writes_lock_retry - Replaces l2_writes_hdr_miss. Number of times
l2arc_write_done() failed to acquire hash_lock (and re-tries).
arc_meta_min - Shows the value of the zfs_arc_meta_min module
parameter (see below).
The "index" column of the "dbuf" kstat has been removed since it doesn't
have a direct analog in the new multilist scheme. Additional multilist-
related stats could be added in the future but would likely require
extensions to the mulilist API.
The following module parameters have been added:
zfs_arc_evict_batch_limit - Number of ARC headers to free per sub-list
before moving on to the next sub-list.
zfs_arc_meta_min - Enforce a floor on the amount of metadata in
the ARC.
zfs_arc_num_sublists_per_state - Number of multilist sub-lists per
ARC state.
zfs_arc_overflow_shift - Controls amount by which the ARC must exceed
the target size to be considered "overflowing".
Ported-by: Tim Chase <tim@chase2k.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov
5408 managing ZFS cache devices requires lots of RAM
Reviewed by: Christopher Siden <christopher.siden@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Don Brady <dev.fs.zfs@gmail.com>
Reviewed by: Josef 'Jeff' Sipek <josef.sipek@nexenta.com>
Approved by: Garrett D'Amore <garrett@damore.org>
Porting notes:
Due to the restructuring of the ARC-related structures, this
patch conflicts with at least the following existing ZoL commits:
6e1d7276c9
Fix inaccurate arcstat_l2_hdr_size calculations
The ARC_SPACE_HDRS constant no longer exists and has been
somewhat equivalently replaced by HDR_L2ONLY_SIZE.
e0b0ca983d
Add visibility in to cached dbufs
The new layering of l{1,2}arc_buf_hdr_t within the arc_buf_hdr
struct requires additional structure member names to be used
when referencing the inner items. Also, the presence of L1 or L2
inner member is indicated by flags using the new HDR_HAS_L{1,2}HDR
macros.
Ported by: Tim Chase <tim@chase2k.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
5369 arc flags should be an enum
5370 consistent arc_buf_hdr_t naming scheme
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Alex Reece <alex.reece@delphix.com>
Reviewed by: Sebastien Roy <sebastien.roy@delphix.com>
Reviewed by: Richard Elling <richard.elling@richardelling.com>
Approved by: Richard Lowe <richlowe@richlowe.net>
Porting notes:
ZoL has moved some ARC definitions into arc_impl.h.
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Ported by: Tim Chase <tim@chase2k.com>
The general strategy used by ZFS to verify that blocks are valid is
to checksum everything. This has the advantage of being extremely
robust and generically applicable regardless of the contents of
the block. If a blocks checksum is valid then its contents are
trusted by the higher layers.
This system works exceptionally well as long as bad data is never
written with a valid checksum. If this does somehow occur due to
a software bug or a memory bit-flip on a non-ECC system it may
result in kernel panic.
One such place where this could occur is if somehow the logical
size stored in a block pointer exceeds the maximum block size.
This will result in an attempt to allocate a buffer greater than
the maximum block size causing a system panic.
To prevent this from happening the arc_read() function has been
updated to detect this specific case. If a block pointer with an
invalid logical size is passed it will treat the block as if it
contained a checksum error.
Signed-off-by: Tim Chase <tim@chase2k.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#2678
4631 zvol_get_stats triggering too many reads
Reviewed by: Adam Leventhal <ahl@delphix.com>
Reviewed by: Sebastien Roy <sebastien.roy@delphix.com>
Reviewed by: Matt Ahrens <mahrens@delphix.com>
Approved by: Dan McDonald <danmcd@omniti.com>
References:
https://www.illumos.org/issues/4631https://github.com/illumos/illumos-gate/commit/bbfa8ea
Ported-by: Boris Protopopov <bprotopopov@hotmail.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#2612Closes#2480
4914 zfs on-disk bookmark structure should be named *_phys_t
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Christopher Siden <christopher.siden@delphix.com>
Reviewed by: Richard Lowe <richlowe@richlowe.net>
Reviewed by: Saso Kiselkov <skiselkov.ml@gmail.com>
Approved by: Robert Mustacchi <rm@joyent.com>
References:
https://www.illumos.org/issues/4914https://github.com/illumos/illumos-gate/commit/7802d7b
Porting notes:
There were a number of zfsonlinux-specific uses of zbookmark_t which
needed to be updated. This should reduce the likelihood of further
problems like issue #2094 from occurring.
Ported by: Tim Chase <tim@chase2k.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#2558
Previously, the "data_size" field in the arcstats kstat contained the
amount of cached "metadata" and "data" in the ARC. The problem is this
then made it difficult to extract out just the "metadata" size, or just
the "data" size.
To make it easier to distinguish the two values, "data_size" has been
modified to count only buffers of type ARC_BUFC_DATA, and "meta_size"
was added to count only buffers of type ARC_BUFC_METADATA. If one wants
the old "data_size" value, simply sum the new "data_size" and
"meta_size" values.
Signed-off-by: Prakash Surya <surya1@llnl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Issue #2110
To maintain a strict limit on the metadata contained in the arc, while
preventing the arc buffer headers from completely consuming the
"arc_meta_used" space, we need to evict metadata buffers from the arc's
ghost lists along with the regular lists.
This change modifies arc_adjust_meta such that it more closely models
the adjustments made in arc_adjust. "arc_meta_used" is used similarly to
"arc_size", and "arc_meta_limit" is used similarly to "arc_c".
Testing metadata intensive workloads (e.g. creating, copying, and
removing millions of small files and/or directories) has shown this
change to make a dramatic improvement to the hit rate maintained in the
arc. While I think there is still room for improvement, this is a big
step in the right direction.
In addition, zpl_free_cached_objects was made into a no-op as I'm not
yet sure how to properly implement that function.
Signed-off-by: Prakash Surya <surya1@llnl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Issue #2110
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
3112 ztest does not honor ZFS_DEBUG
3113 ztest should use watchpoints to protect frozen arc bufs
3114 some leaked nvlists in zfsdev_ioctl
Reviewed by: Adam Leventhal <ahl@delphix.com>
Reviewed by: Matt Amdur <Matt.Amdur@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Christopher Siden <chris.siden@delphix.com>
Approved by: Eric Schrock <eric.schrock@delphix.com>
References:
https://www.illumos.org/issues/3112https://www.illumos.org/issues/3113https://www.illumos.org/issues/3114illumos/illumos-gate@cd1c8b85eb
The /proc/self/cmd watchpoint interface is specific to Solaris.
Therefore, the #3113 implementation was reworked to use the more
portable mprotect(2) system call. When the pages are watched they
are marked read-only for protection. Any write to the protected
address range immediately trigger a SIGSEGV. The pages are marked
writable again when they are unwatched.
Ported-by: Brian Behlendorf <behlendorf1@llnl.gov>
Issue #1489
Currently there is no mechanism to inspect which dbufs are being
cached by the system. There are some coarse counters in arcstats
by they only give a rough idea of what's being cached. This patch
aims to improve the current situation by adding a new dbufs kstat.
When read this new kstat will walk all cached dbufs linked in to
the dbuf_hash. For each dbuf it will dump detailed information
about the buffer. It will also dump additional information about
the referenced arc buffer and its related dnode. This provides a
more complete view in to exactly what is being cached.
With this generic infrastructure in place utilities can be written
to post-process the data to understand exactly how the caching is
working. For example, the data could be processed to show a list
of all cached dnodes and how much space they're consuming. Or a
similar list could be generated based on dnode type. Many other
ways to interpret the data exist based on what kinds of questions
you're trying to answer.
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Prakash Surya <surya1@llnl.gov>
3137 L2ARC compression
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Approved by: Dan McDonald <danmcd@nexenta.com>
References:
illumos/illumos-gate@aad02571bchttps://www.illumos.org/issues/3137http://wiki.illumos.org/display/illumos/L2ARC+Compression
Notes for Linux port:
A l2arc_nocompress module option was added to prevent the
compression of l2arc buffers regardless of how a dataset's
compression property is set. This allows the legacy behavior
to be preserved.
Ported by: James H <james@kagisoft.co.uk>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#1379
These functions are used in neither Illumos nor ZFSOnLinux. They appear
to have been replaced by arc_buf_alloc()/arc_buf_free(), so lets remove
them.
Signed-off-by: Richard Yao <ryao@gentoo.org>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Issue #1614
3805 arc shouldn't cache freed blocks
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Christopher Siden <christopher.siden@delphix.com>
Reviewed by: Richard Elling <richard.elling@dey-sys.com>
Reviewed by: Will Andrews <will@firepipe.net>
Approved by: Dan McDonald <danmcd@nexenta.com>
References:
illumos/illumos-gate@6e6d5868f5https://www.illumos.org/issues/3805
ZFS should proactively evict freed blocks from the cache.
On dcenter, we saw that we were caching ~256GB of metadata, while the
pool only had <4GB of metadata on disk. We were wasting about half the
system's RAM (252GB) on blocks that have been freed.
Even though these freed blocks will never be used again, and thus will
eventually be evicted, this causes us to use memory inefficiently for 2
reasons:
1. A block that is freed has no chance of being accessed again, but will
be kept in memory preferentially to a block that was accessed before it
(and is thus older) but has not been freed and thus has at least some
chance of being accessed again.
2. We partition the ARC into several buckets:
user data that has been accessed only once (MRU)
metadata that has been accessed only once (MRU)
user data that has been accessed more than once (MFU)
metadata that has been accessed more than once (MFU)
The user data vs metadata split is somewhat arbitrary, and the primary
control on how much memory is used to cache data vs metadata is to
simply try to keep the proportion the same as it has been in the past
(each bucket "evicts against" itself). The secondary control is to
evict data before evicting metadata.
Because of this bucketing, we may end up with one bucket mostly
containing freed blocks that are very old, while another bucket has more
recently accessed, still-allocated blocks. Data in the useful bucket
(with still-allocated blocks) may be evicted in preference to data in
the useless bucket (with old, freed blocks).
On dcenter, we saw that the MFU metadata bucket was 230MB, while the MFU
data bucket was 27GB and the MRU metadata bucket was 256GB. However,
the vast majority of data in the MRU metadata bucket (256GB) was freed
blocks, and thus useless. Meanwhile, the MFU metadata bucket (230MB)
was constantly evicting useful blocks that will be soon needed.
The problem of cache segmentation is a larger problem that needs more
investigation. However, if we stop caching freed blocks, it should
reduce the impact of this more fundamental issue.
Ported-by: Richard Yao <ryao@cs.stonybrook.edu>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#1503
Commit c409e4647f introduced a
number of module parameters. This required several types to be
changed to accomidate the required module parameters Linux macros.
Unfortunately, arc.c contained its own extern definition of the
zfs_write_limit_max variable and its type was not updated to be
consistent with its dsl_pool.c counterpart. If the variable had
been properly marked extern in a common header, then gcc would
have generated a warning and this would not have slipped through.
The result of this was that the ARC unconditionally expected
zfs_write_limit_max to be 64-bit. Unfortunately, the largest size
integer module parameter that Linux supports is unsigned long, which
varies in size depending on the host system's native word size. The
effect was that on 32-bit systems, ARC incorrectly performed 64-bit
operations on a 32-bit value by reading the neighboring 32 bits as
the upper 32 bits of the 64-bit value.
We correct that by changing the extern declaration to use the unsigned
long type and move these extern definitions in to the common arc.h
header. This should make ARC correctly treat zfs_write_limit_max as a
32-bit value on 32-bit systems.
Reported-by: Jorgen Lundman <lundman@lundman.net>
Signed-off-by: Richard Yao <ryao@cs.stonybrook.edu>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#749
The Linux 3.1 kernel has introduced the concept of per-filesystem
shrinkers which are directly assoicated with a super block. Prior
to this change there was one shared global shrinker.
The zfs code relied on being able to call the global shrinker when
the arc_meta_limit was exceeded. This would cause the VFS to drop
references on a fraction of the dentries in the dcache. The ARC
could then safely reclaim the memory used by these entries and
honor the arc_meta_limit. Unfortunately, when per-filesystem
shrinkers were added the old interfaces were made unavailable.
This change adds support to use the new per-filesystem shrinker
interface so we can continue to honor the arc_meta_limit. The
major benefit of the new interface is that we can now target
only the zfs filesystem for dentry and inode pruning. Thus we
can minimize any impact on the caching of other filesystems.
In the context of making this change several other important
issues related to managing the ARC were addressed, they include:
* The dnlc_reduce_cache() function which was called by the ARC
to drop dentries for the Posix layer was replaced with a generic
zfs_prune_t callback. The ZPL layer now registers a callback to
drop these dentries removing a layering violation which dates
back to the Solaris code. This callback can also be used by
other ARC consumers such as Lustre.
arc_add_prune_callback()
arc_remove_prune_callback()
* The arc_reduce_dnlc_percent module option has been changed to
arc_meta_prune for clarity. The dnlc functions are specific to
Solaris's VFS and have already been largely eliminated already.
The replacement tunable now represents the number of bytes the
prune callback will request when invoked.
* Less aggressively invoke the prune callback. We used to call
this whenever we exceeded the arc_meta_limit however that's not
strictly correct since it results in over zeleous reclaim of
dentries and inodes. It is now only called once the arc_meta_limit
is exceeded and every effort has been made to evict other data from
the ARC cache.
* More promptly manage exceeding the arc_meta_limit. When reading
meta data in to the cache if a buffer was unable to be recycled
notify the arc_reclaim thread to invoke the required prune.
* Added arcstat_prune kstat which is incremented when the ARC
is forced to request that a consumer prune its cache. Remember
this will only occur when the ARC has no other choice. If it
can evict buffers safely without invoking the prune callback
it will.
* This change is also expected to resolve the unexpect collapses
of the ARC cache. This would occur because when exceeded just the
arc_meta_limit reclaim presure would be excerted on the arc_c
value via arc_shrink(). This effectively shrunk the entire cache
when really we just needed to reclaim meta data.
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#466Closes#292
One of the neat tricks an autoconf style project is capable of
is allow configurion/building in a directory other than the
source directory. The major advantage to this is that you can
build the project various different ways while making changes
in a single source tree.
For example, this project is designed to work on various different
Linux distributions each of which work slightly differently. This
means that changes need to verified on each of those supported
distributions perferably before the change is committed to the
public git repo.
Using nfs and custom build directories makes this much easier.
I now have a single source tree in nfs mounted on several different
systems each running a supported distribution. When I make a
change to the source base I suspect may break things I can
concurrently build from the same source on all the systems each
in their own subdirectory.
wget -c http://github.com/downloads/behlendorf/zfs/zfs-x.y.z.tar.gz
tar -xzf zfs-x.y.z.tar.gz
cd zfs-x-y-z
------------------------- run concurrently ----------------------
<ubuntu system> <fedora system> <debian system> <rhel6 system>
mkdir ubuntu mkdir fedora mkdir debian mkdir rhel6
cd ubuntu cd fedora cd debian cd rhel6
../configure ../configure ../configure ../configure
make make make make
make check make check make check make check
This change also moves many of the include headers from individual
incude/sys directories under the modules directory in to a single
top level include directory. This has the advantage of making
the build rules cleaner and logically it makes a bit more sense.
