[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2014-02-07 Thread Hudson (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13894338#comment-13894338
 ] 

Hudson commented on HBASE-9892:
---

FAILURE: Integrated in hbase-0.96 #283 (See 
[https://builds.apache.org/job/hbase-0.96/283/])
HBASE-10340 [BACKPORT] HBASE-9892 Add info port to ServerName to support multi 
instances in a node (stack: rev 1565546)
* 
/hbase/branches/0.96/hbase-protocol/src/main/java/org/apache/hadoop/hbase/protobuf/generated/HBaseProtos.java
* /hbase/branches/0.96/hbase-protocol/src/main/protobuf/HBase.proto
* 
/hbase/branches/0.96/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/MasterStatusTmpl.jamon
* 
/hbase/branches/0.96/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/RegionServerListTmpl.jamon
* 
/hbase/branches/0.96/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/regionserver/RSStatusTmpl.jamon
* 
/hbase/branches/0.96/hbase-server/src/main/java/org/apache/hadoop/hbase/LocalHBaseCluster.java
* 
/hbase/branches/0.96/hbase-server/src/main/java/org/apache/hadoop/hbase/master/HMaster.java
* 
/hbase/branches/0.96/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
* 
/hbase/branches/0.96/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/RSDumpServlet.java
* 
/hbase/branches/0.96/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/RSStatusServlet.java
* 
/hbase/branches/0.96/hbase-server/src/main/java/org/apache/hadoop/hbase/zookeeper/RegionServerTracker.java
* 
/hbase/branches/0.96/hbase-server/src/main/resources/hbase-webapps/master/table.jsp


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Critical
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-0.94-v6.diff, HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2014-02-07 Thread Hudson (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13894463#comment-13894463
 ] 

Hudson commented on HBASE-9892:
---

SUCCESS: Integrated in hbase-0.96-hadoop2 #196 (See 
[https://builds.apache.org/job/hbase-0.96-hadoop2/196/])
HBASE-10340 [BACKPORT] HBASE-9892 Add info port to ServerName to support multi 
instances in a node (stack: rev 1565546)
* 
/hbase/branches/0.96/hbase-protocol/src/main/java/org/apache/hadoop/hbase/protobuf/generated/HBaseProtos.java
* /hbase/branches/0.96/hbase-protocol/src/main/protobuf/HBase.proto
* 
/hbase/branches/0.96/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/MasterStatusTmpl.jamon
* 
/hbase/branches/0.96/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/RegionServerListTmpl.jamon
* 
/hbase/branches/0.96/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/regionserver/RSStatusTmpl.jamon
* 
/hbase/branches/0.96/hbase-server/src/main/java/org/apache/hadoop/hbase/LocalHBaseCluster.java
* 
/hbase/branches/0.96/hbase-server/src/main/java/org/apache/hadoop/hbase/master/HMaster.java
* 
/hbase/branches/0.96/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
* 
/hbase/branches/0.96/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/RSDumpServlet.java
* 
/hbase/branches/0.96/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/RSStatusServlet.java
* 
/hbase/branches/0.96/hbase-server/src/main/java/org/apache/hadoop/hbase/zookeeper/RegionServerTracker.java
* 
/hbase/branches/0.96/hbase-server/src/main/resources/hbase-webapps/master/table.jsp


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Critical
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-0.94-v6.diff, HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2014-02-07 Thread Hudson (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13895189#comment-13895189
 ] 

Hudson commented on HBASE-9892:
---

FAILURE: Integrated in HBase-0.94 #1277 (See 
[https://builds.apache.org/job/HBase-0.94/1277/])
HBASE-10340 [BACKPORT] HBASE-9892 Add info port to ServerName to support multi 
instances in a node (stack: rev 1565828)
* 
/hbase/branches/0.94/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/MasterStatusTmpl.jamon
* /hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/HConstants.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/LocalHBaseCluster.java
* /hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/master/HMaster.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/master/HMasterCommandLine.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/protobuf/generated/HBaseProtos.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/regionserver/RSDumpServlet.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/regionserver/RSStatusServlet.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/zookeeper/RegionServerTracker.java
* /hbase/branches/0.94/src/main/protobuf/hbase.proto
* /hbase/branches/0.94/src/main/resources/hbase-webapps/master/table.jsp


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Critical
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-0.94-v6.diff, HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2014-02-07 Thread Hudson (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13895223#comment-13895223
 ] 

Hudson commented on HBASE-9892:
---

SUCCESS: Integrated in HBase-0.94-security #404 (See 
[https://builds.apache.org/job/HBase-0.94-security/404/])
HBASE-10340 [BACKPORT] HBASE-9892 Add info port to ServerName to support multi 
instances in a node (stack: rev 1565828)
* 
/hbase/branches/0.94/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/MasterStatusTmpl.jamon
* /hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/HConstants.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/LocalHBaseCluster.java
* /hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/master/HMaster.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/master/HMasterCommandLine.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/protobuf/generated/HBaseProtos.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/regionserver/RSDumpServlet.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/regionserver/RSStatusServlet.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/zookeeper/RegionServerTracker.java
* /hbase/branches/0.94/src/main/protobuf/hbase.proto
* /hbase/branches/0.94/src/main/resources/hbase-webapps/master/table.jsp


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Critical
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-0.94-v6.diff, HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2014-02-07 Thread Hudson (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13895271#comment-13895271
 ] 

Hudson commented on HBASE-9892:
---

FAILURE: Integrated in HBase-0.94-on-Hadoop-2 #13 (See 
[https://builds.apache.org/job/HBase-0.94-on-Hadoop-2/13/])
HBASE-10340 [BACKPORT] HBASE-9892 Add info port to ServerName to support multi 
instances in a node (stack: rev 1565828)
* 
/hbase/branches/0.94/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/MasterStatusTmpl.jamon
* /hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/HConstants.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/LocalHBaseCluster.java
* /hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/master/HMaster.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/master/HMasterCommandLine.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/protobuf/generated/HBaseProtos.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/regionserver/RSDumpServlet.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/regionserver/RSStatusServlet.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/zookeeper/RegionServerTracker.java
* /hbase/branches/0.94/src/main/protobuf/hbase.proto
* /hbase/branches/0.94/src/main/resources/hbase-webapps/master/table.jsp


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Critical
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-0.94-v6.diff, HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2014-02-07 Thread Hudson (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13895288#comment-13895288
 ] 

Hudson commented on HBASE-9892:
---

SUCCESS: Integrated in HBase-0.94-JDK7 #42 (See 
[https://builds.apache.org/job/HBase-0.94-JDK7/42/])
HBASE-10340 [BACKPORT] HBASE-9892 Add info port to ServerName to support multi 
instances in a node (stack: rev 1565828)
* 
/hbase/branches/0.94/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/MasterStatusTmpl.jamon
* /hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/HConstants.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/LocalHBaseCluster.java
* /hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/master/HMaster.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/master/HMasterCommandLine.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/protobuf/generated/HBaseProtos.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/regionserver/RSDumpServlet.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/regionserver/RSStatusServlet.java
* 
/hbase/branches/0.94/src/main/java/org/apache/hadoop/hbase/zookeeper/RegionServerTracker.java
* /hbase/branches/0.94/src/main/protobuf/hbase.proto
* /hbase/branches/0.94/src/main/resources/hbase-webapps/master/table.jsp


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Critical
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-0.94-v6.diff, HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2014-01-14 Thread Lars Hofhansl (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13871014#comment-13871014
 ] 

Lars Hofhansl commented on HBASE-9892:
--

This is currently marked fixed, but I do not think it was commited to 0.96.
I would like this in 0.94 as well. Should I reopen, or rather create a 
backporting issue for 0.94 and 0.96?

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Critical
 Fix For: 0.98.0, 0.96.2, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-0.94-v6.diff, HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2014-01-14 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13871354#comment-13871354
 ] 

stack commented on HBASE-9892:
--

[~lhofhansl] I opened HBASE-10340  I'd like to try it first.  Will do soon.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Critical
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-0.94-v6.diff, HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2014-01-13 Thread Enis Soztutar (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13870362#comment-13870362
 ] 

Enis Soztutar commented on HBASE-9892:
--

[~saint@gmail.com] are we going to have this in 0.96? 

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-0.94-v6.diff, HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2014-01-13 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13870451#comment-13870451
 ] 

stack commented on HBASE-9892:
--

I'd like it in 0.96 if it does not break compatibility.  I've not tried it.  I 
can.  Let me file it against 0.96.2 and try it before commit.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-0.94-v6.diff, HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-18 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13852300#comment-13852300
 ] 

stack commented on HBASE-9892:
--

[~apurtell] You want this boss?

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-18 Thread Andrew Purtell (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13852329#comment-13852329
 ] 

Andrew Purtell commented on HBASE-9892:
---

Yes please [~stack]

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-18 Thread Lars Hofhansl (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13852350#comment-13852350
 ] 

Lars Hofhansl commented on HBASE-9892:
--

Would this have any backwards compatibility issues?

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-18 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13852353#comment-13852353
 ] 

stack commented on HBASE-9892:
--

Not going by what [~liushaohui] writes above regards how it works but yeah, in 
0.94, would want to try a rolling upgrade before commit I'd say.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-18 Thread Hudson (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13852487#comment-13852487
 ] 

Hudson commented on HBASE-9892:
---

SUCCESS: Integrated in HBase-0.98 #22 (See 
[https://builds.apache.org/job/HBase-0.98/22/])
HBASE-9892 Add info port to ServerName to support multi instances in a node 
(stack: rev 1552169)
* 
/hbase/branches/0.98/hbase-protocol/src/main/java/org/apache/hadoop/hbase/protobuf/generated/HBaseProtos.java
* /hbase/branches/0.98/hbase-protocol/src/main/protobuf/HBase.proto
* 
/hbase/branches/0.98/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/MasterStatusTmpl.jamon
* 
/hbase/branches/0.98/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/RegionServerListTmpl.jamon
* 
/hbase/branches/0.98/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/regionserver/RSStatusTmpl.jamon
* 
/hbase/branches/0.98/hbase-server/src/main/java/org/apache/hadoop/hbase/LocalHBaseCluster.java
* 
/hbase/branches/0.98/hbase-server/src/main/java/org/apache/hadoop/hbase/master/HMaster.java
* 
/hbase/branches/0.98/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
* 
/hbase/branches/0.98/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/RSDumpServlet.java
* 
/hbase/branches/0.98/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/RSStatusServlet.java
* 
/hbase/branches/0.98/hbase-server/src/main/java/org/apache/hadoop/hbase/zookeeper/RegionServerTracker.java
* 
/hbase/branches/0.98/hbase-server/src/main/resources/hbase-webapps/master/table.jsp


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-18 Thread Liu Shaohui (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13852530#comment-13852530
 ] 

Liu Shaohui commented on HBASE-9892:


[~stack] [~lhofhansl]
I will rebase the code and check the rolling upgrade compatibility for 0.94 
today.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-18 Thread Hudson (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13852591#comment-13852591
 ] 

Hudson commented on HBASE-9892:
---

FAILURE: Integrated in HBase-0.98-on-Hadoop-1.1 #19 (See 
[https://builds.apache.org/job/HBase-0.98-on-Hadoop-1.1/19/])
HBASE-9892 Add info port to ServerName to support multi instances in a node 
(stack: rev 1552169)
* 
/hbase/branches/0.98/hbase-protocol/src/main/java/org/apache/hadoop/hbase/protobuf/generated/HBaseProtos.java
* /hbase/branches/0.98/hbase-protocol/src/main/protobuf/HBase.proto
* 
/hbase/branches/0.98/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/MasterStatusTmpl.jamon
* 
/hbase/branches/0.98/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/RegionServerListTmpl.jamon
* 
/hbase/branches/0.98/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/regionserver/RSStatusTmpl.jamon
* 
/hbase/branches/0.98/hbase-server/src/main/java/org/apache/hadoop/hbase/LocalHBaseCluster.java
* 
/hbase/branches/0.98/hbase-server/src/main/java/org/apache/hadoop/hbase/master/HMaster.java
* 
/hbase/branches/0.98/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
* 
/hbase/branches/0.98/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/RSDumpServlet.java
* 
/hbase/branches/0.98/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/RSStatusServlet.java
* 
/hbase/branches/0.98/hbase-server/src/main/java/org/apache/hadoop/hbase/zookeeper/RegionServerTracker.java
* 
/hbase/branches/0.98/hbase-server/src/main/resources/hbase-webapps/master/table.jsp


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-17 Thread Hudson (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13850256#comment-13850256
 ] 

Hudson commented on HBASE-9892:
---

FAILURE: Integrated in HBase-TRUNK #4730 (See 
[https://builds.apache.org/job/HBase-TRUNK/4730/])
HBASE-9892 Add info port to ServerName to support multi instances in a node 
(stack: rev 1551458)
* 
/hbase/trunk/hbase-protocol/src/main/java/org/apache/hadoop/hbase/protobuf/generated/HBaseProtos.java
* /hbase/trunk/hbase-protocol/src/main/protobuf/HBase.proto
* 
/hbase/trunk/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/MasterStatusTmpl.jamon
* 
/hbase/trunk/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/RegionServerListTmpl.jamon
* 
/hbase/trunk/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/regionserver/RSStatusTmpl.jamon
* 
/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/LocalHBaseCluster.java
* 
/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/master/HMaster.java
* 
/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
* 
/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/RSDumpServlet.java
* 
/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/RSStatusServlet.java
* 
/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/wal/FSHLog.java
* 
/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/zookeeper/RegionServerTracker.java
* /hbase/trunk/hbase-server/src/main/resources/hbase-webapps/master/table.jsp
* 
/hbase/trunk/hbase-server/src/test/java/org/apache/hadoop/hbase/regionserver/wal/HLogPerformanceEvaluation.java


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-17 Thread Hudson (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13851233#comment-13851233
 ] 

Hudson commented on HBASE-9892:
---

FAILURE: Integrated in HBase-TRUNK-on-Hadoop-1.1 #9 (See 
[https://builds.apache.org/job/HBase-TRUNK-on-Hadoop-1.1/9/])
HBASE-9892 Add info port to ServerName to support multi instances in a node 
(stack: rev 1551458)
* 
/hbase/trunk/hbase-protocol/src/main/java/org/apache/hadoop/hbase/protobuf/generated/HBaseProtos.java
* /hbase/trunk/hbase-protocol/src/main/protobuf/HBase.proto
* 
/hbase/trunk/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/MasterStatusTmpl.jamon
* 
/hbase/trunk/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/RegionServerListTmpl.jamon
* 
/hbase/trunk/hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/regionserver/RSStatusTmpl.jamon
* 
/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/LocalHBaseCluster.java
* 
/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/master/HMaster.java
* 
/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
* 
/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/RSDumpServlet.java
* 
/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/RSStatusServlet.java
* 
/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/wal/FSHLog.java
* 
/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/zookeeper/RegionServerTracker.java
* /hbase/trunk/hbase-server/src/main/resources/hbase-webapps/master/table.jsp
* 
/hbase/trunk/hbase-server/src/test/java/org/apache/hadoop/hbase/regionserver/wal/HLogPerformanceEvaluation.java


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-16 Thread Enis Soztutar (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13849602#comment-13849602
 ] 

Enis Soztutar commented on HBASE-9892:
--

Go for it!

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-16 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13849648#comment-13849648
 ] 

stack commented on HBASE-9892:
--

[~liushaohui] One thought I had before commit is what happens for say the case 
where it is an existing cluster and the znode has empty data?  Will we use the 
info port from the configuration?  What if we do a rolling restart when the 
znode has no data in it?  Who writes the znode data? Will new servers be able 
to work if the znode has no data in it?   Thanks.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-16 Thread Liu Shaohui (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13850039#comment-13850039
 ] 

Liu Shaohui commented on HBASE-9892:


@stack
{quote}
One thought I had before commit is what happens for say the case where it is an 
existing cluster and the znode has empty data? Will we use the info port from 
the configuration? 
{quote}
Yes.  See the code in RegionServerTracker. if  the znode has empty data or 
invalid data, default infoPort will be 0, and getRegionServerInfoPort in 
HMaster will return the REGIONSERVER_INFO_PORT from config.

{quote}
What if we do a rolling restart when the znode has no data in it? Who writes 
the znode data? Will new servers be able to work if the znode has no data in it?
{quote}
RegionServerTracker will handle this situation.
For rs with old version, its znode has no data and hmaster will use info port 
from config.  When new region servers start up, it will write the RSinfo data 
to it's znode and RegionServerTracker will watch this and use the data from 
znode.


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-16 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13850099#comment-13850099
 ] 

stack commented on HBASE-9892:
--

Committed to trunk.

[~apurtell] You want this?  Good for operators.  [~liushaohui] needs it to run 
multiple RS on the one node.  I will take it in 0.96 if you take it in 0.98.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-16 Thread Hadoop QA (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13850128#comment-13850128
 ] 

Hadoop QA commented on HBASE-9892:
--

{color:red}-1 overall{color}.  Here are the results of testing the latest 
attachment 
  
http://issues.apache.org/jira/secure/attachment/12619023/HBASE-9892-trunk-v3.diff
  against trunk revision .

{color:green}+1 @author{color}.  The patch does not contain any @author 
tags.

{color:red}-1 tests included{color}.  The patch doesn't appear to include 
any new or modified tests.
Please justify why no new tests are needed for this 
patch.
Also please list what manual steps were performed to 
verify this patch.

{color:green}+1 hadoop1.0{color}.  The patch compiles against the hadoop 
1.0 profile.

{color:green}+1 hadoop1.1{color}.  The patch compiles against the hadoop 
1.1 profile.

{color:green}+1 javadoc{color}.  The javadoc tool did not generate any 
warning messages.

{color:green}+1 javac{color}.  The applied patch does not increase the 
total number of javac compiler warnings.

{color:green}+1 findbugs{color}.  The patch does not introduce any new 
Findbugs (version 1.3.9) warnings.

{color:green}+1 release audit{color}.  The applied patch does not increase 
the total number of release audit warnings.

{color:green}+1 lineLengths{color}.  The patch does not introduce lines 
longer than 100

{color:red}-1 site{color}.  The patch appears to cause mvn site goal to 
fail.

{color:green}+1 core tests{color}.  The patch passed unit tests in .

Test results: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8197//testReport/
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8197//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-protocol.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8197//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-thrift.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8197//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-client.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8197//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-hadoop2-compat.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8197//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-examples.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8197//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-prefix-tree.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8197//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-common.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8197//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-server.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8197//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-hadoop-compat.html
Console output: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8197//console

This message is automatically generated.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, 
 HBASE-9892-trunk-v3.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think 

[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-13 Thread Hadoop QA (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13847555#comment-13847555
 ] 

Hadoop QA commented on HBASE-9892:
--

{color:red}-1 overall{color}.  Here are the results of testing the latest 
attachment 
  
http://issues.apache.org/jira/secure/attachment/12618593/HBASE-9892-trunk-v2.patch
  against trunk revision .

{color:green}+1 @author{color}.  The patch does not contain any @author 
tags.

{color:red}-1 tests included{color}.  The patch doesn't appear to include 
any new or modified tests.
Please justify why no new tests are needed for this 
patch.
Also please list what manual steps were performed to 
verify this patch.

{color:green}+1 hadoop1.0{color}.  The patch compiles against the hadoop 
1.0 profile.

{color:green}+1 hadoop1.1{color}.  The patch compiles against the hadoop 
1.1 profile.

{color:green}+1 javadoc{color}.  The javadoc tool did not generate any 
warning messages.

{color:green}+1 javac{color}.  The applied patch does not increase the 
total number of javac compiler warnings.

{color:green}+1 findbugs{color}.  The patch does not introduce any new 
Findbugs (version 1.3.9) warnings.

{color:green}+1 release audit{color}.  The applied patch does not increase 
the total number of release audit warnings.

{color:green}+1 lineLengths{color}.  The patch does not introduce lines 
longer than 100

{color:red}-1 site{color}.  The patch appears to cause mvn site goal to 
fail.

{color:green}+1 core tests{color}.  The patch passed unit tests in .

Test results: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8156//testReport/
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8156//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-hadoop2-compat.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8156//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-prefix-tree.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8156//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-client.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8156//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-common.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8156//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-protocol.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8156//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-server.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8156//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-examples.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8156//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-thrift.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8156//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-hadoop-compat.html
Console output: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8156//console

This message is automatically generated.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.96.1, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will 

[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-13 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13847663#comment-13847663
 ] 

stack commented on HBASE-9892:
--

+1 from me.  Ok by you [~enis]?

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.96.1, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-trunk-v2.patch, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-09 Thread Enis Soztutar (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13843757#comment-13843757
 ] 

Enis Soztutar commented on HBASE-9892:
--

Left some minor comments to address. Sorry that this is taking so long. Thanks  
Liu for pushing this. 

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.96.1, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-07 Thread Liu Shaohui (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13842168#comment-13842168
 ] 

Liu Shaohui commented on HBASE-9892:


[~stack]

Sorry for the trunk patch makes the test TestMetricsRegionServerSourceImpl 
fail. 

I will find the reason and update the trunk patch later.


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.96.1, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-06 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13841897#comment-13841897
 ] 

stack commented on HBASE-9892:
--

[~enis] You good w/ patch as is?  I am.  Wouldn't mind getting it into 0.96.1.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.96.1, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-06 Thread Hadoop QA (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13841919#comment-13841919
 ] 

Hadoop QA commented on HBASE-9892:
--

{color:red}-1 overall{color}.  Here are the results of testing the latest 
attachment 
  
http://issues.apache.org/jira/secure/attachment/12617503/HBASE-9892-trunk-v1.patch
  against trunk revision .

{color:green}+1 @author{color}.  The patch does not contain any @author 
tags.

{color:red}-1 tests included{color}.  The patch doesn't appear to include 
any new or modified tests.
Please justify why no new tests are needed for this 
patch.
Also please list what manual steps were performed to 
verify this patch.

{color:green}+1 hadoop1.0{color}.  The patch compiles against the hadoop 
1.0 profile.

{color:green}+1 hadoop1.1{color}.  The patch compiles against the hadoop 
1.1 profile.

{color:green}+1 javadoc{color}.  The javadoc tool did not generate any 
warning messages.

{color:green}+1 javac{color}.  The applied patch does not increase the 
total number of javac compiler warnings.

{color:red}-1 findbugs{color}.  The patch appears to introduce 2 new 
Findbugs (version 1.3.9) warnings.

{color:green}+1 release audit{color}.  The applied patch does not increase 
the total number of release audit warnings.

{color:green}+1 lineLengths{color}.  The patch does not introduce lines 
longer than 100

{color:red}-1 site{color}.  The patch appears to cause mvn site goal to 
fail.

 {color:red}-1 core tests{color}.  The patch failed these unit tests:
   
org.apache.hadoop.hbase.regionserver.TestMetricsRegionServerSourceImpl

Test results: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8080//testReport/
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8080//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-protocol.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8080//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-thrift.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8080//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-client.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8080//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-hadoop2-compat.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8080//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-examples.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8080//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-prefix-tree.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8080//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-common.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8080//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-server.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8080//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-hadoop-compat.html
Console output: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8080//console

This message is automatically generated.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.96.1, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,

[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-06 Thread Hadoop QA (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13841971#comment-13841971
 ] 

Hadoop QA commented on HBASE-9892:
--

{color:red}-1 overall{color}.  Here are the results of testing the latest 
attachment 
  
http://issues.apache.org/jira/secure/attachment/12617512/HBASE-9892-trunk-v1.patch
  against trunk revision .

{color:green}+1 @author{color}.  The patch does not contain any @author 
tags.

{color:red}-1 tests included{color}.  The patch doesn't appear to include 
any new or modified tests.
Please justify why no new tests are needed for this 
patch.
Also please list what manual steps were performed to 
verify this patch.

{color:green}+1 hadoop1.0{color}.  The patch compiles against the hadoop 
1.0 profile.

{color:green}+1 hadoop1.1{color}.  The patch compiles against the hadoop 
1.1 profile.

{color:green}+1 javadoc{color}.  The javadoc tool did not generate any 
warning messages.

{color:green}+1 javac{color}.  The applied patch does not increase the 
total number of javac compiler warnings.

{color:red}-1 findbugs{color}.  The patch appears to introduce 2 new 
Findbugs (version 1.3.9) warnings.

{color:green}+1 release audit{color}.  The applied patch does not increase 
the total number of release audit warnings.

{color:green}+1 lineLengths{color}.  The patch does not introduce lines 
longer than 100

{color:red}-1 site{color}.  The patch appears to cause mvn site goal to 
fail.

 {color:red}-1 core tests{color}.  The patch failed these unit tests:
   
org.apache.hadoop.hbase.regionserver.TestMetricsRegionServerSourceImpl

Test results: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8082//testReport/
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8082//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-protocol.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8082//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-thrift.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8082//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-client.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8082//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-hadoop2-compat.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8082//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-examples.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8082//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-prefix-tree.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8082//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-common.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8082//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-server.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8082//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-hadoop-compat.html
Console output: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8082//console

This message is automatically generated.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Fix For: 0.98.0, 0.96.1, 0.99.0

 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-trunk-v1.patch, 
 HBASE-9892-trunk-v1.patch, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we 

[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-05 Thread Hadoop QA (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13841019#comment-13841019
 ] 

Hadoop QA commented on HBASE-9892:
--

{color:red}-1 overall{color}.  Here are the results of testing the latest 
attachment 
  
http://issues.apache.org/jira/secure/attachment/12617325/HBASE-9892-trunk-v1.diff
  against trunk revision .

{color:green}+1 @author{color}.  The patch does not contain any @author 
tags.

{color:red}-1 tests included{color}.  The patch doesn't appear to include 
any new or modified tests.
Please justify why no new tests are needed for this 
patch.
Also please list what manual steps were performed to 
verify this patch.

{color:green}+1 hadoop1.0{color}.  The patch compiles against the hadoop 
1.0 profile.

{color:green}+1 hadoop1.1{color}.  The patch compiles against the hadoop 
1.1 profile.

{color:red}-1 javadoc{color}.  The javadoc tool appears to have generated 1 
warning messages.

{color:green}+1 javac{color}.  The applied patch does not increase the 
total number of javac compiler warnings.

{color:red}-1 findbugs{color}.  The patch appears to introduce 2 new 
Findbugs (version 1.3.9) warnings.

{color:green}+1 release audit{color}.  The applied patch does not increase 
the total number of release audit warnings.

{color:green}+1 lineLengths{color}.  The patch does not introduce lines 
longer than 100

{color:red}-1 site{color}.  The patch appears to cause mvn site goal to 
fail.

 {color:red}-1 core tests{color}.  The patch failed these unit tests:
   
org.apache.hadoop.hbase.regionserver.TestMetricsRegionServerSourceImpl

Test results: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8071//testReport/
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8071//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-hadoop2-compat.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8071//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-prefix-tree.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8071//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-client.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8071//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-common.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8071//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-protocol.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8071//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-server.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8071//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-examples.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8071//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-thrift.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8071//artifact/trunk/patchprocess/newPatchFindbugsWarningshbase-hadoop-compat.html
Console output: 
https://builds.apache.org/job/PreCommit-HBASE-Build/8071//console

This message is automatically generated.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff, HBASE-9892-0.94-v5.diff, 
 HBASE-9892-trunk-v1.diff, HBASE-9892-v5.txt


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  

[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-12-02 Thread Enis Soztutar (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13836989#comment-13836989
 ] 

Enis Soztutar commented on HBASE-9892:
--

bq. info port is an attribute that does not vary while server is up and 
running. It does not belong in the cluster status set. Can you not get it from 
zk Steve Loughran? Or failing that, you want us to write you a script that will 
read it out of zk for you – this and stuff like current master?
We still need to make the info ports and maybe other static attributes for the 
cluster to the clients as an API I think. It belongs together with liveServers 
data, since for each live server, the client may need the info port. 

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-28 Thread Liu Shaohui (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13834752#comment-13834752
 ] 

Liu Shaohui commented on HBASE-9892:


[~stack] [~enis]
Please help to review the patch for 0.94 again?

If there is no other problem, I will start to do the patch for trunk. Happy 
Thanksgiving!

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-27 Thread Steve Loughran (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13833639#comment-13833639
 ] 

Steve Loughran commented on HBASE-9892:
---

I had a quick look and while the intricacies of HBase code escape me , it looks 
like the Masters get the port info via ZK. Does this propagate as far as the 
hbase status data you get with {{HBaseAdmin.getClusterStatus()}}? That's where 
I need to pick it up from -along with the infoserver port of the master itself

-steve

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-27 Thread Liu Shaohui (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13833685#comment-13833685
 ] 

Liu Shaohui commented on HBASE-9892:


[~ste...@apache.org] 

After checking the code of HBaseAdmin.getClusterStatus(), it seems that it does 
not return the info ports of regionservers. 
 


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-27 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13834378#comment-13834378
 ] 

stack commented on HBASE-9892:
--

info port is an attribute that does not vary while server is up and running.  
It does not belong in the cluster status set.  Can you not get it from zk 
[~stev...@iseran.com]?  Or failing that, you want us to write you a script that 
will read it out of zk for you -- this and stuff like current master?

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff, HBASE-9892-0.94-v4.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-11 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13819516#comment-13819516
 ] 

stack commented on HBASE-9892:
--

[~liushaohui] What do you think of the [~enis] comment above regards patch for 
0.94?

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-07 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13816208#comment-13816208
 ] 

stack commented on HBASE-9892:
--

[~liushaohui] For trunk, removal of HBASE-7027 and replacing it with a pb'd and 
trunk version of the attached patch would be best going forward.  For 0.94, 
yes, ServerLoad is versioned but changing it you will be into a thorny area for 
you will have to be able to demonstrate you have not for sure broken rolling 
updates.  Thanks for working on this ugly issue.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-07 Thread Enis Soztutar (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13816320#comment-13816320
 ] 

Enis Soztutar commented on HBASE-9892:
--

bq. Since HServerLoad have verison field, I think we can add a info port field 
and keep the compatibility.
The problem with the versioned writables is that, on the serializing front, you 
have to know the version of the deserializing side. For example, in a rolling 
restart scenario, even if you version the HServerLoad, the master will throw 
exceptions if it has not updated yet because it will not know about the new 
version number. 
bq. For trunk, removal of HBASE-7027 and replacing it with a pb'd and trunk 
version of the attached patch would be best going forward.
Agreed. Let's forward port this patch to trunk and go with it. We have to PB 
the RegionServerInfo object in the RB patch. Undoing HBASE-7027 can be done in 
a subtask after this is committed. 

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-07 Thread Liu Shaohui (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13816927#comment-13816927
 ] 

Liu Shaohui commented on HBASE-9892:


Thanks. [~stack] [~enish]
Please help to review patch for 0.94 in RB again.
If there is no  problems, I will start to make it to trunk.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-06 Thread Liu Shaohui (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13814752#comment-13814752
 ] 

Liu Shaohui commented on HBASE-9892:


Upload a new patch for [~enis]'s review. 
If this patch is ok, i will start to make a trunk patch. 

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-06 Thread Steve Loughran (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13815174#comment-13815174
 ] 

Steve Loughran commented on HBASE-9892:
---

As Enis says, currently we know the problem is there but don't try to fix it. 

The issue we have there is not just if YARN assigns 1 region server to the 
same node (it doesn't currently support anti-affinity in allocation requests), 
but that someone else may be running their own application, HBase or otherwise, 
on the same machine. If you hard code a port it can fail -any port. The sole 
advantage we have is that will trigger a new container request/review

Because this also affects the masters, we have to leave that UI at port 0 too 
-which is the worst issue. I would really like to get hold of that via ZK, from 
where we can bootstrap the rest of the cluster information 

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-06 Thread Nick Dimiduk (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13815502#comment-13815502
 ] 

Nick Dimiduk commented on HBASE-9892:
-

Hi [~liushaohui]

I cannot reproduce the problem you describe on 0.96 or trunk. There it appears 
the info port comes from ServerName or ServerLoad objects for live server link 
or dead server list, respectively. Could we not backport an existing fix rather 
than add all this new plumbing?

For context, I have multiple RS processes deployed on each RS host. Their RPC 
and info ports are set explicitly for each process. Perhaps your configuration 
is different?

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-06 Thread Enis Soztutar (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13815541#comment-13815541
 ] 

Enis Soztutar commented on HBASE-9892:
--

If we do not have the problem in 0.96, I would be -0 for fixing it in 0.94. 

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-06 Thread Liu Shaohui (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13815575#comment-13815575
 ] 

Liu Shaohui commented on HBASE-9892:


Thanks [~ndimiduk][~enis]
HBASE-7027 partly have fixed the problem by report the info port to hmaster via 
serverLoad indead. 

But the  infoPort in table.jps still be got from config.
{code}
  // HARDCODED FOR NOW TODO: FIX GET FROM ZK
  // This port might be wrong if RS actually ended up using something else.
  int infoPort = conf.getInt(hbase.regionserver.info.port, 60030);
{code}

What about backporting HBASE-7027 to 0.94 and fixing the issue in table.jsp? 


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-06 Thread Enis Soztutar (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13815583#comment-13815583
 ] 

Enis Soztutar commented on HBASE-9892:
--

bq. What about backporting HBASE-7027 to 0.94 and fixing the issue in table.jsp?
It makes sense, but the challenge is that we cannot easily backport HBASE-7027 
without breaking BC. HServerLoad does not have extra fields we can use i fear. 

Your patch at RB is actually better than 7027, if we do this for trunk, we 
should undo 7027. 




 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-06 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13815594#comment-13815594
 ] 

stack commented on HBASE-9892:
--

HBASE-7027 is an admitted hack; would be good to do it better.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-06 Thread Liu Shaohui (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13815619#comment-13815619
 ] 

Liu Shaohui commented on HBASE-9892:


[~enis]
{quote}
What about backporting HBASE-7027 to 0.94 and fixing the issue in table.jsp?

It makes sense, but the challenge is that we cannot easily backport HBASE-7027 
without breaking BC. HServerLoad does not have extra fields we can use i fear. 
{quote}
Since HServerLoad have verison field, I think we can add a info port field and 
keep the compatibility.

[~stack] [~enis] Could you give a suggestion? Which method is acceptable? 
This patch or backporting 7027 and fixing the small issues left.


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff, 
 HBASE-9892-0.94-v3.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-05 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13813739#comment-13813739
 ] 

stack commented on HBASE-9892:
--

bq. Now, there is no data in regionserver's ephemeral node. It's a good idea to 
write static attributes like info port there. 

It is a suggestion.

Could be tricky setting this value then triggering watches.  Will have to reset 
them.

Maybe znode is not the right place?  It is too awkward and if only this one 
attribute, its a bit of work adding it there.

It could be added to the server JMX bean but you'd have to do rmi to find it 
which requires a port (IIRC)

There is the RS heartbeat.  Currently we send load.  Seems a bit silly sending 
over constant attributes on each heartbeat but might be easy to do.

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-05 Thread Liu Shaohui (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13813753#comment-13813753
 ] 

Liu Shaohui commented on HBASE-9892:


{quote}
Could be tricky setting this value then triggering watches. Will have to reset 
them.
{quote}
No need to reset them. RegionServerTracker only get data from zk once.

{quote}
It could be added to the server JMX bean but you'd have to do rmi to find it 
which requires a port (IIRC)
There is the RS heartbeat. Currently we send load. Seems a bit silly sending 
over constant attributes on each heartbeat but might be easy to do.
{quote}
I think the info port of a rs will not change after it starts up, so no need to 
send over constant attributes on each heartbeat. 


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-05 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13813991#comment-13813991
 ] 

stack commented on HBASE-9892:
--

Patch looks fine.

Did you intend to include this in the patch?

Index: src/main/java/org/apache/hadoop/hbase/regionserver/RSDumpServlet.java

... and this?

Index: src/main/java/org/apache/hadoop/hbase/regionserver/RSStatusServlet.java

Should this be public?  Can it be package protected?

getRegionServerInfoPort

Can we write the znode content as protobuf?  Will make it easier adding new 
attributes and in trunk, all znodes are pb:

+String nodePath = ZKUtil.joinZNode(watcher.rsZNode, n);
+infoPort = Bytes.toInt(ZKUtil.getData(watcher, nodePath));

If you need help, I can help do the trunk patch np.

Good stuff.



 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-05 Thread Enis Soztutar (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13814285#comment-13814285
 ] 

Enis Soztutar commented on HBASE-9892:
--

Great. Left some comments at RB. 

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-05 Thread Liu Shaohui (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13814674#comment-13814674
 ] 

Liu Shaohui commented on HBASE-9892:


[~stack]
{quote}
Did you intend to include this in the patch?
Index: src/main/java/org/apache/hadoop/hbase/regionserver/RSDumpServlet.java
... and this?
Index: src/main/java/org/apache/hadoop/hbase/regionserver/RSStatusServlet.java
Should this be public? Can it be package protected?
{quote}
I did not get your meaning. All these http Servlets are public. I think no need 
to change it to package protected.

{quote}
Can we write the znode content as protobuf? Will make it easier adding new 
attributes and in trunk, all znodes are pb:
+ String nodePath = ZKUtil.joinZNode(watcher.rsZNode, n);
+ infoPort = Bytes.toInt(ZKUtil.getData(watcher, nodePath));
{quote}
Yes. I will write a writable class: RegionServerInfo for 0.94 and pb for trunk. 
[~enis]


 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff, HBASE-9892-0.94-v2.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-04 Thread Enis Soztutar (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13813605#comment-13813605
 ] 

Enis Soztutar commented on HBASE-9892:
--

We don't have to hold the info port as a part of ServerName I think. ServerName 
is for unique identifying the RS process. We can keep the info port as reported 
by the RS in ServerManager, and the web UI can get the port by ServerName from 
there. 
Hoya, right now, does not start multiple region servers of the same host for 
the same cluster I believe. 

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-04 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13813646#comment-13813646
 ] 

stack commented on HBASE-9892:
--

[~liushaohui] I commented similar to Enis over in RB.

The issue w/ RS reporting its info port only when it registers is that should 
the master fail, how does the new master figure what the info port for all 
running RSs are?

Is the info port availble in jmx server attributes?  If not, could add there?  
Or RS could write static attributes like info port into the body of its 
ephemeral znode?

There are different formats in trunk for SN in znode and meta?

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-04 Thread Liu Shaohui (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13813649#comment-13813649
 ] 

Liu Shaohui commented on HBASE-9892:


[~enis] Good idea.

I will try to add an api : getRegionServerInfoPort(ServerName) in HMaster for 
get the port by ServerName from ServerManager. 

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)


[jira] [Commented] (HBASE-9892) Add info port to ServerName to support multi instances in a node

2013-11-04 Thread Liu Shaohui (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-9892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13813688#comment-13813688
 ] 

Liu Shaohui commented on HBASE-9892:


[~stack]

{quote}
The issue w/ RS reporting its info port only when it registers is that should 
the master fail, how does the new master figure what the info port for all 
running RSs are?
Is the info port availble in jmx server attributes? If not, could add there? Or 
RS could write static attributes like info port into the body of its ephemeral 
znode?
{quote}

Thanks [~stack]. I did not notice this problem.
Now, there is no data in regionserver's ephemeral node. It's a good idea to 
write static attributes like info port there. 
What do you think? [~enis] 


{quote}
There are different formats in trunk for SN in znode and meta?
{quote}
For meta, there are two columns: server(ip:port) and serverstartcode for SN.
For root-region-server zk node,  the format is ip,port,serverstartcode

 Add info port to ServerName to support multi instances in a node
 

 Key: HBASE-9892
 URL: https://issues.apache.org/jira/browse/HBASE-9892
 Project: HBase
  Issue Type: Improvement
Reporter: Liu Shaohui
Assignee: Liu Shaohui
Priority: Minor
 Attachments: HBASE-9892-0.94-v1.diff


 The full GC time of  regionserver with big heap( 30G ) usually  can not be 
 controlled in 30s. At the same time, the servers with 64G memory are normal. 
 So we try to deploy multi rs instances(2-3 ) in a single node and the heap of 
 each rs is about 20G ~ 24G.
 Most of the things works fine, except the hbase web ui. The master get the RS 
 info port from conf, which is suitable for this situation of multi rs  
 instances in a node. So we add info port to ServerName.
 a. at the startup, rs report it's info port to Hmaster.
 b, For root region, rs write the servername with info port ro the zookeeper 
 root-region-server node.
 c, For meta regions, rs write the servername with info port to root region 
 d. For user regions,  rs write the servername with info port to meta regions 
 So hmaster and client can get info port from the servername.
 To test this feature, I change the rs num from 1 to 3 in standalone mode, so 
 we can test it in standalone mode,
 I think Hoya(hbase on yarn) will encounter the same problem.  Anyone knows 
 how Hoya handle this problem?
 PS: There are  different formats for servername in zk node and meta table, i 
 think we need to unify it and refactor the code.



--
This message was sent by Atlassian JIRA
(v6.1#6144)