[jira] [Commented] (HBASE-7092) RegionServer OOM related to onlineregions
[ https://issues.apache.org/jira/browse/HBASE-7092?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13491712#comment-13491712 ] stack commented on HBASE-7092: -- How many regions per server? 4G or RAM? I should add a note to important configs. in refguide Jimmy? In 0.96 slab is on by default. RegionServer OOM related to onlineregions - Key: HBASE-7092 URL: https://issues.apache.org/jira/browse/HBASE-7092 Project: HBase Issue Type: Bug Reporter: Jimmy Xiang Assignee: Jimmy Xiang Priority: Blocker Fix For: 0.96.0 Attachments: heap-dump.png One instance of java.util.concurrent.ConcurrentHashMap loaded by system class loader occupies 3,972,154,848 (92.88%) bytes. The instance is referenced by org.apache.hadoop.hbase.regionserver.HRegionServer @ 0x7038d3798 , loaded by sun.misc.Launcher$AppClassLoader @ 0x703994668. The memory is accumulated in one instance of java.util.concurrent.ConcurrentHashMap$Segment[] loaded by system class loader. Keywords sun.misc.Launcher$AppClassLoader @ 0x703994668 java.util.concurrent.ConcurrentHashMap java.util.concurrent.ConcurrentHashMap$Segment[] -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HBASE-7092) RegionServer OOM related to onlineregions
[ https://issues.apache.org/jira/browse/HBASE-7092?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13491715#comment-13491715 ] Jimmy Xiang commented on HBASE-7092: I have about 2600 regions per server, and just 4G max heap. Slab is on by default in hbase-default.xml, not the code. RegionServer OOM related to onlineregions - Key: HBASE-7092 URL: https://issues.apache.org/jira/browse/HBASE-7092 Project: HBase Issue Type: Bug Reporter: Jimmy Xiang Assignee: Jimmy Xiang Priority: Blocker Fix For: 0.96.0 Attachments: heap-dump.png One instance of java.util.concurrent.ConcurrentHashMap loaded by system class loader occupies 3,972,154,848 (92.88%) bytes. The instance is referenced by org.apache.hadoop.hbase.regionserver.HRegionServer @ 0x7038d3798 , loaded by sun.misc.Launcher$AppClassLoader @ 0x703994668. The memory is accumulated in one instance of java.util.concurrent.ConcurrentHashMap$Segment[] loaded by system class loader. Keywords sun.misc.Launcher$AppClassLoader @ 0x703994668 java.util.concurrent.ConcurrentHashMap java.util.concurrent.ConcurrentHashMap$Segment[] -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HBASE-7092) RegionServer OOM related to onlineregions
[ https://issues.apache.org/jira/browse/HBASE-7092?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13491786#comment-13491786 ] stack commented on HBASE-7092: -- I added a note to the MSLAB section that lots of regions could cause you OOME (MSLAB is on since 0.92, not 0.96 as I said above) RegionServer OOM related to onlineregions - Key: HBASE-7092 URL: https://issues.apache.org/jira/browse/HBASE-7092 Project: HBase Issue Type: Bug Reporter: Jimmy Xiang Assignee: Jimmy Xiang Priority: Blocker Fix For: 0.96.0 Attachments: heap-dump.png One instance of java.util.concurrent.ConcurrentHashMap loaded by system class loader occupies 3,972,154,848 (92.88%) bytes. The instance is referenced by org.apache.hadoop.hbase.regionserver.HRegionServer @ 0x7038d3798 , loaded by sun.misc.Launcher$AppClassLoader @ 0x703994668. The memory is accumulated in one instance of java.util.concurrent.ConcurrentHashMap$Segment[] loaded by system class loader. Keywords sun.misc.Launcher$AppClassLoader @ 0x703994668 java.util.concurrent.ConcurrentHashMap java.util.concurrent.ConcurrentHashMap$Segment[] -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HBASE-7092) RegionServer OOM related to onlineregions
[ https://issues.apache.org/jira/browse/HBASE-7092?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13490810#comment-13490810 ] Jimmy Xiang commented on HBASE-7092: I got the same issue with jdk 1.6. So it is not a jdk issue. Uploaded the head-dump image shows it may related to onlineregions. RegionServer OOM related to onlineregions - Key: HBASE-7092 URL: https://issues.apache.org/jira/browse/HBASE-7092 Project: HBase Issue Type: Bug Reporter: Jimmy Xiang Assignee: Jimmy Xiang Fix For: 0.96.0 Attachments: heap-dump.png One instance of java.util.concurrent.ConcurrentHashMap loaded by system class loader occupies 3,972,154,848 (92.88%) bytes. The instance is referenced by org.apache.hadoop.hbase.regionserver.HRegionServer @ 0x7038d3798 , loaded by sun.misc.Launcher$AppClassLoader @ 0x703994668. The memory is accumulated in one instance of java.util.concurrent.ConcurrentHashMap$Segment[] loaded by system class loader. Keywords sun.misc.Launcher$AppClassLoader @ 0x703994668 java.util.concurrent.ConcurrentHashMap java.util.concurrent.ConcurrentHashMap$Segment[] -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HBASE-7092) RegionServer OOM related to onlineregions
[ https://issues.apache.org/jira/browse/HBASE-7092?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13490915#comment-13490915 ] Jimmy Xiang commented on HBASE-7092: Actually, it is not block cache. It is memstore buffer. RegionServer OOM related to onlineregions - Key: HBASE-7092 URL: https://issues.apache.org/jira/browse/HBASE-7092 Project: HBase Issue Type: Bug Reporter: Jimmy Xiang Assignee: Jimmy Xiang Priority: Blocker Fix For: 0.96.0 Attachments: heap-dump.png One instance of java.util.concurrent.ConcurrentHashMap loaded by system class loader occupies 3,972,154,848 (92.88%) bytes. The instance is referenced by org.apache.hadoop.hbase.regionserver.HRegionServer @ 0x7038d3798 , loaded by sun.misc.Launcher$AppClassLoader @ 0x703994668. The memory is accumulated in one instance of java.util.concurrent.ConcurrentHashMap$Segment[] loaded by system class loader. Keywords sun.misc.Launcher$AppClassLoader @ 0x703994668 java.util.concurrent.ConcurrentHashMap java.util.concurrent.ConcurrentHashMap$Segment[] -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
