[jira] [Created] (HBASE-26235) We could start RegionServerTracker before becoming active master

[Duo Zhang (Jira)]

Duo Zhang created HBASE-26235:
-

 Summary: We could start RegionServerTracker before becoming active 
master
 Key: HBASE-26235
 URL: https://issues.apache.org/jira/browse/HBASE-26235
 Project: HBase
  Issue Type: Sub-task
Reporter: Duo Zhang


The direct reason is that, after decoupling, HMaster does not extends 
HRegionServer, so we do not have RegionServerAddressTracker now, so for backup 
masters we do not have a way to get all the region server which means it can 
not serve as connection registry any more, so we need to find a way to let 
backup masters to also have a way to get the region server list.

But in general, since we already have a class in HMaster which is used to track 
region servers, we'd better reuse it. And after this change, the backup masters 
could also be used as the fallback servers for getting region server list in 
HBASE-26220, which could also reduce the load of active master.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

[jira] [Resolved] (HBASE-26232) SEEK_NEXT_USING_HINT is ignored on reversed Scans

[Duo Zhang (Jira)]

 [ 
https://issues.apache.org/jira/browse/HBASE-26232?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Duo Zhang resolved HBASE-26232.
---
Hadoop Flags: Reviewed
  Resolution: Fixed

Pushed to branch-2.3+.

Thanks [~keyserbrian1] for contributing.

> SEEK_NEXT_USING_HINT is ignored on reversed Scans
> -
>
> Key: HBASE-26232
> URL: https://issues.apache.org/jira/browse/HBASE-26232
> Project: HBase
>  Issue Type: Bug
>  Components: Filters
>Affects Versions: 3.0.0-alpha-1, 2.4.5
> Environment: Tested, confirmed, and fixed using the in-memory 
> mini-cluster on 64-bit Ubuntu. Originally discovered with the in-memory 
> mini-cluster on RHEL 7 when writing a custom filter for an unrelated project.
>Reporter: Brian Keyser
>Assignee: Brian Keyser
>Priority: Critical
> Fix For: 2.5.0, 3.0.0-alpha-2, 2.4.6, 2.3.7
>
>
> It turns out that the {{SEEK_NEXT_USING_HINT}} feature is broken for reversed 
> scans. That's because the implementation hard-codes an assumption that the 
> scan is forward.
> Here's the mistake: 
> [https://github.com/keyserbrian1/hbase/blob/735bcf85e9b59a71981babe2f5da51978d61d8d3/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/StoreScanner.java#L740|https://github.com/keyserbrian1/hbase/blob/735bcf85e9b59a71981babe2f5da51978d61d8d3/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/StoreScanner.java#L740]
> The bug sneaks past the existing tests because the mistake isn't detectable 
> by looking at the result of a naive reversed scan.
> Attached is a PR that includes a fix and tests designed to detect if filter 
> hints are actually used.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

[jira] [Created] (HBASE-26236) Simple travis build for hbase-filesystem

[Josh Elser (Jira)]

Josh Elser created HBASE-26236:
--

 Summary: Simple travis build for hbase-filesystem
 Key: HBASE-26236
 URL: https://issues.apache.org/jira/browse/HBASE-26236
 Project: HBase
  Issue Type: Improvement
  Components: hboss
Reporter: Josh Elser
Assignee: Josh Elser
 Fix For: hbase-filesystem-1.0.0-alpha2


Noticed that we don't have any kind of precommit checks. Time to make a quick 
one.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

[VOTE] Merge feature branch HBASE-25853 into branch-2

[Tak Lon (Stephen) Wu]

Hi everyone,

I'm writing this request and propose a merge of HBASE-25853
 "Backport HBASE-22120
Replace HTrace with OpenTelemetry" to branch-2. The goal is to remove
HTrace and uses OpenTelemetry also in branch-2.5.0+.

Highlights
* These changes only support async clients and calls
* We will have a separate thread for support sync client in branch-2 , see
HBASE-26141 

PRs
* The merge PR with 18 reviewed commits
https://github.com/apache/hbase/pull/3637
* see all those reviewed commits at feature branch HBASE-25853


Please vote:
[+1]  Agree
[+/-0]Neutral
[-1]   Disagree (please include actionable feedback)

Thanks,
Stephen

Re: [VOTE] Merge feature branch HBASE-25853 into branch-2

[Duo Zhang]

+1.

The original implementation on master was done by me, and thanks
Stephen for the great work on backporting them to branch-2.

Tak Lon (Stephen) Wu  于2021年8月30日周一 上午11:22写道：

> Hi everyone,
>
> I'm writing this request and propose a merge of HBASE-25853
>  "Backport HBASE-22120
> Replace HTrace with OpenTelemetry" to branch-2. The goal is to remove
> HTrace and uses OpenTelemetry also in branch-2.5.0+.
>
> Highlights
> * These changes only support async clients and calls
> * We will have a separate thread for support sync client in branch-2 , see
> HBASE-26141 
>
> PRs
> * The merge PR with 18 reviewed commits
> https://github.com/apache/hbase/pull/3637
> * see all those reviewed commits at feature branch HBASE-25853
> 
>
> Please vote:
> [+1]  Agree
> [+/-0]Neutral
> [-1]   Disagree (please include actionable feedback)
>
> Thanks,
> Stephen
>

[jira] [Created] (HBASE-26237) Improve computation complexity for primaryRegionCountSkewCostFunctio

[Clara Xiong (Jira)]

Clara Xiong created HBASE-26237:
---

 Summary: Improve computation complexity for 
primaryRegionCountSkewCostFunctio
 Key: HBASE-26237
 URL: https://issues.apache.org/jira/browse/HBASE-26237
 Project: HBase
  Issue Type: Bug
  Components: Balancer
Reporter: Clara Xiong


Recomputation of primaryRegionCountSkewCostFunction can be reduced from O(n ) 
to O(1) by only incrementing the destination and decrementing the source 
instead of full recompute.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

[jira] [Resolved] (HBASE-26235) We could start RegionServerTracker before becoming active master

[Duo Zhang (Jira)]

 [ 
https://issues.apache.org/jira/browse/HBASE-26235?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Duo Zhang resolved HBASE-26235.
---
Fix Version/s: 3.0.0-alpha-2
   2.5.0
 Hadoop Flags: Reviewed
   Resolution: Fixed

Pushed to master and branch-2.

Thanks [~niuyulin] for reviewing.

> We could start RegionServerTracker before becoming active master
> 
>
> Key: HBASE-26235
> URL: https://issues.apache.org/jira/browse/HBASE-26235
> Project: HBase
>  Issue Type: Sub-task
>  Components: master, Zookeeper
>Reporter: Duo Zhang
>Assignee: Duo Zhang
>Priority: Major
> Fix For: 2.5.0, 3.0.0-alpha-2
>
>
> The direct reason is that, after decoupling, HMaster does not extends 
> HRegionServer, so we do not have RegionServerAddressTracker now, so for 
> backup masters we do not have a way to get all the region server which means 
> it can not serve as connection registry any more, so we need to find a way to 
> let backup masters to also have a way to get the region server list.
> But in general, since we already have a class in HMaster which is used to 
> track region servers, we'd better reuse it. And after this change, the backup 
> masters could also be used as the fallback servers for getting region server 
> list in HBASE-26220, which could also reduce the load of active master.
> And this could also speed up the active master start process as we could save 
> one zookeeper scan when becoming active master.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)