HDFS-17321. RBF: Add RouterAutoMsyncService for auto msync in Router

[LiuGuH]

Description of PR

Router should have the ability to to auto msync to a nameservice. And it can ensure router periodically refreshes its record of a namespace's state.

Different from HDFS-17027, this is controled by router itself without configuring with AbstractNNFailoverProxyProvider.
And HDFS-16890 maybe lead to many read requests into active NN at the same time.

This PR provides a new way to implememts auto msync in Router.

[hadoop-yetus]

🎊 +1 overall

Vote	Subsystem	Runtime	Logfile	Comment
+0 🆗	reexec	0m 20s		Docker mode activated.
			_ Prechecks _
+1 💚	dupname	0m 0s		No case conflicting files found.
+0 🆗	codespell	0m 0s		codespell was not available.
+0 🆗	detsecrets	0m 0s		detect-secrets was not available.
+0 🆗	xmllint	0m 0s		xmllint was not available.
+1 💚	@author	0m 0s		The patch does not contain any @author tags.
+1 💚	test4tests	0m 0s		The patch appears to include 3 new or modified test files.
			_ trunk Compile Tests _
+1 💚	mvninstall	31m 53s		trunk passed
+1 💚	compile	0m 22s		trunk passed with JDK Ubuntu-11.0.21+9-post-Ubuntu-0ubuntu120.04
+1 💚	compile	0m 19s		trunk passed with JDK Private Build-1.8.0_392-8u392-ga-1~20.04-b08
+1 💚	checkstyle	0m 17s		trunk passed
+1 💚	mvnsite	0m 28s		trunk passed
+1 💚	javadoc	0m 27s		trunk passed with JDK Ubuntu-11.0.21+9-post-Ubuntu-0ubuntu120.04
+1 💚	javadoc	0m 22s		trunk passed with JDK Private Build-1.8.0_392-8u392-ga-1~20.04-b08
+1 💚	spotbugs	0m 49s		trunk passed
+1 💚	shadedclient	19m 28s		branch has no errors when building and testing our client artifacts.
			_ Patch Compile Tests _
+1 💚	mvninstall	0m 20s		the patch passed
+1 💚	compile	0m 18s		the patch passed with JDK Ubuntu-11.0.21+9-post-Ubuntu-0ubuntu120.04
+1 💚	javac	0m 18s		the patch passed
+1 💚	compile	0m 16s		the patch passed with JDK Private Build-1.8.0_392-8u392-ga-1~20.04-b08
+1 💚	javac	0m 16s		the patch passed
+1 💚	blanks	0m 0s		The patch has no blanks issues.
+1 💚	checkstyle	0m 11s		the patch passed
+1 💚	mvnsite	0m 21s		the patch passed
+1 💚	javadoc	0m 17s		the patch passed with JDK Ubuntu-11.0.21+9-post-Ubuntu-0ubuntu120.04
+1 💚	javadoc	0m 17s		the patch passed with JDK Private Build-1.8.0_392-8u392-ga-1~20.04-b08
+1 💚	spotbugs	0m 48s		the patch passed
+1 💚	shadedclient	19m 43s		patch has no errors when building and testing our client artifacts.
			_ Other Tests _
+1 💚	unit	19m 20s		hadoop-hdfs-rbf in the patch passed.
+1 💚	asflicense	0m 24s		The patch does not generate ASF License warnings.
		99m 48s

Subsystem	Report/Notes
Docker	ClientAPI=1.43 ServerAPI=1.43 base: https://ci-hadoop.apache.org/job/hadoop-multibranch/job/PR-6404/1/artifact/out/Dockerfile
GITHUB PR	#6404
Optional Tests	dupname asflicense compile javac javadoc mvninstall mvnsite unit shadedclient spotbugs checkstyle codespell detsecrets xmllint
uname	Linux 715c46e9e3f8 5.15.0-88-generic #98-Ubuntu SMP Mon Oct 2 15:18:56 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux
Build tool	maven
Personality	dev-support/bin/hadoop.sh
git revision	trunk / f8183f8
Default Java	Private Build-1.8.0_392-8u392-ga-1~20.04-b08
Multi-JDK versions	/usr/lib/jvm/java-11-openjdk-amd64:Ubuntu-11.0.21+9-post-Ubuntu-0ubuntu120.04 /usr/lib/jvm/java-8-openjdk-amd64:Private Build-1.8.0_392-8u392-ga-1~20.04-b08
Test Results	https://ci-hadoop.apache.org/job/hadoop-multibranch/job/PR-6404/1/testReport/
Max. process+thread count	2620 (vs. ulimit of 5500)
modules	C: hadoop-hdfs-project/hadoop-hdfs-rbf U: hadoop-hdfs-project/hadoop-hdfs-rbf
Console output	https://ci-hadoop.apache.org/job/hadoop-multibranch/job/PR-6404/1/console
versions	git=2.25.1 maven=3.6.3 spotbugs=4.2.2
Powered by	Apache Yetus 0.14.0 https://yetus.apache.org

This message was automatically generated.

[slfan1989]

@simbadzina Can you help review this PR? Thank you very much! I see that HDFS-16890 and HDFS-17027 were both completed by you.

[slfan1989]

In unit tests, we'd better not use sleep.

[LiuGuH]

Changed with GenericTestUtils . Thanks.

[simbadzina]

@LiuGuH can you expand more on the following

And HDFS-16890 maybe lead to many read requests into active NN at the same time.

I see how multiple concurrent calls can read the same false value from isNamespaceStateIdFresh(nsId) before the accumulator is updated, but the windows for these reads should be very small. I'm wondering if we can instead fix the existing mechanism such that only a single read is sent to the active, vs. adding a new mechanism.

Additionally, the periodic redirection of calls to the active only happens in the case when there are no calls going to the active already so having some reads be sent to the active should not overload it.

[LiuGuH]

I'm wondering if we can instead fix the existing mechanism such that only a single read is sent to the active, vs. adding a new mechanism.

Yes, it can. But for only a single read is send to the active, we should add synchronized. And this maybe have performance impact. Add a seperate RouterAutoMsyncService maybe a way to slove it.

Additionally, the periodic redirection of calls to the active only happens in the case when there are no calls going to the active already so having some reads be sent to the active should not overload it.

In most cases, that's true. But I think add RouterAutoMsyncService will be more robustness. Thanks

[simbadzina]

I'm wondering if we can instead fix the existing mechanism such that only a single read is sent to the active, vs. adding a new mechanism.

Yes, it can. But for only a single read is send to the active, we should add synchronized. And this maybe have performance impact. Add a seperate RouterAutoMsyncService maybe a way to slove it.

Additionally, the periodic redirection of calls to the active only happens in the case when there are no calls going to the active already so having some reads be sent to the active should not overload it.

In most cases, that's true. But I think add RouterAutoMsyncService will be more robustness. Thanks

That is fair. I'll take a closer look at the implementation now that we've discussed the higher level details. Please also address the unit test comment by @slfan1989

[hadoop-yetus]

🎊 +1 overall

Vote	Subsystem	Runtime	Logfile	Comment
+0 🆗	reexec	0m 21s		Docker mode activated.
			_ Prechecks _
+1 💚	dupname	0m 0s		No case conflicting files found.
+0 🆗	codespell	0m 0s		codespell was not available.
+0 🆗	detsecrets	0m 0s		detect-secrets was not available.
+0 🆗	xmllint	0m 0s		xmllint was not available.
+1 💚	@author	0m 0s		The patch does not contain any @author tags.
+1 💚	test4tests	0m 0s		The patch appears to include 3 new or modified test files.
			_ trunk Compile Tests _
+1 💚	mvninstall	40m 54s		trunk passed
+1 💚	compile	0m 22s		trunk passed with JDK Ubuntu-11.0.21+9-post-Ubuntu-0ubuntu120.04
+1 💚	compile	0m 20s		trunk passed with JDK Private Build-1.8.0_392-8u392-ga-1~20.04-b08
+1 💚	checkstyle	0m 19s		trunk passed
+1 💚	mvnsite	0m 27s		trunk passed
+1 💚	javadoc	0m 29s		trunk passed with JDK Ubuntu-11.0.21+9-post-Ubuntu-0ubuntu120.04
+1 💚	javadoc	0m 18s		trunk passed with JDK Private Build-1.8.0_392-8u392-ga-1~20.04-b08
+1 💚	spotbugs	0m 49s		trunk passed
+1 💚	shadedclient	19m 31s		branch has no errors when building and testing our client artifacts.
			_ Patch Compile Tests _
+1 💚	mvninstall	0m 18s		the patch passed
+1 💚	compile	0m 18s		the patch passed with JDK Ubuntu-11.0.21+9-post-Ubuntu-0ubuntu120.04
+1 💚	javac	0m 18s		the patch passed
+1 💚	compile	0m 18s		the patch passed with JDK Private Build-1.8.0_392-8u392-ga-1~20.04-b08
+1 💚	javac	0m 18s		the patch passed
+1 💚	blanks	0m 0s		The patch has no blanks issues.
+1 💚	checkstyle	0m 10s		the patch passed
+1 💚	mvnsite	0m 22s		the patch passed
+1 💚	javadoc	0m 19s		the patch passed with JDK Ubuntu-11.0.21+9-post-Ubuntu-0ubuntu120.04
+1 💚	javadoc	0m 17s		the patch passed with JDK Private Build-1.8.0_392-8u392-ga-1~20.04-b08
+1 💚	spotbugs	0m 52s		the patch passed
+1 💚	shadedclient	19m 45s		patch has no errors when building and testing our client artifacts.
			_ Other Tests _
+1 💚	unit	19m 24s		hadoop-hdfs-rbf in the patch passed.
+1 💚	asflicense	0m 24s		The patch does not generate ASF License warnings.
		108m 38s

Subsystem	Report/Notes
Docker	ClientAPI=1.43 ServerAPI=1.43 base: https://ci-hadoop.apache.org/job/hadoop-multibranch/job/PR-6404/2/artifact/out/Dockerfile
GITHUB PR	#6404
Optional Tests	dupname asflicense compile javac javadoc mvninstall mvnsite unit shadedclient spotbugs checkstyle codespell detsecrets xmllint
uname	Linux 8a23deefa7cb 5.15.0-88-generic #98-Ubuntu SMP Mon Oct 2 15:18:56 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux
Build tool	maven
Personality	dev-support/bin/hadoop.sh
git revision	trunk / efcc74f
Default Java	Private Build-1.8.0_392-8u392-ga-1~20.04-b08
Multi-JDK versions	/usr/lib/jvm/java-11-openjdk-amd64:Ubuntu-11.0.21+9-post-Ubuntu-0ubuntu120.04 /usr/lib/jvm/java-8-openjdk-amd64:Private Build-1.8.0_392-8u392-ga-1~20.04-b08
Test Results	https://ci-hadoop.apache.org/job/hadoop-multibranch/job/PR-6404/2/testReport/
Max. process+thread count	2631 (vs. ulimit of 5500)
modules	C: hadoop-hdfs-project/hadoop-hdfs-rbf U: hadoop-hdfs-project/hadoop-hdfs-rbf
Console output	https://ci-hadoop.apache.org/job/hadoop-multibranch/job/PR-6404/2/console
versions	git=2.25.1 maven=3.6.3 spotbugs=4.2.2
Powered by	Apache Yetus 0.14.0 https://yetus.apache.org

This message was automatically generated.

[slfan1989]

@simbadzina Thanks for helping review the code!

@LiuGuH Thanks for the contribution! but I think it's better if we don't introduce a new one if the original one works. Personally, I'm worried about the synchronization state of a single thread, it doesn't seem to be a good design. If some boundaries occur (such as NN restart, hang, etc.), how do we deal with it? If we are going to design a more complex mechanism to handle boundary issues, I think it is better to use the mechanisms that are already available.

[hadoop-yetus]

🎊 +1 overall

Vote	Subsystem	Runtime	Logfile	Comment
			_ Prechecks _
+1 💚	dupname	0m 01s		No case conflicting files found.
+0 🆗	spotbugs	0m 01s		spotbugs executables are not available.
+0 🆗	codespell	0m 01s		codespell was not available.
+0 🆗	detsecrets	0m 01s		detect-secrets was not available.
+0 🆗	xmllint	0m 01s		xmllint was not available.
+1 💚	@author	0m 00s		The patch does not contain any @author tags.
+1 💚	test4tests	0m 00s		The patch appears to include 3 new or modified test files.
			_ trunk Compile Tests _
+1 💚	mvninstall	89m 58s		trunk passed
+1 💚	compile	4m 55s		trunk passed
+1 💚	checkstyle	4m 31s		trunk passed
+1 💚	mvnsite	5m 00s		trunk passed
+1 💚	javadoc	4m 40s		trunk passed
+1 💚	shadedclient	146m 45s		branch has no errors when building and testing our client artifacts.
			_ Patch Compile Tests _
+1 💚	mvninstall	2m 53s		the patch passed
+1 💚	compile	2m 24s		the patch passed
+1 💚	javac	2m 24s		the patch passed
+1 💚	blanks	0m 00s		The patch has no blanks issues.
+1 💚	checkstyle	2m 02s		the patch passed
+1 💚	mvnsite	2m 26s		the patch passed
+1 💚	javadoc	2m 07s		the patch passed
+1 💚	shadedclient	176m 27s		patch has no errors when building and testing our client artifacts.
			_ Other Tests _
+1 💚	asflicense	8m 22s		The patch does not generate ASF License warnings.
		439m 42s

Subsystem	Report/Notes
GITHUB PR	#6404
Optional Tests	dupname asflicense compile javac javadoc mvninstall mvnsite unit shadedclient spotbugs checkstyle codespell detsecrets xmllint
uname	MINGW64_NT-10.0-17763 daf58be70672 3.4.10-87d57229.x86_64 2024-02-14 20:17 UTC x86_64 Msys
Build tool	maven
Personality	/c/hadoop/dev-support/bin/hadoop.sh
git revision	trunk / efcc74f
Default Java	Azul Systems, Inc.-1.8.0_332-b09
Test Results	https://ci-hadoop.apache.org/job/hadoop-multibranch-windows-10/job/PR-6404/1/testReport/
modules	C: hadoop-hdfs-project/hadoop-hdfs-rbf U: hadoop-hdfs-project/hadoop-hdfs-rbf
Console output	https://ci-hadoop.apache.org/job/hadoop-multibranch-windows-10/job/PR-6404/1/console
versions	git=2.44.0.windows.1
Powered by	Apache Yetus 0.14.0 https://yetus.apache.org

This message was automatically generated.

[github-actions]

We're closing this stale PR because it has been open for 100 days with no activity. This isn't a judgement on the merit of the PR in any way. It's just a way of keeping the PR queue manageable.
If you feel like this was a mistake, or you would like to continue working on it, please feel free to re-open it and ask for a committer to remove the stale tag and review again.
Thanks all for your contribution.