You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@kudu.apache.org by al...@apache.org on 2020/09/09 15:20:33 UTC
[kudu] branch branch-1.13.x updated (f052657 -> 265c41c)
This is an automated email from the ASF dual-hosted git repository.
alexey pushed a change to branch branch-1.13.x
in repository https://gitbox.apache.org/repos/asf/kudu.git.
from f052657 Bump version to 1.13.0
new ea4b91e [threadpool] optimize queue overload detection
new 265c41c [docs] remove duplicate entry from 1.13 release notes
The 2 revisions listed above as "new" are entirely new to this
repository and will be described in separate emails. The revisions
listed as "add" were already present in the repository and have only
been added to this reference.
Summary of changes:
docs/release_notes.adoc | 2 --
src/kudu/util/threadpool.cc | 28 ++++++++++++++++------------
src/kudu/util/threadpool.h | 11 ++++++++---
3 files changed, 24 insertions(+), 17 deletions(-)
[kudu] 01/02: [threadpool] optimize queue overload detection
Posted by al...@apache.org.
This is an automated email from the ASF dual-hosted git repository.
alexey pushed a commit to branch branch-1.13.x
in repository https://gitbox.apache.org/repos/asf/kudu.git
commit ea4b91e38298a9cb75f9919adc62deebfbc9b788
Author: Alexey Serbin <al...@apache.org>
AuthorDate: Sun Sep 6 17:15:41 2020 -0700
[threadpool] optimize queue overload detection
This patch simplifies the queue overload detection and makes it a bit
more robust, according to 'perf stat' running
KUDU_ALLOW_SLOW_TESTS=1 perf stat ./bin/threadpool-test \
--gtest_filter='*ThreadPoolPerformanceTest.ConcurrentAndSerialTasksMix/1'
Before:
147699.960062 task-clock # 44.348 CPUs utilized
371,519 context-switches # 0.003 M/sec
653 cpu-migrations # 0.004 K/sec
3,664 page-faults # 0.025 K/sec
423,794,029,838 cycles # 2.869 GHz
94,656,186,092 instructions # 0.22 insns per cycle
34,018,899,188 branches # 230.324 M/sec
22,374,862 branch-misses # 0.07% of all branches
3.330492839 seconds time elapsed
After:
126357.374737 task-clock # 40.768 CPUs utilized
350,907 context-switches # 0.003 M/sec
3,167 cpu-migrations # 0.025 K/sec
3,478 page-faults # 0.028 K/sec
362,348,476,629 cycles # 2.868 GHz
82,964,368,394 instructions # 0.23 insns per cycle
29,553,336,891 branches # 233.887 M/sec
16,945,558 branch-misses # 0.06% of all branches
3.099419461 seconds time elapsed
Change-Id: Ic01ca7419beba92d7067c60ef520811136d67587
Reviewed-on: http://gerrit.cloudera.org:8080/16424
Tested-by: Kudu Jenkins
Reviewed-by: Grant Henke <gr...@apache.org>
Reviewed-by: Andrew Wong <aw...@cloudera.com>
(cherry picked from commit c4997af4b90b4a102597775e575320c95eb3c1ba)
Reviewed-on: http://gerrit.cloudera.org:8080/16429
Reviewed-by: Attila Bukor <ab...@apache.org>
---
src/kudu/util/threadpool.cc | 28 ++++++++++++++++------------
src/kudu/util/threadpool.h | 11 ++++++++---
2 files changed, 24 insertions(+), 15 deletions(-)
diff --git a/src/kudu/util/threadpool.cc b/src/kudu/util/threadpool.cc
index 6e2097d..569c968 100644
--- a/src/kudu/util/threadpool.cc
+++ b/src/kudu/util/threadpool.cc
@@ -771,6 +771,7 @@ ThreadPool::QueueLoadMeter::QueueLoadMeter(
: tpool_(tpool),
queue_time_threshold_(queue_time_threshold),
queue_time_history_length_(queue_time_history_length),
+ over_queue_threshold_num_(0),
queue_head_submit_time_(MonoTime()),
overloaded_since_(MonoTime()),
has_spare_thread_(true) {
@@ -826,18 +827,23 @@ void ThreadPool::QueueLoadMeter::UpdateQueueInfoUnlocked(
const MonoTime& queue_head_submit_time,
bool has_spare_thread) {
tpool_.lock_.AssertAcquired();
+
if (task_queue_time.Initialized()) {
- // TODO(aserbin): any better way of tracking the running minimum of N numbers?
queue_times_.emplace_back(task_queue_time);
- queue_times_ordered_.insert(task_queue_time);
+ // Given the criterion to detect whether the queue is overloaded, it's not
+ // exactly necessary to track the running minimum of queue times for the
+ // specified history window. It's enough just to keep an eye on whether the
+ // window contains at least one element that's not over the threshold.
+ if (task_queue_time > queue_time_threshold_) {
+ ++over_queue_threshold_num_;
+ }
if (queue_times_.size() > queue_time_history_length_) {
const auto& elem = queue_times_.front();
- auto it = queue_times_ordered_.find(elem);
- DCHECK(it != queue_times_ordered_.end());
- queue_times_ordered_.erase(it);
+ if (elem > queue_time_threshold_) {
+ --over_queue_threshold_num_;
+ }
queue_times_.pop_front();
}
- min_queue_time_ = *queue_times_.begin();
}
queue_head_submit_time_.store(queue_head_submit_time);
has_spare_thread_.store(has_spare_thread);
@@ -848,11 +854,10 @@ void ThreadPool::QueueLoadMeter::UpdateQueueInfoUnlocked(
const bool queue_empty = !queue_head_submit_time.Initialized();
const auto queue_time = queue_empty
? MonoDelta::FromSeconds(0) : now - queue_head_submit_time;
- const auto min_queue_time = min_queue_time_.Initialized()
- ? min_queue_time_ : MonoDelta::FromSeconds(0);
const bool was_overloaded = overloaded_since_.load().Initialized();
const bool overloaded = !has_spare_thread &&
- std::max(min_queue_time, queue_time) > queue_time_threshold_;
+ (over_queue_threshold_num_ == queue_time_history_length_ ||
+ queue_time > queue_time_threshold_);
// Track the state transitions and update overloaded_since_.
if (!was_overloaded && overloaded) {
VLOG(3) << Substitute("state transition: normal --> overloaded");
@@ -863,13 +868,12 @@ void ThreadPool::QueueLoadMeter::UpdateQueueInfoUnlocked(
}
VLOG(4) << Substitute("state refreshed: overloaded since $0, queue $1, "
"has $2 thread, queue head submit time $3, "
- "queue time $4, min queue time $5",
+ "queue time $4",
overloaded_since_.load().ToString(),
queue_empty ? "empty" : "not empty",
has_spare_thread ? "spare" : "no spare",
queue_head_submit_time.ToString(),
- queue_time.ToString(),
- min_queue_time.ToString());
+ queue_time.ToString());
}
std::ostream& operator<<(std::ostream& o, ThreadPoolToken::State s) {
diff --git a/src/kudu/util/threadpool.h b/src/kudu/util/threadpool.h
index 403acbd..ff1fae8 100644
--- a/src/kudu/util/threadpool.h
+++ b/src/kudu/util/threadpool.h
@@ -16,13 +16,14 @@
// under the License.
#pragma once
+#include <sys/types.h>
+
#include <atomic>
#include <cstddef>
#include <deque>
#include <functional>
#include <iosfwd>
#include <memory>
-#include <set>
#include <string>
#include <unordered_set>
@@ -340,11 +341,15 @@ class ThreadPool {
// to the parameter 'N' in the description of the algorithm above.
const size_t queue_time_history_length_;
+ // Number of elements in the queue history measurement window which are
+ // over the threshold specified by 'queue_time_threshold_'. Using the
+ // terminology from above, (min(QT_historic(M) > T_overload) iff
+ // (over_queue_threshold_num_ == M).
+ ssize_t over_queue_threshold_num_;
+
// Queue timings of the most recent samples. The size of these containers
// is kept under queue_time_history_length_ limit.
std::deque<MonoDelta> queue_times_;
- std::multiset<MonoDelta> queue_times_ordered_;
- MonoDelta min_queue_time_;
// Below fields are to store the latest snapshot of the information about
// the task queue of the pool the meter is attached to.
[kudu] 02/02: [docs] remove duplicate entry from 1.13 release notes
Posted by al...@apache.org.
This is an automated email from the ASF dual-hosted git repository.
alexey pushed a commit to branch branch-1.13.x
in repository https://gitbox.apache.org/repos/asf/kudu.git
commit 265c41c87c82103a4646e3bbde11bcc75275a47b
Author: Alexey Serbin <al...@apache.org>
AuthorDate: Tue Sep 8 14:13:34 2020 -0700
[docs] remove duplicate entry from 1.13 release notes
Removed a note that was duplicate to the following one:
* The Spark KuduContext accumulator metrics now track operation
counts per table instead of cumulatively for all tables.
Change-Id: I186f7160d71e94ed54c0e583655e67e2c0486ae9
Reviewed-on: http://gerrit.cloudera.org:8080/16430
Tested-by: Kudu Jenkins
Reviewed-by: Attila Bukor <ab...@apache.org>
---
docs/release_notes.adoc | 2 --
1 file changed, 2 deletions(-)
diff --git a/docs/release_notes.adoc b/docs/release_notes.adoc
index 044d389..cfd8223 100644
--- a/docs/release_notes.adoc
+++ b/docs/release_notes.adoc
@@ -95,8 +95,6 @@
the `--tablet_apply_pool_overload_threshold_ms` Tablet Server’s flag to
appropriate value, e.g. 250 (see
link:https://issues.apache.org/jira/browse/KUDU-1587[KUDU-1587]).
-* Operation accumulators in Spark KuduContext are now tracked on a per-table
- basis.
* Java client’s error collector can be resized (see
link:https://issues.apache.org/jira/browse/KUDU-1422[KUDU-1422]).
* Calls to the Kudu master server are now drastically reduced when using scan