[NOTICE] The minimum supported java version will be 17 starting from 3.0.0

2024-06-18 Thread Duo Zhang
After several rounds of discussions[1][2], we finally decided to set the minimum supported java version for HBase 3.0.0+ to 17. Notice that, we will keep HBase 2.x to work with both java 8, java 11 and java 17, so it is suggested to upgrade to use java 17 first on HBase 2.x, and then upgrade to

[ANNOUNCE] New HBase committer Andor Molnár

2024-05-29 Thread Duo Zhang
On behalf of the Apache HBase PMC, I am pleased to announce that Andor Molnár(andor) has accepted the PMC's invitation to become a committer on the project. We appreciate all of Andor Molnár's generous contributions thus far and look forward to his continued involvement. Congratulations and

[ANNOUNCE] Apache HBase 2.4.18 is now available for download

2024-05-25 Thread Duo Zhang
The HBase team is happy to announce the immediate availability of HBase 2.4.18. Apache HBase™ is an open-source, distributed, versioned, non-relational database. Apache HBase gives you low latency random access to billions of rows with millions of columns atop non-specialized hardware. To learn

Re: [ANNOUNCE] Apache HBase 2.6.0 is now available for download

2024-05-20 Thread Duo Zhang
Congratulations! But it seems you missed the replacement for the first '_version_' placeholder... Bryan Beaudreault 于2024年5月21日周二 00:44写道: > > The HBase team is happy to announce the immediate availability of HBase > _version_. > > Apache HBase™ is an open-source, distributed, versioned,

HBase Quarterly report Jan-Mar 2024

2024-05-18 Thread Duo Zhang
Hi all, HBase submits a report to the ASF board once a quarter, to inform the board about project health. I'm sending the report to the user@ and dev@ mailing lists because you are the project, and for transparency. If you have any questions about the report or the running of the project, you can

[ANNOUNCE] New HBase committer Istvan Toth

2024-04-02 Thread Duo Zhang
On behalf of the Apache HBase PMC, I am pleased to announce that Istvan Toth(stoty) has accepted the PMC's invitation to become a committer on the project. We appreciate all of Istvan Toth's generous contributions thus far and look forward to his continued involvement. Congratulations and

HBase Quarterly report Oct-Dec 2023

2024-01-28 Thread Duo Zhang
Hi all, HBase submits a report to the ASF board once a quarter, to inform the board about project health. I'm sending the report to the user@ and dev@ mailing lists because you are the project, and for transparency. If you have any questions about the report or the running of the project, you can

[ANNOUNCE] Apache HBase 3.0.0-beta-1 is now available for download

2024-01-19 Thread Duo Zhang
The HBase team is happy to announce the immediate availability of HBase 3.0.0-beta-1. Apache HBase™ is an open-source, distributed, versioned, non-relational database. Apache HBase gives you low latency random access to billions of rows with millions of columns atop non-specialized hardware. To

Re: Stripe Compaction 策略遇到的一些问题请教

2023-11-28 Thread Duo Zhang
果然 https://github.com/apache/hbase/blob/4d90b918a3702b4e4ae2f9ee890c14665e821c01/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/StripeStoreEngine.java#L84 这个方法里压根就没用过那个 forceMajor 参数 可以开个 issue 吧,看看这块怎么改一下,至少得把 forceMajor 这个参数传递下去,在 stripe 内部做 compact 的时候需要能拿到这个参数 张铎(Duo Zhang

Re: Stripe Compaction 策略遇到的一些问题请教

2023-11-28 Thread Duo Zhang
我印象里应该是有这个机制的,是不是 StripeCompaction 里没考虑这个参数 leojie 于2023年11月21日周二 11:42写道: > > 张老师, > 您好! > 请教一个HBase的问题,我们线上有张表,应用了Stripe > Compaction策略,每个region均值40G,被分成8个Stripe,每个Stripe5g,业务删除大量数据,表整体和单个region手动触发major > compaction不起作用,挑选不出来合适的文件参与合并。 > >

[ANNOUNCE] Please welcome Bryan Beaudreault to the HBase PMC

2023-10-16 Thread Duo Zhang
On behalf of the Apache HBase PMC I am pleased to announce that Bryan Beaudreault has accepted our invitation to become a PMC member on the Apache HBase project. We appreciate Bryan Beaudreault stepping up to take more responsibility in the HBase project. Please join me in welcoming Bryan

[ANNOUNCE] New HBase committer Hui Ruan(阮辉)

2023-09-15 Thread Duo Zhang
On behalf of the Apache HBase PMC, I am pleased to announce that Hui Ruan(frostruan) has accepted the PMC's invitation to become a committer on the project. We appreciate all of Hui's generous contributions thus far and look forward to his continued involvement. Congratulations and welcome, Hui

Re: [DISCUSS] Reschedule July meetup?

2023-07-25 Thread Duo Zhang
We'd better discuss the topics we want to discuss in the online meeting on the mailing list first, and post them out when scheduling the meeting. In this way I think we can attract more attendees. And for me, I prefer a night session, as I have to take care of my children in the morning...

[ANNOUNCE] Apache HBase 3.0.0-alpha-4 is now available for download

2023-06-08 Thread Duo Zhang
The HBase team is happy to announce the immediate availability of HBase 3.0.0-alpha-4. Apache HBase™ is an open-source, distributed, versioned, non-relational database. Apache HBase gives you low latency random access to billions of rows with millions of columns atop non-specialized hardware. To

Re: Hmaster starting issue.

2023-06-07 Thread Duo Zhang
Do you want to start a single node pseudo distributed hbase cluster, or a fully distributed cluster? For the former one, you do not need to start hdfs or zookeeper, just type ./start-hbase.sh with default configuration is enough. If you want a fully distributed cluster, and want it to use pre

Re: Call for Presentations, Community Over Code Asia 2023

2023-06-05 Thread Duo Zhang
native, no other self deployed services other than HBase itself - New WAL implementation: BookKeeper? Embedded WAL service? Plan to create a google doc online so others in the community could also help polishing the slides, as I do not know all the details about these features. Thanks. Duo

Fwd: Call for Presentations, Community Over Code Asia 2023

2023-06-05 Thread Duo Zhang
FYI. The CFP deadline has been extended to 6.18, please submit your representations :) 会议的中文介绍见:https://www.bagevent.com/event/8409854 -- Forwarded message - 发件人: Rich Bowen Date: 2023年6月6日周二 00:09 Subject: Call for Presentations, Community Over Code Asia 2023 To:

Re: HBase2 Orphan Regions on RegionServer

2023-06-05 Thread Duo Zhang
建议 grep 一下 master 的日志,这种大概率是已经被合并的 region,只是没有正确的 offline,可以看看当前表里和他重叠的那个 region 是怎么合并出来的 在 2.3 以后这种情况会少很多,主要是 AM-v2 和 procedure 存储的部分做了不少改进和优化 leojie 于2023年6月5日周一 16:09写道: > hi all > 请教社区一些HBase2中 Orphan Regions的问题,使用的hbase版本是2.2.6,hadoop版本是3.3.2 > HBase的hbck report页面出现了一些Orphan Regions, >

Re: TimeoutIOException: Failed to get sync result after 300000 ms for txid=16920651960, WAL system stuck?

2023-05-31 Thread Duo Zhang
, > > 停止该任务后,这两天集群指标比较稳定了,没有发现WAL system stuck?的异常了。 > > > > 感觉DN有压力是引起此现象的一个导火索,但可能不是sync > > wal操作hang住一两个小时后,才导致RS挂掉的直接原因,我试试升级jdk,再研读下这块的代码 > > > > 张铎(Duo Zhang) 于2023年5月13日周六 18:58写道: > > > >> 这看起来还是网络有抖动所以连接断了?建议还是查查出这个 log 时候集群的各种指标 > >

Re: TimeoutIOException: Failed to get sync result after 300000 ms for txid=16920651960, WAL system stuck?

2023-05-13 Thread Duo Zhang
iled > org.apache.hbase.thirdparty.io.netty.channel.unix.Errors$NativeIoException: > syscall:read(..) failed: Connection reset by peer > at > > org.apache.hbase.thirdparty.io.netty.channel.unix.FileDescriptor.readAddress(..)(Unknown > Source) > 2023-05-11 10:59:33,229 INFO [HBase

Re: TimeoutIOException: Failed to get sync result after 300000 ms for txid=16920651960, WAL system stuck?

2023-05-10 Thread Duo Zhang
你往上翻翻有没有别的异常?这个看起来应该就是 AsyncFSWAL 有 bug 导致 hang 住不动了,不过我翻了一下,2.2.6 之后似乎没有跟这个有关的 fix 了,之前倒是有一些。 另外你的 jdk 版本是多少?我印象里 jdk8 早期版本 synchronized 有个 bug 可能会导致执行顺序错乱 leojie 于2023年5月9日周二 14:45写道: > hi all > 向社区求助一个问题,这两天总是在12:50左右遇到一个异常,描述如下: > hbase版本:2.2.6 > hadoop版本:3.3.1 >

Re: 关于offPeakCompactionTracker的疑惑

2023-05-04 Thread Duo Zhang
可以的,建 issue 吧,先不着急改具体代码 或者你可以内部先改改测试一下,然后把结果啥的也贴到 issue 上 章啸 于2023年5月4日周四 13:35写道: > 我是发现OffPeak compaction并不能充分利用低峰期的集群资源去合并高峰期生成的hfile,因为同一时间只能有一个 off peak > compact。 > 似乎这样并不是很合理,我想尝试修改这里(去掉static,改成一个store > 内同一时间只有一个),但是不理解这样设计的初衷,以及我这样修改是否有问题。 > 我想我可以创建一个jira,把我的修改思路提上来。 > > > 在

Re: 关于offPeakCompactionTracker的疑惑

2023-05-03 Thread Duo Zhang
我 blame 翻了一下 https://github.com/apache/hbase/blob/5998a0f349824adf823f79a52530e97dfc624b92/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/compactions/OffPeakCompactions.java 这个 AtomicBoolean 的作用其实就是替代这个文件里的一些逻辑的,在 HBASE-7437 的改动里把这个文件去掉了

Re: [ANNOUNCE] New HBase committer Nihal Jain

2023-05-03 Thread Duo Zhang
Congratulations! Viraj Jasani 于2023年5月3日周三 23:47写道: > Congratulations Nihal!! Very well deserved!! > > On Wed, May 3, 2023 at 5:12 AM Nick Dimiduk wrote: > > > Hello! > > > > On behalf of the Apache HBase PMC, I am pleased to announce that Nihal > Jain > > has accepted the PMC's invitation to

Re: Re: Re: hbase集群单个region flush时间过长问题求助

2023-04-02 Thread Duo Zhang
2.1 上的代码我记的不太清楚了,不过 region assign 相关 procedure 在 2.2 重新写了一遍,之前确实有 bug 很难搞定,常见的就是 region 不下线,或者一个 region 被 assign 到两台 RS 上,只能重启搞定。另外后来又在 2.3 上把存储 procedure 的部分重新实现了一遍,才算比较稳定。 所以 2.1 在大集群特别是有些过载的情况下跑 merge 确实有可能出 bug,hbck 的处理没啥问题,也只能这样搞。。。 邢* 于2023年3月31日周五 15:15写道: > >

Re: 大表快照scan场景中,遇到快照删除,但是快照引用的hfile未删除的现象

2023-04-02 Thread Duo Zhang
嗯,逻辑上没问题,就是写法,你可以先在自己的仓库上修了试一下 leojie 于2023年4月3日周一 09:04写道: > 感谢张老师的回复,这个PR是我们组之前一个同事提的,,这个修复没啥问题,应该是可以直接应用的吧 > > 张铎(Duo Zhang) 于2023年3月30日周四 12:09写道: > > > https://issues.apache.org/jira/browse/HBASE-25510 > > > > 这个,不过还没合,也很久没动静了,我再跟一下吧 > > > > leoj

Re: Re: hbase集群单个region flush时间过长问题求助

2023-03-30 Thread Duo Zhang
> is closed by DFSClient_NONMAPREDUCE_-1935786627_1 > 2023-03-30 12:22:53,311 DEBUG org.apache.hadoop.net.NetworkTopology: > Choosing random from 13 available nodes on node /, scope=, > excludedScope=/TX/T01/03, excludeNodes=[x.x.x.x:9866]. numOfDatanodes=13. > 2023-03-30 12:22:53,311 DEBUG org.apache.h

Re: 大表快照scan场景中,遇到快照删除,但是快照引用的hfile未删除的现象

2023-03-29 Thread Duo Zhang
https://issues.apache.org/jira/browse/HBASE-25510 这个,不过还没合,也很久没动静了,我再跟一下吧 leojie 于2023年3月30日周四 11:43写道: > 好的,感谢张老师回复,我搜下ISSUE列表 > > 张铎(Duo Zhang) 于2023年3月29日周三 21:44写道: > > > 这都是 cleaner 的线程,能跑到这里说明还是在扫描看文件是否可以被删除的,并不是判断有 snapshot 正在执行就直接跳过执行? > > > > 可以看看是不是就是删

Re: hbase集群单个region flush时间过长问题求助

2023-03-29 Thread Duo Zhang
这个要看你 flush 出去的数据有多大吧?以及写 HDFS 的速度,十几秒通常看还行吧,不算特别慢。有监控打 snapshot 时候集群的整体负载和 IO 情况吗? 邢* 于2023年3月29日周三 19:20写道: > hi all, > > > 向社区求助一个HBase的情况,情况描述如下:我们hbase集群某张大表最近做snapshot的时间过长,耗时约半小时。梳理日志发现每个region > flush的时间在十几秒左右,想咨询下flush的时间为什么会这么久? > > 集群现状:目前是有14台cpu内存都是96c

Re: 大表快照scan场景中,遇到快照删除,但是快照引用的hfile未删除的现象

2023-03-29 Thread Duo Zhang
ter.cleaner.CleanerChore.deleteAction(CleanerChore.java:442) > at > org.apache.hadoop.hbase.master.cleaner.CleanerChore.traverseAndDelete(CleanerChore.java:387) > at > org.apache.hadoop.hbase.master.cleaner.CleanerChore.lambda$null$2(CleanerChore.java:396) > at > org.apache.had

Re: 大表快照scan场景中,遇到快照删除,但是快照引用的hfile未删除的现象

2023-03-27 Thread Duo Zhang
那个日志的意思是有 snapshot 正在执行,所以不会跑 cleaner。可以查一下是否有正在执行的 snapshot 操作卡住了之类的 leojie 于2023年3月27日周一 20:53写道: > hi all, > > 向社区求助一个HBase的情况,情况描述如下:在我们快照scan的场景中,有些大表会较为频繁的做快照(如天级),但是这些快照元数据删除后,快照引用的hfile貌似未清理,这体现在,我们的集群archive目录空间占用嗖嗖往上涨。hmaster中,只找到如下貌似相关的日志: > [image: image.png] > 2023-03-27

[ANNOUNCE] New HBase committer Tianhang Tang(唐天航)

2023-03-15 Thread Duo Zhang
On behalf of the Apache HBase PMC, I am pleased to announce that Tianhang Tang(thangTang) has accepted the PMC's invitation to become a committer on the project. We appreciate all of Tianhang's generous contributions thus far and look forward to his continued involvement. Congratulations and

[ANNOUNCE] Apache HBase 2.4.16 is now available for download

2023-02-05 Thread Duo Zhang
The HBase team is happy to announce the immediate availability of HBase 2.4.16. Apache HBase™ is an open-source, distributed, versioned, non-relational database. Apache HBase gives you low latency random access to billions of rows with millions of columns atop non-specialized hardware. To learn

HBase Quarterly report Oct-Dec 2022

2023-02-05 Thread Duo Zhang
Hi all, HBase submits a report to the ASF board once a quarter, to inform the board about project health. I'm sending the report to the user@ and dev@ mailing lists because you are the project, and for transparency. If you have any questions about the report or the running of the project, you can

[ANNOUNCE] Please welcome Tak Lon (Stephen) Wu to the HBase PMC

2023-01-29 Thread Duo Zhang
On behalf of the Apache HBase PMC I am pleased to announce that Tak Lon (Stephen) Wu has accepted our invitation to become a PMC member on the Apache HBase project. We appreciate Tak Lon (Stephen) Wu stepping up to take more responsibility in the HBase project. Please join me in welcoming Tak Lon

Re: scan 引发未捕获运行时异常导致RS宕机

2023-01-27 Thread Duo Zhang
instanceof HFileBlock)) { > comparison = ((HFileBlock) existing).getNextBlockOnDiskSize() > - ((HFileBlock) newBlock).getNextBlockOnDiskSize(); > } > return comparison; > } > > 这里comparison != 0这个判断逻辑是必须的? > > > 张铎(Duo Zhang) 于2023年1月28日周六 11:09写道: > >

Re: scan 引发未捕获运行时异常导致RS宕机

2023-01-27 Thread Duo Zhang
建议先升级到2.4或者2.5的最新版试试?2.2.6 已经是比较老的版本了,后面 BucketCache 修过不少小 bug https://issues.apache.org/jira/browse/HBASE-26281 比如这个就有可能导致BucketCache里的内容错乱,出现各种奇怪的错误 leojie 于2023年1月28日周六 10:22写道: > > Hi all, > 向社区求助一个HBase的scan异常导致RS宕机的问题,异常日志如下: > > 2023-01-19 03:19:06,986 ERROR >

[ANNOUNCE] New HBase committer Rushabh Shah

2022-12-14 Thread Duo Zhang
On behalf of the Apache HBase PMC, I am pleased to announce that Rushabh Shah(shahrs87) has accepted the PMC's invitation to become a committer on the project. We appreciate all of Rushabh's generous contributions thus far and look forward to his continued involvement. Congratulations and

[ANNOUNCE] Apache HBase 2.5.2 is now available for download

2022-12-04 Thread Duo Zhang
The HBase team is happy to announce the immediate availability of HBase 2.5.2. Apache HBase™ is an open-source, distributed, versioned, non-relational database. Apache HBase gives you low latency random access to billions of rows with millions of columns atop non-specialized hardware. To learn

Re: [ANNOUNCE] New HBase Committer Liangjun He

2022-12-03 Thread Duo Zhang
Congratulations! Yu Li 于2022年12月3日周六 21:51写道: > > Hi All, > > On behalf of the Apache HBase PMC, I am pleased to announce that Liangjun > He (heliangjun) has accepted the PMC's invitation to become a committer on > the project. We appreciate all of Liangjun's generous contributions thus > far

HBase Quarterly report Jul-Sep 2022

2022-10-22 Thread Duo Zhang
Hi all, HBase submits a report to the ASF board once a quarter, to inform the board about project health. I'm sending the report to the user@ and dev@ mailing lists because you are the project, and for transparency. If you have any questions about the report or the running of the project, you can

[ANNOUNCE] Apache HBase 1.7.2 is now available for download

2022-08-09 Thread Duo Zhang
The HBase team is happy to announce the immediate availability of HBase1.7.2. Apache HBase™ is an open-source, distributed, versioned, non-relational database. Apache HBase gives you low latency random access to billions of rows with millions of columns atop non-specialized hardware. To learn

[ANNOUNCE] Apache HBase 3.0.0-alpha-3 is now available for download

2022-06-30 Thread Duo Zhang
The HBase team is happy to announce the immediate availability of HBase 3.0.0-alpha-3. Apache HBase™ is an open-source, distributed, versioned, non-relational database. Apache HBase gives you low latency random access to billions of rows with millions of columns atop non-specialized hardware. To

HBase Quarterly report Jan-Mar 2022

2022-04-26 Thread Duo Zhang
Hi all, HBase submits a report to the ASF board once a quarter, to inform the board about project health. I'm sending the report to the user@ and dev@ mailing lists because you are the project, and for transparency. If you have any questions about the report or the running of the project, you can

[ANNOUNCE] Please welcome Xiaolin Ha(哈晓琳) to the HBase PMC

2022-04-09 Thread Duo Zhang
On behalf of the Apache HBase PMC I am pleased to announce that Xiaolin Ha has accepted our invitation to become a PMC member on the Apache HBase project. We appreciate Xiaolin Ha stepping up to take more responsibility in the HBase project. Please join me in welcoming Xiaolin Ha to the HBase

[ANNOUNCE] New HBase committer Bryan Beaudreault

2022-04-09 Thread Duo Zhang
On behalf of the Apache HBase PMC, I am pleased to announce that Bryan Beaudreault(bbeaudreault) has accepted the PMC's invitation to become a committer on the project. We appreciate all of Bryan's generous contributions thus far and look forward to his continued involvement. Congratulations and

[通知] Apache HBase 2.3 结束维护

2022-03-25 Thread Duo Zhang
Apache HBase 2.3 已经在 2021 年 10 月进入停止维护状态。很抱歉通知发的晚了一些。 Apache HBase 2.3 的最后一个版本 2.3.7 可以在 archive.apache.org 下载到,后续我们不会再发布新的 2.3 版本。 建议使用 HBase 2.3.z 的用户尽快升级到当前的稳定版本(目前是 2.4.10)。 感谢所有 HBase 2.3 的贡献者。

[ANNOUNCE] New HBase committer Yutong Xiao(肖禹同)

2022-03-01 Thread Duo Zhang
On behalf of the Apache HBase PMC, I am pleased to announce that Yutong Xiao(YutSean) has accepted the PMC's invitation to become a committer on the project. We appreciate all of Yutong's generous contributions thus far and look forward to his continued involvement. Congratulations and welcome,

[ANNOUNCE] New HBase committer Lei Cheng(程磊)

2022-03-01 Thread Duo Zhang
On behalf of the Apache HBase PMC, I am pleased to announce that Lei Cheng(comnetwork) has accepted the PMC's invitation to become a committer on the project. We appreciate all of Lei's generous contributions thus far and look forward to his continued involvement. Congratulations and welcome, Lei

HBase Quarterly report Oct-Dec 2021

2022-02-06 Thread Duo Zhang
Hi all, HBase submits a report to the ASF board once a quarter, to inform the board about project health. I'm sending the report to the user@ and dev@ mailing lists because you are the project, and for transparency. If you have any questions about the report or the running of the project, you can

[ANNOUNCE] Apache HBase 3.0.0-alpha-2 is now available for download

2021-12-23 Thread Duo Zhang
The HBase team is happy to announce the immediate availability of HBase 3.0.0-alpha-2. Apache HBase™ is an open-source, distributed, versioned, non-relational database. Apache HBase gives you low latency random access to billions of rows with millions of columns atop non-specialized hardware. To

Re: [NOTICE] Apache log4j2 security vulnerability

2021-12-10 Thread Duo Zhang
io/docs/blog/log4j-zero-day/ > > -Stephen > > On Thu, Dec 9, 2021 at 8:09 PM 张铎(Duo Zhang) > wrote: > > > See this PR > > > > https://github.com/apache/logging-log4j2/pull/608 > > > > Although the final 2.15.0 release for log4j2 has not been publish

[NOTICE] Apache log4j2 security vulnerability

2021-12-09 Thread Duo Zhang
See this PR https://github.com/apache/logging-log4j2/pull/608 Although the final 2.15.0 release for log4j2 has not been published yet, at least on the Chinese internet the details and how to make use of this vulnerability has already been public[1]. HBase 3.0.0-alpha-1 is affected, so once

Re: HBase Quarterly report Aug-Oct 2021

2021-11-07 Thread Duo Zhang
Oh, sorry, the title should be Jul-Sep,I forgot to send it last month :( 张铎(Duo Zhang) 于2021年11月8日周一 上午7:58写道: > > Hi all, > > HBase submits a report to the ASF board once a quarter, to inform the board > about project health. I'm sending the report to the user@ and dev@ mailing

HBase Quarterly report Aug-Oct 2021

2021-11-07 Thread Duo Zhang
Hi all, HBase submits a report to the ASF board once a quarter, to inform the board about project health. I'm sending the report to the user@ and dev@ mailing lists because you are the project, and for transparency. If you have any questions about the report or the running of the project, you can

[ANNOUNCE] Apache HBase 1.4.14 is now available for download

2021-10-26 Thread Duo Zhang
The HBase team is happy to announce the immediate availability of HBase1.4.14. Apache HBase™ is an open-source, distributed, versioned, non-relational database. Apache HBase gives you low latency random access to billions of rows with millions of columns atop non-specialized hardware. To learn

[ANNOUNCE] New HBase Committer Xiaolin Ha(哈晓琳)

2021-05-15 Thread Duo Zhang
On behalf of the Apache HBase PMC, I am pleased to announce that Xiaolin Ha(sunhelly) has accepted the PMC's invitation to become a committer on the project. We appreciate all of Xiaolin's generous contributions thus far and look forward to her continued involvement. Congratulations and welcome,

Re: 求助HBase社区关于HBase2.1.0 主备复制的一系列异常现象

2021-01-05 Thread Duo Zhang
@gmail.com; > 发送时间:2021年1月4日(星期一) 晚上10:01 > 收件人:"user-zh" > 主题:Re: 求助HBase社区关于HBase2.1.0 主备复制的一系列异常现象 > > > > 哦,看到了,在最上面有 > > 你这个是开了serial replication?有必要吗?感觉有可能是这个地方出了bug之类的,导致某些日志没推出去然后后面就全卡了 > > 张铎(Duo Zhang) > shell里敲 > > list_peers >

Re: 求助HBase社区关于HBase2.1.0 主备复制的一系列异常现象

2021-01-05 Thread Duo Zhang
< > palomino...@gmail.com; > 发送时间:2021年1月4日(星期一) 晚上10:01 > 收件人:"user-zh" > 主题:Re: 求助HBase社区关于HBase2.1.0 主备复制的一系列异常现象 > > > > 哦,看到了,在最上面有 > > 你这个是开了serial replication?有必要吗?感觉有可能是这个地方出了bug之类的,导致某些日志没推出去然后后面就全卡了 > >

Re: 求助HBase社区关于HBase2.1.0 主备复制的一系列异常现象

2021-01-04 Thread Duo Zhang
哦,看到了,在最上面有 你这个是开了serial replication?有必要吗?感觉有可能是这个地方出了bug之类的,导致某些日志没推出去然后后面就全卡了 张铎(Duo Zhang) 于2021年1月4日周一 下午9:58写道: > shell里敲 > > list_peers > > sudo rm -rf /* <2326130...@qq.com> 于2021年1月4日周一 下午9:33写道: > >> hi >> 感谢您的回复,日志中没有Replication的ERROR信息,replicati

Re: 求助HBase社区关于HBase2.1.0 主备复制的一系列异常现象

2021-01-04 Thread Duo Zhang
shell里敲 list_peers sudo rm -rf /* <2326130...@qq.com> 于2021年1月4日周一 下午9:33写道: > hi > 感谢您的回复,日志中没有Replication的ERROR信息,replication peer是啥状态?这句话的意思是? 截图放在附件中,麻烦查收 > > hbase(main):002:0> list_peers > PEER_ID CLUSTER_KEY ENDPOINT_CLASSNAME STATE REPLICATE_ALL NAMESPACES > TABLE_CFS BANDWIDTH SERIAL

Re: 求助HBase社区关于HBase2.1.0 主备复制的一系列异常现象

2021-01-04 Thread Duo Zhang
你这个只有第一条是和replication直接相关的,这个就是说queue太长了,有replication报错的日志吗? 另外你具体的replication peer是啥状态? sudo rm -rf /* <2326130...@qq.com> 于2021年1月4日周一 上午11:27写道: > hi > 刚刚截取了一些日志中频繁打印的与Replication相关的日志信息,您帮忙看看对分析问题是否有帮助。图片在附件,对应文字在下文 > > 上午11点21:01.959分 WARN ReplicationSource > > WAL group >

Re: 求助HBase社区关于HBase2.1.0 主备复制的一系列异常现象

2021-01-03 Thread Duo Zhang
图片还是看不到。。。 日志里有什么有用的信息吗?是根本推不过去还是推的很慢?oldWALs是所有regionserver的日志都在堆积,还是只有某些在堆积? sudo rm -rf /* <2326130...@qq.com> 于2021年1月4日周一 上午10:45写道: > hi, > 非常抱歉,需要再次打扰,HBase replication 目前处于阻塞状态,某一WAL > log已经卡在一端有三个多月,我截图了ReplicationSource > 相关线程的一些WATTING和BLOCKED的状态,需要社区帮忙分析定位一下原因。 > 图片请查看附件,感谢感谢 >

Re: Hbase2.2.3写入卡主

2020-12-16 Thread Duo Zhang
是只要一跑checkAndPut就必现吗?还是跑着跑着就会出错?这个看着就是某次写wal完了(不一定写成功)没有推MVCC 凡尔 于2020年12月16日周三 下午9:25写道: > 问题现象 > > 在使用kylin2.6.6+hbase2.2.3中 > > 服务端出现大量如下警告: > > TRACE [handler=88] ipc.RpcServer: called: 165045 service :ClientService > methodName: Mutate size…… > > > > > WARN[handler=88]

[ANNOUNCE] New HBase Committer Zheng Wang(王正)

2020-09-23 Thread Duo Zhang
On behalf of the Apache HBase PMC, I am pleased to announce that Zheng Wang has accepted the PMC's invitation to become a committer on the project. We appreciate all of Zheng's generous contributions thus far and look forward to his continued involvement. Congratulations and welcome, Zheng Wang!

Re: hbase 主备集群同时开启replication后WALs目录越积越大

2020-07-23 Thread Duo Zhang
对,看不到,一片白。 另外zheng wang你用的是什么邮件客户端,为啥我在gmail里每次看到的你邮件都是一堆转义字符,空格变成nbsp之类的。。。 zheng wang <18031...@qq.com> 于2020年7月23日周四 下午2:21写道: > 看不到你贴的东西,一片白的。 > > > > > --原始邮件-- > 发件人: > "user-zh" >

Re: hbase 主备集群同时开启replication后WALs目录越积越大

2020-07-21 Thread Duo Zhang
对,你看看卡在哪了到底 sudo rm -rf /* 于2020年7月22日周三 上午11:54写道: > 机器已经重启过了,现在replication是没有延时的,昨天的异常场景 > 现在没办法重现,只是重启过机器后,replication已经没有延时了,WALs 还在持续增长,现在WALs的量从昨晚到现在已经增至670G, > 还在持续增长,我们RegionServer是5台,RS的堆内存设置的是32G,按理说数据持续flush > ,replication又没有延时,WALS应该及时被清理才对。我们的MajorCompact是手动触发的,目前禁用了Major > > > > >

Re: hbase 主备集群同时开启replication后WALs目录越积越大

2020-07-21 Thread Duo Zhang
你看看为啥有一个region server的replication延迟很大? 是写入有热点?还是对面集群不收? sudo rm -rf /* 于2020年7月22日周三 上午10:59写道: > hi: > 我们线上集群的版本是cdh6.3.2-hbase2.1.0, > 主备双集群,备集群作为容灾集群。两集群开启了replication,且互相replication。最近发现,主集群的WALs目录持续增加, > 十几天膨胀到十几个T,排查后发现某一个RegionServer节点的replication延迟很大,复制队列中积累了大量的数据,延迟卡在了十几天前的WALs >

Re: 开启replication后,hbase读写性能下降问题

2020-07-20 Thread Duo Zhang
写入的很猛吗?需要看看瓶颈在哪,我们之前遇到过peer很多的情况下,replication对HDFS的压力很大,因为每个peer都要读一遍WAL。 你这个是SSD的话,看起来是replication本身对region server的压力很大? zheng wang <18031...@qq.com> 于2020年7月21日周二 下午12:05写道: > 可以考虑对replication限速。 > > > > > --原始邮件-- > 发件人: > >

Re: HBase资源隔离队列参数貌似不生效

2020-06-02 Thread Duo Zhang
图好像看不到? sudo rm -rf /* 于2020年6月3日周三 上午10:00写道: > 配置了队列隔离参数,参数如下: > hbase.regionserver.handler.count 30 > hbase.ipc.server.callqueue.handler.factor 1 > hbase.ipc.server.callqueue.read.ratio 0.5 > hbase.ipc.server.callqueue.scan.ratio 0.1 > > 这样我控制scan的线程为1, > 此时,查看线程的分配情况 > > >

Re: 关于ServerSideScanMetrics类设计的问题

2020-05-14 Thread Duo Zhang
这个东西目前看主要是mapreduce的framework会用,所以暴露了比较多的内部方法,你可以看一下TableRecordReaderImpl的updateCounters方法 理论上说,直接返回AtomicLong肯定是不好的,因为没法控制用户到底怎么操作,但修改一个IA.Public的类总是比较麻烦的,因为很有可能已经有用户在以这种方式使用了。。。 韩聪聪 于2020年5月14日周四 下午1:57写道: > hi,all > >

[通知] 莊偉赳(Wei-Chiu Chuang)成为 HBase Committer

2020-05-13 Thread Duo Zhang
非常荣幸代表 Apache HBase 项目管理委员会宣布,莊偉赳(Wei-Chiu Chuang)已经接受了我们的邀请,成为了一名HBase的Committer。 我们感谢莊偉赳一直以来为社区做出的巨大贡献,并期待他在将来继续发挥更大的价值。 请和我一起祝贺莊偉赳。 谢谢! 张铎,代表 Apache HBase 项目管理委员会

Re: Re: Re: Re: hbase读耗时问题排查

2020-03-21 Thread Duo Zhang
不加这个参数会导致Unsafe无法使用吧,目前很少有人测试过没有Unsafe的情况下HBase是啥表现,可能会有些bug,你加上试试看? 于2020年3月21日 周六10:40写道: > 没加这个参数,需要加上吗? > - 原始邮件 - > 发件人:张铎(Duo Zhang) > 收件人:ww112...@sina.com > 抄送人:user-zh > 主题:Re: Re: Re: hbase读耗时问题排查 > 日期:2020年03月16日 12点22分 > > 应该不会的,这里也没有用什么字节码生成的东西 &

Re: Re: Re: hbase读耗时问题排查

2020-03-15 Thread Duo Zhang
应该不会的,这里也没有用什么字节码生成的东西 我怀疑还是Unsafe那块export的不对?你们在启动的时候有加类似这样的参数吗? --add-exports java.base/jdk.internal.misc=ALL-UNNAMED 于2020年3月13日周五 上午10:08写道: > 有可能每次序列化的类不一样吗? > > > - 原始邮件 ----- > 发件人:张铎(Duo Zhang) > 收件人:ww112...@sina.com > 抄送人:user-zh > 主题:Re: Re: hbase读耗时问

Re: Re: hbase读耗时问题排查

2020-03-12 Thread Duo Zhang
这是Class.forName特别慢?不太对劲啊,按理说class只要load一次就不会再load了? 不过JDK11的话,是不是Unsafe这个类没有export出来啊。。。 比如netty,如果想完整的把所有性能都发挥出来,得加一堆参数 https://github.com/netty/netty/issues/7838 于2020年3月12日周四 下午3:21写道: > > hbase版本是2.1.0,jdk版本是11,回收器使用的zgc。 > > > > 当时的火焰图,可以参照附件: > > > ----- 原始

Re: hbase读耗时问题排查

2020-03-12 Thread Duo Zhang
这个格式有点儿乱,不太能看清楚,是只追到这一个函数耗时很长吗?他调用的函数没有时间长的? 你用的HBase版本是啥? 于2020年3月5日周四 下午4:03写道: > 你好: > > 在排查读超时的时候,追踪到org.apache.hadoop.hbase.io.hfile.HFileBlock$1:deserialize这个方法,发现这个方法经常出现耗时在几百毫秒,甚至出现1s多的情况,这个有什么解决办法。 > > `---ts=2020-03-05 >

Re: 关于HBase-23286 Improve MTTR: Split WAL to HFile的一些想法

2020-03-02 Thread Duo Zhang
我理解,kill -9之后zk是不会马上就知道rs挂了的吧,本来也需要等一段时间?所以那个40s,是从kill -9开始算起,还是从master真正开始split开始算起? mintao 于2020年2月28日周五 下午5:06写道: > 1.测试表只有1个cf; > 2. hbase.split.writer.creation.bounded没有配置,但是我配置了 > (1)hbase.regionserver.hlog.splitlog.writer.threads=20 > (2)hbase.regionserver.wal.maxsplitters=10 > > > | |

Re: Re: HBase 代码 Intellij IDEA 模板是什么

2020-02-26 Thread Duo Zhang
你需要用你订阅的那个邮箱来发邮件,不然都会被拦。。。 chenyec...@cmss.chinamobile.com 于2020年2月27日周四 下午1:58写道: > 谢谢,之前已经订阅了。 > > 测试可以收到回复。 > > -- > chenyec...@cmss.chinamobile.com > > > *发件人:* 张铎(Duo Zhang) > *发送时间:* 2020-02-27 11:27 > *收件人:* user-zh ; c

Re: HBase 代码 Intellij IDEA 模板是什么

2020-02-26 Thread Duo Zhang
Code Template是自动生成文件头和各种注释之类的 Formatter是表示代码风格的,比如用tab还是space缩进,一行最长多少之类 另外你得订阅邮件组再发邮件,不然别人回复你都收不到,而且不订阅邮件组发邮件的话默认会被拦,还得专门放行 chenyec...@cmss.chinamobile.com 于2020年2月27日周四 上午11:24写道: > 大家好: > > 如题,在HBase 代码仓库中发现 > > 1、 >