Currently, a region_listener is added during construction and removed
during destruction. This was done to mimick the old region(region_group&)
constructor, as region_listener replaces region_group.
However, this makes moving the binomial heap handle outside logalloc
difficult. The natural place for the handle is in a derived class
of logalloc::region (e.g. memtable), but members of this derived class
will be destroyed earlier than the logalloc::region here. We could play
trickes with an earlier base class but it's better to just decouple
region lifecycle from listener lifecycle.
Do that be adding listen()/unlisten() methods. Some small awkwardness
remains in that merge() implicitly unlistens (see comment in
region::unlisten).
Unit tests are adjusted.
region_group is an abstraction that allows accounting for groups of
regions, but the cost/benefit ratio of maintaining the abstraction
is poor. Each time we need to change decision algorithm of memtable
flushing (admittedly rarely), we need to distill that into an abstraction
for region_groups and then use it. An example is virtual regions groups;
we wanted to account for the partially flushed memtables and had to
invent region groups to stand in their place.
Rather than continuing to invest in the abstraction, break it now
and move it to the memtable dirty memory manager which is responsible
for making those decisions. The relevant code is moved to
dirty_memory_manager.hh and dirty_memory_manager.cc (new file), and
a new unit test file is added as well.
A downside of the change is that unit testing will be more difficult.
Right now tracker_reclaim_lock uses tracker::impl::reclaiming_lock,
which won't be visible if we want to expose tracker_reclaim_lock and
use it from another translation unit. However, it's simple to switch
to an implementation that doesn't require an unknown-size data member,
and instead increment a counter via a pointer, so do that.
- add conversions between region and region_impl
- add accessor for the binomial heap handle
- add accessor for region_impl::id()
- remove friend declarations
This helps in moving region_group to a different source file, where
the definitions of region_impl will not be visible.
As a first step in moving region_group away from logalloc, decouple
communications between region and region_group. We introduce region_listener,
that listens for the events that region passed directly to region_group.
A region_group now installs a region_listener in a region, instead of
having region know about the region_group directly.
This decoupling is still leaky:
- merge() chooses to forget the merged-from region's region_listener.
This happens to be suitable for the only user of merge().
- We're still embedding the binomial heap handle, used by region_group
to keep track of region sizes, in regions. A complete decoupling would
transfer that responsibility to region_group.
Rename reclaim_timer::_reserve_segments to _segments_to_release
as it is clearer and more suitable for later patches
that will add reclaim_timers in more functions.
Signed-off-by: Benny Halevy <bhalevy@scylladb.com>
"
Some templates put constraints onto the involved types with the help of
static assertions. Having them in form of concepts is much better.
tests: unit(dev)
"
* 'br-static-assert-to-concept' of https://github.com/xemul/scylla:
sstables: Remove excessive type-match assertions
mutation_reader: Sanitize invocable asserion and concept
code: Convert is_future result_of assertions into invoke_result concept
code: Convert is_same+result_of assertions into invocable concepts
code: Convert nothrow construction assertions into concepts
code: Convert is_integral assertions to concepts
allocated through the region
It indicates alloc-dealloc mismatch, and can cause other problems in
the systems like unable to reclaim memory. We want to catch this at
the deallocation site to be able to quickly indentify the offender.
Misbehavior of this sort can cause fake OOMs due to underflow of
_non_lsa_memory_in_use. When it underflows enough,
shard_segment_pool.total_memory() will become 0 and memory reclamation
will stop doing anything.
Refs #10056
Instead of lengthy blurbs, switch to single-line, machine-readable
standardized (https://spdx.dev) license identifiers. The Linux kernel
switched long ago, so there is strong precedent.
Three cases are handled: AGPL-only, Apache-only, and dual licensed.
For the latter case, I chose (AGPL-3.0-or-later and Apache-2.0),
reasoning that our changes are extensive enough to apply our license.
The changes we applied mechanically with a script, except to
licenses/README.md.
Closes#9937
seastar::later() was recently deprecated and replaced with two
alternatives: a cheap seastar::yield() and an expensive (but more
powerful) seastar::check_for_io_immediately(), that corresponds to
the original later().
This patch replaces all later() calls with the weaker yield(). In
all cases except one, it's unambiguously correct. In one case
(test/perf scheduling_latency_measurer::stop()) it's not so ambiguous,
since check_for_io_immediately() will additionally force a poll and
so will cause more work to be done (but no additional tasks to be
executed). However, I think that any measurement that relies on
the measuring the work on the last tick to be inaccurate (you need
thousands of ticks to get any amount of confidence in the
measurement) that in the end it doesn't matter what we pick.
Tests: unit (dev)
Closes#9904
alloc_buf() calls new_buf_active() when there is no active segment to
allocate a new active segment. new_buf_active() allocates memory
(e.g. a new segment) so may cause memory reclamation, which may cause
segment compaction, which may call alloc_buf() and re-enter
new_buf_active(). The first call to new_buf_active() would then
override _buf_active and cause the segment allocated during segment
compaction to be leaked.
This then causes abort when objects from the leaked segment are freed
because the segment is expected to be present in _closed_segments, but
isn't. boost::intrusive::list::erase() will fail on assertion that the
object being erased is linked.
Introduced in b5ca0eb2a2.
Fixes#9821Fixes#9192Fixes#9825Fixes#9544Fixes#9508
Refs #9573
Message-Id: <20211229201443.119812-1-tgrabiec@scylladb.com>
The stall report uses the millisecond unit, but actually reports
nanoseconds.
Switch to microseconds (milliseconds are a bit too coarse) and
use the safer "duration / 1us" style rather than "duration::count()"
that leads to unit confusion.
Fixes#9733.
Closes#9734
We've been observing hard to explain crashes recently around
lsa_buffer destruction, where the containing segment is absent in
_segment_descs which causes log_heap::adjust_up to abort. Add more
checks to catch certain impossible senarios which can lead to this
sooner.
Refs #9192.
Message-Id: <20211116122346.814437-1-tgrabiec@scylladb.com>
We cannot recover from a failure in this method. The implementation
makes sure it never happens. Invariants will be broken if this
throws. Detect violations early by marking as noexcept.
We could make it exception safe and try to leave the data structures
in a consistent state but the reclaimer cannot make progress if this throws, so
it's pointless.
Refs #9192
Message-Id: <20211116122019.813418-1-tgrabiec@scylladb.com>
Reuse the existing `reclaim_timer` for stall detection.
* Since a timer is now set around every reclaim and compaction, use a
coarse one for speed.
* Set log level according to conditions (stalls deserve a warning).
* Add compaction/migration/eviction/allocation stats.
Refs #4186.
Signed-off-by: Michael Livshin <michael.livshin@scylladb.com>
Introduced in d72b91053b.
If region was not compactible, for example because it has dense
segments, we would keep evicting even though the target for reclaimed
segments was met. In the worst case we may have to evict whole cache.
Refs #9038 (unlikely to be the cause though)
Message-Id: <20210720104039.463662-1-tgrabiec@scylladb.com>
lsa_buffer allocations are aligned to 4K. If smaller size is
requested, whole 4K is used. However, only requested size was used in
accounting segment occupancy. This can confuse reclaimer which may
think the segment is sparse while it is actually dense, and compacting
it will yield no or little gain. This can cause inefficient memory
reclamation or lack of progress.
Refs #9038
Message-Id: <20210720104110.463812-1-tgrabiec@scylladb.com>
_free_space may be initialized with garbage so kind() getter should
only look at the bit which corresponds to the kind. Misclasification
of segment as being of different kind may result in a hang during
segment compaction.
Surfaced in debug mode build where the field is filled with 0xbebebebe.
Introduced in b5ca0eb2a2.
Fixes#9057
Message-Id: <20210719232734.443964-1-tgrabiec@scylladb.com>
"
The main goal of this series is to improve efficiency of reads from large partitions by
reducing amount of I/O needed to read the sstable index. This is achieved by caching
index file pages and partition index entries in memory.
Currently, the pages are cached by individual reads only for the duration of the read.
This was done to facilitate binary search in the promoted index (intra-partition index).
After this series, all reads share the index file page cache, which stays around even after reads stop.
The page cache is subject to eviction. It uses the same region as the current row cache and shares
the LRU with row cache entries. This means that LRU objects need to be virtualized. This series takes
an easy approach and does this by introducing a virtual base class. This adds an overhead to row cache
entry to store the vtable pointer.
SStable indexes have a hierarchy. There is a summary, which is a sparse partition key index into the
full partition index. This one is already kept in memory. The partition index is divided by the summary
into pages. Each entry in the partition index contains promoted index, which is a sparse index into atoms
identified by the clustering key (rows, tombstones).
In order to read the promoted index, the reader needs to read the partition index entry first.
To speed this up, this series also adds caching of partition index entries. This cache survives
reads and is subject to eviction, just like the index file page cache. The unit of caching is
the partition index page. Without this cache, each access to promoted index would have to be
preceded with the parsing of the partition index page containing the partition key.
Performance testing results follow.
1) scylla-bench large partition reads
Populated with:
perf_fast_forward --run-tests=large-partition-skips --datasets=sb-large-part-ds1 \
-c1 -m1G --populate --value-size=1024 --rows=10000000
Single partition, 9G data file, 4MB index file
Test execution:
build/release/scylla -c1 -m4G
scylla-bench -workload uniform -mode read -limit 1 -concurrency 100 -partition-count 1 \
-clustering-row-count 10000000 -duration 60m
TL;DR: after: 2x throughput, 0.5 median latency
Before (c1daf2bb24):
Results
Time (avg): 5m21.033180213s
Total ops: 966951
Total rows: 966951
Operations/s: 3011.997048812112
Rows/s: 3011.997048812112
Latency:
max: 74.055679ms
99.9th: 63.569919ms
99th: 41.320447ms
95th: 38.076415ms
90th: 37.158911ms
median: 34.537471ms
mean: 33.195994ms
After:
Results
Time (avg): 5m14.706669345s
Total ops: 2042831
Total rows: 2042831
Operations/s: 6491.22243800942
Rows/s: 6491.22243800942
Latency:
max: 60.096511ms
99.9th: 35.520511ms
99th: 27.000831ms
95th: 23.986175ms
90th: 21.659647ms
median: 15.040511ms
mean: 15.402076ms
2) scylla-bench small partitions
I tested several scenarios with a varying data set size, e.g. data fully fitting in memory,
half fitting, and being much larger. The improvement varied a bit but in all cases the "after"
code performed slightly better.
Below is a representative run over data set which does not fit in memory.
scylla -c1 -m4G
scylla-bench -workload uniform -mode read -concurrency 400 -partition-count 10000000 \
-clustering-row-count 1 -duration 60m -no-lower-bound
Before:
Time (avg): 51.072411913s
Total ops: 3165885
Total rows: 3165885
Operations/s: 61988.164024260645
Rows/s: 61988.164024260645
Latency:
max: 34.045951ms
99.9th: 25.985023ms
99th: 23.298047ms
95th: 19.070975ms
90th: 17.530879ms
median: 3.899391ms
mean: 6.450616ms
After:
Time (avg): 50.232410679s
Total ops: 3778863
Total rows: 3778863
Operations/s: 75227.58014424688
Rows/s: 75227.58014424688
Latency:
max: 37.027839ms
99.9th: 24.805375ms
99th: 18.219007ms
95th: 14.090239ms
90th: 12.124159ms
median: 4.030463ms
mean: 5.315111ms
The results include the warmup phase which populates the partition index cache, so the hot-cache effect
is dampened in the statistics. See the 99th percentile. Latency gets better after the cache warms up which
moves it lower.
3) perf_fast_forward --run-tests=large-partition-skips
Caching is not used here, included to show there are no regressions for the cold cache case.
TL;DR: No significant change
perf_fast_forward --run-tests=large-partition-skips --datasets=large-part-ds1 -c1 -m1G
Config: rows: 10000000, value size: 2000
Before:
read skip time (s) iterations frags frag/s mad f/s max f/s min f/s avg aio aio (KiB) blocked dropped idx hit idx miss idx blk c hit c miss c blk cpu
1 0 36.429822 4 10000000 274500 62 274521 274429 153889.2 153883 19696986 153853 0 0 0 0 0 0 0 22.5%
1 1 36.856236 4 5000000 135662 7 135670 135650 155652.0 155652 19704117 139326 1 0 1 1 0 0 0 38.1%
1 8 36.347667 4 1111112 30569 0 30570 30569 155652.0 155652 19704117 139071 1 0 1 1 0 0 0 19.5%
1 16 36.278866 4 588236 16214 1 16215 16213 155652.0 155652 19704117 139073 1 0 1 1 0 0 0 16.6%
1 32 36.174784 4 303031 8377 0 8377 8376 155652.0 155652 19704117 139056 1 0 1 1 0 0 0 12.3%
1 64 36.147104 4 153847 4256 0 4256 4256 155652.0 155652 19704117 139109 1 0 1 1 0 0 0 11.1%
1 256 9.895288 4 38911 3932 1 3933 3930 100869.2 100868 3178298 59944 38912 0 1 1 0 0 0 14.3%
1 1024 2.599921 4 9757 3753 0 3753 3753 26604.0 26604 801850 15071 9758 0 1 1 0 0 0 14.6%
1 4096 0.784568 4 2441 3111 1 3111 3109 7982.0 7982 205946 3772 2442 0 1 1 0 0 0 13.8%
64 1 36.553975 4 9846154 269359 10 269369 269337 155663.8 155652 19704117 139230 1 0 1 1 0 0 0 28.2%
64 8 36.509694 4 8888896 243467 8 243475 243449 155652.0 155652 19704117 139120 1 0 1 1 0 0 0 26.5%
64 16 36.466282 4 8000000 219381 4 219385 219374 155652.0 155652 19704117 139232 1 0 1 1 0 0 0 24.8%
64 32 36.395926 4 6666688 183171 6 183180 183165 155652.0 155652 19704117 139158 1 0 1 1 0 0 0 21.8%
64 64 36.296856 4 5000000 137753 4 137757 137737 155652.0 155652 19704117 139105 1 0 1 1 0 0 0 17.7%
64 256 20.590392 4 2000000 97133 18 97151 94996 135248.8 131395 7877402 98335 31282 0 1 1 0 0 0 15.7%
64 1024 6.225773 4 588288 94492 1436 95434 88748 46066.5 41321 2324378 30360 9193 0 1 1 0 0 0 15.8%
64 4096 1.856069 4 153856 82893 54 82948 82721 16115.0 16043 583674 11574 2675 0 1 1 0 0 0 16.3%
After:
read skip time (s) iterations frags frag/s mad f/s max f/s min f/s avg aio aio (KiB) blocked dropped idx hit idx miss idx blk c hit c miss c blk cpu
1 0 36.429240 4 10000000 274505 38 274515 274417 153887.8 153883 19696986 153849 0 0 0 0 0 0 0 22.4%
1 1 36.933806 4 5000000 135377 15 135385 135354 155658.0 155658 19704085 139398 1 0 1 1 0 0 0 40.0%
1 8 36.419187 4 1111112 30509 2 30510 30507 155658.0 155658 19704085 139233 1 0 1 1 0 0 0 22.0%
1 16 36.353475 4 588236 16181 0 16182 16181 155658.0 155658 19704085 139183 1 0 1 1 0 0 0 19.2%
1 32 36.251356 4 303031 8359 0 8359 8359 155658.0 155658 19704085 139120 1 0 1 1 0 0 0 14.8%
1 64 36.203692 4 153847 4249 0 4250 4249 155658.0 155658 19704085 139071 1 0 1 1 0 0 0 13.0%
1 256 9.965876 4 38911 3904 0 3906 3904 100875.2 100874 3178266 60108 38912 0 1 1 0 0 0 17.9%
1 1024 2.637501 4 9757 3699 1 3700 3697 26610.0 26610 801818 15071 9758 0 1 1 0 0 0 19.5%
1 4096 0.806745 4 2441 3026 1 3027 3024 7988.0 7988 205914 3773 2442 0 1 1 0 0 0 18.3%
64 1 36.611243 4 9846154 268938 5 268942 268921 155669.8 155705 19704085 139330 2 0 1 1 0 0 0 29.9%
64 8 36.559471 4 8888896 243135 11 243156 243124 155658.0 155658 19704085 139261 1 0 1 1 0 0 0 28.1%
64 16 36.510319 4 8000000 219116 15 219126 219101 155658.0 155658 19704085 139173 1 0 1 1 0 0 0 26.3%
64 32 36.439069 4 6666688 182954 9 182964 182943 155658.0 155658 19704085 139274 1 0 1 1 0 0 0 23.2%
64 64 36.334808 4 5000000 137609 11 137612 137596 155658.0 155658 19704085 139258 2 0 1 1 0 0 0 19.1%
64 256 20.624759 4 2000000 96971 88 97059 92717 138296.0 131401 7877370 98332 31282 0 1 1 0 0 0 17.2%
64 1024 6.260598 4 588288 93967 1429 94905 88051 45939.5 41327 2324346 30361 9193 0 1 1 0 0 0 17.8%
64 4096 1.881338 4 153856 81780 140 81920 81520 16109.8 16092 582714 11617 2678 0 1 1 0 0 0 18.2%
4) perf_fast_forward --run-tests=large-partition-slicing
Caching enabled, each line shows the median run from many iterations
TL;DR: We can observe reduction in IO which translates to reduction in execution time,
especially for slicing in the middle of partition.
perf_fast_forward --run-tests=large-partition-slicing --datasets=large-part-ds1 -c1 -m1G --keep-cache-across-test-cases
Config: rows: 10000000, value size: 2000
Before:
offset read time (s) iterations frags frag/s mad f/s max f/s min f/s avg aio aio (KiB) blocked dropped idx hit idx miss idx blk c hit c miss c blk allocs tasks insns/f cpu
0 1 0.000491 127 1 2037 24 2109 127 4.0 4 128 2 2 0 1 1 0 0 0 157 80 3058208 15.0%
0 32 0.000561 1740 32 56995 410 60031 47208 5.0 5 160 3 2 0 1 1 0 0 0 386 111 113353 17.5%
0 256 0.002052 488 256 124736 7111 144762 89053 16.6 17 672 14 2 0 1 1 0 0 0 2113 446 52669 18.6%
0 4096 0.016437 61 4096 249199 692 252389 244995 69.4 69 8640 57 5 0 1 1 0 0 0 26638 1717 23321 22.4%
5000000 1 0.002171 221 1 461 2 466 221 25.0 25 268 3 3 0 1 1 0 0 0 638 376 14311524 10.2%
5000000 32 0.002392 404 32 13376 48 13528 13015 27.0 27 332 5 3 0 1 1 0 0 0 931 432 489691 11.9%
5000000 256 0.003659 279 256 69967 764 73130 52563 39.5 41 780 19 3 0 1 1 0 0 0 2689 825 93756 15.8%
5000000 4096 0.018592 55 4096 220313 433 234214 218803 94.2 94 9484 62 9 0 1 1 0 0 0 27349 2213 26562 21.0%
After:
offset read time (s) iterations frags frag/s mad f/s max f/s min f/s avg aio aio (KiB) blocked dropped idx hit idx miss idx blk c hit c miss c blk allocs tasks insns/f cpu
0 1 0.000229 115 1 4371 85 4585 115 2.1 2 64 1 1 1 0 0 0 0 0 90 31 1314749 22.2%
0 32 0.000277 2174 32 115674 1015 128109 14144 3.0 3 96 2 1 1 0 0 0 0 0 319 62 52508 26.1%
0 256 0.001786 576 256 143298 5534 179142 113715 14.7 17 544 15 1 1 0 0 0 0 0 2110 453 45419 21.4%
0 4096 0.015498 61 4096 264289 2006 268850 259342 67.4 67 8576 59 4 1 0 0 0 0 0 26657 1738 22897 23.7%
5000000 1 0.000415 233 1 2411 15 2456 234 4.1 4 128 2 2 1 0 0 0 0 0 199 72 2644719 16.8%
5000000 32 0.000635 1413 32 50398 349 51149 46439 6.0 6 192 4 2 1 0 0 0 0 0 458 128 125893 18.6%
5000000 256 0.002028 486 256 126228 3024 146327 82559 17.8 18 1024 13 4 1 0 0 0 0 0 2123 385 51787 19.6%
5000000 4096 0.016836 61 4096 243294 814 263434 241660 73.0 73 9344 62 8 1 0 0 0 0 0 26922 1920 24389 22.4%
Future work:
- Check the impact on non-uniform workloads. Caching sstable indexes takes space away from the row cache
which may reduce the hit ratio.
- Reduce memory footprint of partition index cache. Currently, about 8x bloat over the on-disk size.
- Disable cache population for "bypass cache" reads
- Add a switch to disable sstable index caching, per-node, maybe per-table
- Better sstable index format. Current format leads to inefficiency in caching since only some elements of the cached
page can be hot. A B-tree index would be more efficient. Same applies to the partition index. Only some elements in
the partition index page can be hot.
- Add heuristic for reducing index file IO size when large partitions are anticipated. If we're bound by disk's
bandwidth it's wasteful to read the front of promoted index using 32K IO, better use 4K which should cover the
partition entry and then let binary search read the rest.
In V2:
- Fixed perf_fast_forward regression in the number of IOs used to read partition index page
The reader uses 32K reads, which were split by page cache into 4K reads
Fix by propagating IO size hints to page cache and using single IO to populate it.
New patch: "cached_file: Issue single I/O for the whole read range on miss"
- Avoid large allocations to store partition index page entries (due to managed_vector storage).
There is a unit test which detects this and fails.
Fixed by implementing chunked_managed_vector, based on chunked_vector.
- fixed bug in cached_file::evict_gently() where the wrong allocation strategy was used to free btree chunks
- Simplify region_impl::free_buf() according to Avi's suggestions
- Fit segment_kind in segment_descriptor::_free_space and lift requirement that _buf_pointers emptiness determines the kind
- Workaround sigsegv which was most likely due to coroutine miscompilation. Worked around by manipulating local object scope.
- Wire up system/drop_sstable_caches RESTful API
- Fix use-after-move on permit for the old scanning ka/la index reader
- Fixed more cases of double open_data() in tests leading to assert failure
- Adjusted cached_file class doc to account for changes in behavior.
- Rebased
Fixes#7079.
Refs #363.
"
* tag 'sstable-index-caching-v2' of github.com:tgrabiec/scylla: (39 commits)
api: Drop sstable index caches on system/drop_sstable_caches
cached_file: Issue single I/O for the whole read range on miss
row_cache: cache_tracker: Do not register metrics when constructed for tests
sstables, cached_file: Evict cache gently when sstable is destroyed
sstables: Hide partition_index_cache implementation away from sstables.hh
sstables: Drop shared_index_lists alias
sstables: Destroy partition index cache gently
sstables: Cache partition index pages in LSA and link to LRU
utils: Introduce lsa::weak_ptr<>
sstables: Rename index_list to partition_index_page and shared_index_lists to partition_index_cache
sstables, cached_file: Avoid copying buffers from cache when parsing promoted index
cached_file: Introduce get_page_units()
sstables: read: Document that primitive_consumer::read_32() is alloc-free
sstables: read: Count partition index page evictions
sstables: Drop the _use_binary_search flag from index entries
sstables: index_reader: Keep index objects under LSA
lsa: chunked_managed_vector: Adapt more to managed_vector
utils: lsa: chunked_managed_vector: Make LSA-aware
test: chunked_managed_vector_test: Make exception_safe_class standard layout
lsa: Copy chunked_vector to chunked_managed_vector
...
