wwbmmm commented on issue #2000:
URL:
https://github.com/apache/incubator-brpc/issues/2000#issuecomment-1325979641
>
不不,我的重点不是当前bthread挂起并signal其它的worker来继续执行这个bthread带来的性能损耗,我的重点是经过bthread的ProcessEvent并没有切换pthread,为何系统调用函数write和read的开销高于直接调用ProcessEvent时的系统调用函数write和read的开销。
有关系的。当
zhouyanghere commented on issue #2000:
URL:
https://github.com/apache/incubator-brpc/issues/2000#issuecomment-1325952578
不不,我的重点不是当前bthread挂起并signal其它的worker来继续执行这个bthread带来的性能损耗,我的重点是经过bthread的ProcessEvent并没有切换pthread,为何系统调用函数write和read的开销高于直接调用ProcessEvent时的系统调用函数write和read的开销。
--
This
wwbmmm commented on issue #2000:
URL:
https://github.com/apache/incubator-brpc/issues/2000#issuecomment-1325927072
bthread_start_urgent会把当前pthread的栈切到ProcessEvent函数,同时也会将当前bthread挂起并signal其它的worker来继续执行这个bthread,所以会对性能带来一定影响。并不是jumpstack带来的影响。
--
This is an automated message from the Apa
zhouyanghere commented on issue #2000:
URL:
https://github.com/apache/incubator-brpc/issues/2000#issuecomment-1325919196
如果将进程用taskset绑定到固定的CPU上,也没有此差异。是不是brpc
jumpstack的过程导致线程pthread切换了CPU,导致了cache miss,从而影响系统调用的开销。
--
This is an automated message from the Apache Git Service.
To res
zhouyanghere commented on issue #2000:
URL:
https://github.com/apache/incubator-brpc/issues/2000#issuecomment-1325914657
Socket::StartInputEvent 里如果直接调用ProcessEvent函数,也是函数调用过程和栈切换,为何brpc
jumpstack就能影响系统调用。是否和jumpstack的处理有关系?jumpstack是汇编代码,能力有限,看不大懂。
--
This is an automated message from t
zhouyanghere commented on issue #2000:
URL:
https://github.com/apache/incubator-brpc/issues/2000#issuecomment-1325909263
可能我描述的不是很清楚,和多线程应该没有直接关系。其他的类redis数据库,比如DragonFly DB,延时也很低。
我对ProcessEvent加了bvar统计,如果ProcessEvent走bthread,则ProcessEvent函数的处理平均耗时为26us。如果不走,则ProcessEvent只有6us。单链接下最明显。
wwbmmm commented on issue #2013:
URL:
https://github.com/apache/incubator-brpc/issues/2013#issuecomment-1325908659
我觉得可以
--
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
Yes, the RDMA support should be included.
And I'd like the following feature to be included in 1.4.0:
- Early return when size == 0 :
https://github.com/apache/incubator-brpc/pull/2009
- allow IOBuf::append_user_data_with_meta to specify the lkey of rdma
memory region: https://github.com/apache/in
yanglimingcn opened a new issue, #2013:
URL: https://github.com/apache/incubator-brpc/issues/2013
brpc有const concurrenty limiter 和 auto concurrency limiter,用于限制服务的并发度,const
concurrency limiter需要实际对服务压测配置,运营起来比较繁琐。auto concurrency limiter比较灵活,基于little’s
low法则,但是在使用过程中发现会经常报ELIMIT。
在服务正常运营
372046933 commented on code in PR #2009:
URL: https://github.com/apache/incubator-brpc/pull/2009#discussion_r1031004002
##
src/butil/iobuf.cpp:
##
@@ -1211,6 +1211,9 @@ int IOBuf::append_user_data(void* data, size_t size, void
(*deleter)(void*)) {
LOG(FATAL) << "data_s
wwbmmm commented on code in PR #2009:
URL: https://github.com/apache/incubator-brpc/pull/2009#discussion_r1031000482
##
src/butil/iobuf.cpp:
##
@@ -1211,6 +1211,9 @@ int IOBuf::append_user_data(void* data, size_t size, void
(*deleter)(void*)) {
LOG(FATAL) << "data_size
372046933 commented on code in PR #2009:
URL: https://github.com/apache/incubator-brpc/pull/2009#discussion_r1030998420
##
src/butil/iobuf.cpp:
##
@@ -1211,6 +1211,9 @@ int IOBuf::append_user_data(void* data, size_t size, void
(*deleter)(void*)) {
LOG(FATAL) << "data_s
Tuvie commented on issue #1995:
URL:
https://github.com/apache/incubator-brpc/issues/1995#issuecomment-1325850024
@372046933 Thanks for your feedback. About the return value problem, maybe
@wwbmmm can share your opinion.
--
This is an automated message from the Apache Git Service.
To res
372046933 commented on issue #1995:
URL:
https://github.com/apache/incubator-brpc/issues/1995#issuecomment-1325848850
@Tuvie I have tested `append_user_data_with_meta`. It works as expected. As
for the return value of `append_user_data` when size is 0. I added a pre-check
in application an
This is an automated email from the ASF dual-hosted git repository.
jiashunzhu pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-brpc.git
The following commit(s) were added to refs/heads/master by this push:
new d8b21d7c Linux workflow (#2008)
d8b2
zyearn merged PR #2008:
URL: https://github.com/apache/incubator-brpc/pull/2008
--
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-unsubscr...@brpc.apa
thanks, zhaogeng
On 2022/11/22 14:02:03 Zhaogeng Li wrote:
> We are still working on the feature allowing user managed memory.
> To make it easy to use, we keep discussion with our users.
> Hope we can add this feature in the next release.
>
>
>
>
> Thanks,
> Zhaogeng
>
> At 2022-11-21 22:25:
xiaohei520321 opened a new issue, #2012:
URL: https://github.com/apache/incubator-brpc/issues/2012
**Is your feature request related to a problem? (你需要的功能是否与某个问题有关?)**
最近在用brpc server,使用brpc_metrics将数据导入到prometheus上, 但发现没法获取平均latency, 看了代码
自定义的LatencyRecord在brpc_metrics下,能正常导出平均late
18 matches
Mail list logo