You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@rocketmq.apache.org by GitBox <gi...@apache.org> on 2022/08/26 09:21:12 UTC
[GitHub] [rocketmq] isysc1 opened a new pull request, #4903: fix-warmMappedFile
isysc1 opened a new pull request, #4903:
URL: https://github.com/apache/rocketmq/pull/4903
Place a safepoint before the loop, the processing effect takes precedence over sleep
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] lizhanhui commented on pull request #4903: fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
lizhanhui commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1229187159
The existing solution is to write one byte per page as warming. According to https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux_for_real_time/7/html/reference_guide/using_mlock_to_avoid_page_io, it looks sufficient to mlock;
Another thing, actual empty write is actually preferred for some reason, it would be better to iterate from file tail to head, as the head of the file region is more likely to be written with data in the near future.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] shy-share commented on pull request #4903: [ISSUE 4902]fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
shy-share commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1238844605
> @shy-share the article seems deprecated https://stackoverflow.com/questions/72753599/counted-uncounted-loops-and-safepoints-is-while-i-someint-considered-u loop with a long variable becomes COUNTED after JDK16.
@lizhanhui ok，so I think it is still necessary to modify this code. For jdk8, there does not seem to be a big performance gap after modification, but in order to be compatible with subsequent versions of jdk, it seems better to do so
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] shy-share commented on pull request #4903: [ISSUE 4902]fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
shy-share commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1242581308
@lizhanhui @Oliverwqcwrw
i write a benchmark code，refer to [this](https://juejin.cn/post/7141541647447228429)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] kimmking commented on pull request #4903: [ISSUE 4902]fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by "kimmking (via GitHub)" <gi...@apache.org>.
kimmking commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1455065388
for循环退不出的场景，也可以通过-XX:+UseCountedLoopSafepoints解决，不需要改代码。
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] Oliverwqcwrw commented on pull request #4903: fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
Oliverwqcwrw commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1233593164
Would it be better to place safepoint if we had GC problems in the current situation ?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] lizhanhui commented on pull request #4903: fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
lizhanhui commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1229990465
@shy-share The article you posted may talk about anonymous memory regions. For file-backed memory area, mlock should be the behavior discussed in the redhat doc.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] shy-share commented on pull request #4903: fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
shy-share commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1229413684
> The existing solution is to write one byte per page as warming. According to https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux_for_real_time/7/html/reference_guide/using_mlock_to_avoid_page_io, it looks sufficient to mlock;
I checked some information, and some places said that mlock alone cannot guarantee that memory can be locked, because there may be copy-on-write problems. I'm not sure if this is the problem.
[refer to this article](https://www.cnblogs.com/lixiaofei1987/p/3205002.html)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] lizhanhui commented on pull request #4903: fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
lizhanhui commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1233630387
@Oliverwqcwrw Why not perform an A-B test between the current solution and the proposed one, let data talk.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [ISSUE 4902]fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep [rocketmq]
Posted by "yinyinnie (via GitHub)" <gi...@apache.org>.
yinyinnie commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1749900863
还有一个疑问，我理解无论是 1. Thread.sleep 2. 还是把int 换成long, 都是为了手动在for循环体内放置一个安全点。 为什么手动放置一个安全点就达到优化性能的效果了呢？
原因难道是这样：
有安全点的话，虚拟机通过“短暂的停顿”，就能枚举需要被回收的对象，所有“短暂的停顿”所累积的时间记为t1
如果不放置安全点，可能会导致full GC, 将GC的时间记为t2
t1 远小于 t2?
如果原因真是这样，想请问 t1 和 t2之间的量级差多少？
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] Oliverwqcwrw commented on pull request #4903: fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
Oliverwqcwrw commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1233631786
> @Oliverwqcwrw Why not perform an A-B test between the current solution and the proposed one, let data talk.
agree with, The data makes it easier for us to choose specific solutions
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] lizhanhui merged pull request #4903: [ISSUE 4902]fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
lizhanhui merged PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] shy-share commented on pull request #4903: [ISSUE 4902]fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by "shy-share (via GitHub)" <gi...@apache.org>.
shy-share commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1461192275
> for循环退不出的场景，也可以通过-XX:+UseCountedLoopSafepoints解决，不需要改代码。
一般很少修改jdk参数，如果线上用
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] lizhanhui commented on pull request #4903: [ISSUE 4902]fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
lizhanhui commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1238828824
@shy-share Pretty good analysis. Well done.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] lizhanhui commented on pull request #4903: [ISSUE 4902]fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
lizhanhui commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1251772675
@Oliverwqcwrw Either Uncounted Loop or Sleep(0) is fine, IMO. Be sure to add proper documentation, explaining the purpose of `sleep(0)` or `long-uncounted loop`.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] kimmking commented on pull request #4903: [ISSUE 4902]fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by "kimmking (via GitHub)" <gi...@apache.org>.
kimmking commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1455064244
实际上，这里的改动无效：for循环里有方法调用，就不存在这个问题。方法调用栈退出的地方，是可以加check safepoint的。
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] Hccake commented on pull request #4903: fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
Hccake commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1229138861
> > 而 fileSize 又是 int 类型，现在把 i 改为 long 类型完全没有意义
>
> @Hccake filesize是不是int类型和jvm能不能在循环前放置安全点并没有什么关系，它看的就是循环的这个i是不是long类型的 参考这篇[文章](https://juejin.cn/post/6844903878765314061#comment)
受教，学习了
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] shy-share commented on pull request #4903: [ISSUE 4902]fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
shy-share commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1238806436
@lizhanhui @Oliverwqcwrw This [article](https://mp.weixin.qq.com/s/AhoTKmXfRW0RDC3rS5SZcw) has done a simple test, it seems that there is no big difference, and it is only limited to versions below jdk8, this problem has been optimized in jdk10
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] lizhanhui commented on pull request #4903: [ISSUE 4902]fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
lizhanhui commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1238833559
@shy-share the article seems deprecated https://stackoverflow.com/questions/72753599/counted-uncounted-loops-and-safepoints-is-while-i-someint-considered-u
loop with a long variable becomes COUNTED after JDK16.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [ISSUE 4902]fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep [rocketmq]
Posted by "yinyinnie (via GitHub)" <gi...@apache.org>.
yinyinnie commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1749900720
循环体内不是已经有方法调用了嘛： `byteBuffer.put((int) i, (byte) 0);` 难道这个方法调用不能当做“安全点”？
我猜测不能当安全点的原因是：JIT在编译byteBuffer.put((int) i, (byte) 0); 这段代码的时候，会进行内联操作？
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] Oliverwqcwrw commented on pull request #4903: fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
Oliverwqcwrw commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1233632181
> @Oliverwqcwrw Why not perform an A-B test between the current solution and the proposed one, let data talk.
agree with, The data makes it easier for us to choose specific solutions
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] Knowden commented on pull request #4903: fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
Knowden commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1231065887
why the comment of sleep call is 'prevent gc'? In fact it's helping gc, right?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] shy-share commented on pull request #4903: fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
shy-share commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1229131296
> 而 fileSize 又是 int 类型，现在把 i 改为 long 类型完全没有意义
@Hccake filesize是不是int类型和jvm能不能在循环前放置安全点并没有什么关系，它看的就是循环的这个i是不是long类型的
参考这篇[文章](https://juejin.cn/post/6844903878765314061#comment)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] ShadowySpirits commented on a diff in pull request #4903: [ISSUE 4902]fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
ShadowySpirits commented on code in PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#discussion_r967576347
##########
store/src/main/java/org/apache/rocketmq/store/logfile/DefaultMappedFile.java:
##########
@@ -548,15 +548,15 @@ public void warmMappedFile(FlushDiskType type, int pages) {
}
// prevent gc
- if (j % 1000 == 0) {
- log.info("j={}, costTime={}", j, System.currentTimeMillis() - time);
- time = System.currentTimeMillis();
- try {
- Thread.sleep(0);
- } catch (InterruptedException e) {
- log.error("Interrupted", e);
- }
- }
+ // if (j % 1000 == 0) {
+ // log.info("j={}, costTime={}", j, System.currentTimeMillis() - time);
+ // time = System.currentTimeMillis();
+ // try {
+ // Thread.sleep(0);
Review Comment:
IMO, sleep(0) will give CPU time to another thread. These codes should not be deleted to prevent high CPU usage of this warm-up loop.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] shy-share commented on pull request #4903: fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
shy-share commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1231322548
> why the comment of sleep call is 'prevent gc'? In fact it's helping gc, right?
@Knowden In fact, it can't prevent gc, it main purpose is to increase gc frequency and reduce single gc time.
refer to [why-thread-sleep0-can-prevent-gc-in-rocketmq](https://stackoverflow.com/questions/53284031/why-thread-sleep0-can-prevent-gc-in-rocketmq)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] codecov-commenter commented on pull request #4903: fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
codecov-commenter commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1229188508
# [Codecov](https://codecov.io/gh/apache/rocketmq/pull/4903?src=pr&el=h1&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation) Report
> Merging [#4903](https://codecov.io/gh/apache/rocketmq/pull/4903?src=pr&el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation) (8325414) into [develop](https://codecov.io/gh/apache/rocketmq/commit/c4117138367234d6cfbddcbac733ef1c34ab0aa3?el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation) (c411713) will **decrease** coverage by `0.00%`.
> The diff coverage is `0.00%`.
```diff
@@ Coverage Diff @@
## develop #4903 +/- ##
=============================================
- Coverage 43.29% 43.28% -0.01%
+ Complexity 7688 7684 -4
=============================================
Files 991 991
Lines 68784 68776 -8
Branches 9112 9111 -1
=============================================
- Hits 29779 29771 -8
+ Misses 35263 35257 -6
- Partials 3742 3748 +6
```
| [Impacted Files](https://codecov.io/gh/apache/rocketmq/pull/4903?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation) | Coverage Δ | |
|---|---|---|
| [...ache/rocketmq/store/logfile/DefaultMappedFile.java](https://codecov.io/gh/apache/rocketmq/pull/4903/diff?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-c3RvcmUvc3JjL21haW4vamF2YS9vcmcvYXBhY2hlL3JvY2tldG1xL3N0b3JlL2xvZ2ZpbGUvRGVmYXVsdE1hcHBlZEZpbGUuamF2YQ==) | `49.67% <0.00%> (+1.26%)` | :arrow_up: |
| [...a/org/apache/rocketmq/filter/util/BloomFilter.java](https://codecov.io/gh/apache/rocketmq/pull/4903/diff?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-ZmlsdGVyL3NyYy9tYWluL2phdmEvb3JnL2FwYWNoZS9yb2NrZXRtcS9maWx0ZXIvdXRpbC9CbG9vbUZpbHRlci5qYXZh) | `60.43% <0.00%> (-2.20%)` | :arrow_down: |
| [.../apache/rocketmq/controller/ControllerManager.java](https://codecov.io/gh/apache/rocketmq/pull/4903/diff?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-Y29udHJvbGxlci9zcmMvbWFpbi9qYXZhL29yZy9hcGFjaGUvcm9ja2V0bXEvY29udHJvbGxlci9Db250cm9sbGVyTWFuYWdlci5qYXZh) | `72.82% <0.00%> (-2.18%)` | :arrow_down: |
| [...mq/store/ha/autoswitch/AutoSwitchHAConnection.java](https://codecov.io/gh/apache/rocketmq/pull/4903/diff?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-c3RvcmUvc3JjL21haW4vamF2YS9vcmcvYXBhY2hlL3JvY2tldG1xL3N0b3JlL2hhL2F1dG9zd2l0Y2gvQXV0b1N3aXRjaEhBQ29ubmVjdGlvbi5qYXZh) | `73.29% <0.00%> (-1.64%)` | :arrow_down: |
| [...client/consumer/store/RemoteBrokerOffsetStore.java](https://codecov.io/gh/apache/rocketmq/pull/4903/diff?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-Y2xpZW50L3NyYy9tYWluL2phdmEvb3JnL2FwYWNoZS9yb2NrZXRtcS9jbGllbnQvY29uc3VtZXIvc3RvcmUvUmVtb3RlQnJva2VyT2Zmc2V0U3RvcmUuamF2YQ==) | `68.14% <0.00%> (-0.89%)` | :arrow_down: |
| [...apache/rocketmq/store/queue/BatchConsumeQueue.java](https://codecov.io/gh/apache/rocketmq/pull/4903/diff?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-c3RvcmUvc3JjL21haW4vamF2YS9vcmcvYXBhY2hlL3JvY2tldG1xL3N0b3JlL3F1ZXVlL0JhdGNoQ29uc3VtZVF1ZXVlLmphdmE=) | `69.50% <0.00%> (-0.63%)` | :arrow_down: |
| [...rocketmq/broker/processor/PopMessageProcessor.java](https://codecov.io/gh/apache/rocketmq/pull/4903/diff?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-YnJva2VyL3NyYy9tYWluL2phdmEvb3JnL2FwYWNoZS9yb2NrZXRtcS9icm9rZXIvcHJvY2Vzc29yL1BvcE1lc3NhZ2VQcm9jZXNzb3IuamF2YQ==) | `37.63% <0.00%> (-0.54%)` | :arrow_down: |
| [...he/rocketmq/client/impl/consumer/ProcessQueue.java](https://codecov.io/gh/apache/rocketmq/pull/4903/diff?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-Y2xpZW50L3NyYy9tYWluL2phdmEvb3JnL2FwYWNoZS9yb2NrZXRtcS9jbGllbnQvaW1wbC9jb25zdW1lci9Qcm9jZXNzUXVldWUuamF2YQ==) | `61.46% <0.00%> (-0.46%)` | :arrow_down: |
| [...e/rocketmq/remoting/netty/NettyRemotingClient.java](https://codecov.io/gh/apache/rocketmq/pull/4903/diff?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-cmVtb3Rpbmcvc3JjL21haW4vamF2YS9vcmcvYXBhY2hlL3JvY2tldG1xL3JlbW90aW5nL25ldHR5L05ldHR5UmVtb3RpbmdDbGllbnQuamF2YQ==) | `43.73% <0.00%> (-0.44%)` | :arrow_down: |
| [...ava/org/apache/rocketmq/filter/util/BitsArray.java](https://codecov.io/gh/apache/rocketmq/pull/4903/diff?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-ZmlsdGVyL3NyYy9tYWluL2phdmEvb3JnL2FwYWNoZS9yb2NrZXRtcS9maWx0ZXIvdXRpbC9CaXRzQXJyYXkuamF2YQ==) | `59.82% <0.00%> (ø)` | |
| ... and [8 more](https://codecov.io/gh/apache/rocketmq/pull/4903/diff?src=pr&el=tree-more&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation) | |
:mega: We’re building smart automated test selection to slash your CI/CD build times. [Learn more](https://about.codecov.io/iterative-testing/?utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] Oliverwqcwrw commented on pull request #4903: [ISSUE 4902]fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by GitBox <gi...@apache.org>.
Oliverwqcwrw commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1251778070
> @Oliverwqcwrw Either Uncounted Loop or Sleep(0) is fine, IMO. Be sure to add proper documentation, explaining the purpose of `sleep(0)` or `long-uncounted loop`.
agree with :)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [rocketmq] shy-share commented on pull request #4903: [ISSUE 4902]fix-warmMappedFile：Place a safepoint before the loop, the processing effect takes precedence over sleep
Posted by "shy-share (via GitHub)" <gi...@apache.org>.
shy-share commented on PR #4903:
URL: https://github.com/apache/rocketmq/pull/4903#issuecomment-1461191976
> 实际上，这里的改动无效：for循环里有方法调用，就不存在这个问题。方法调用栈退出的地方，是可以加check safepoint的。
方法可能会被JIT编译器优化，从而导致无法正确识别safepoint
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@rocketmq.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org