[jira] [Updated] (HBASE-20144) The shutdown of master will hang if there are no live region server
[ https://issues.apache.org/jira/browse/HBASE-20144?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Duo Zhang updated HBASE-20144: -- Resolution: Fixed Hadoop Flags: Reviewed Status: Resolved (was: Patch Available) Pushed to branch-2.0+. Thanks [~stack] for reviewing. Keep HBASE-19598 open to see if it works. > The shutdown of master will hang if there are no live region server > --- > > Key: HBASE-20144 > URL: https://issues.apache.org/jira/browse/HBASE-20144 > Project: HBase > Issue Type: Bug > Components: Recovery >Reporter: Duo Zhang >Assignee: Duo Zhang >Priority: Major > Fix For: 2.0.0 > > Attachments: HBASE-20144-UT.patch, HBASE-20144-v1.patch, > HBASE-20144.patch > > > In a graceful shutdown, we will set cluster down in ServerManager, and in > ServerManger.expireServer, if the cluster down flag is true and all rses are > gone, we will call master.stop and quit. > But if we have no live rs when shutting down, then expireServer will never be > triggered, so we are dead... -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Updated] (HBASE-20144) The shutdown of master will hang if there are no live region server
[ https://issues.apache.org/jira/browse/HBASE-20144?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Duo Zhang updated HBASE-20144: -- Component/s: Recovery > The shutdown of master will hang if there are no live region server > --- > > Key: HBASE-20144 > URL: https://issues.apache.org/jira/browse/HBASE-20144 > Project: HBase > Issue Type: Bug > Components: Recovery >Reporter: Duo Zhang >Assignee: Duo Zhang >Priority: Major > Fix For: 2.0.0 > > Attachments: HBASE-20144-UT.patch, HBASE-20144-v1.patch, > HBASE-20144.patch > > > In a graceful shutdown, we will set cluster down in ServerManager, and in > ServerManger.expireServer, if the cluster down flag is true and all rses are > gone, we will call master.stop and quit. > But if we have no live rs when shutting down, then expireServer will never be > triggered, so we are dead... -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Updated] (HBASE-20144) The shutdown of master will hang if there are no live region server
[ https://issues.apache.org/jira/browse/HBASE-20144?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Duo Zhang updated HBASE-20144: -- Attachment: HBASE-20144-v1.patch > The shutdown of master will hang if there are no live region server > --- > > Key: HBASE-20144 > URL: https://issues.apache.org/jira/browse/HBASE-20144 > Project: HBase > Issue Type: Bug >Reporter: Duo Zhang >Assignee: Duo Zhang >Priority: Major > Fix For: 2.0.0 > > Attachments: HBASE-20144-UT.patch, HBASE-20144-v1.patch, > HBASE-20144.patch > > > In a graceful shutdown, we will set cluster down in ServerManager, and in > ServerManger.expireServer, if the cluster down flag is true and all rses are > gone, we will call master.stop and quit. > But if we have no live rs when shutting down, then expireServer will never be > triggered, so we are dead... -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Updated] (HBASE-20144) The shutdown of master will hang if there are no live region server
[ https://issues.apache.org/jira/browse/HBASE-20144?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Duo Zhang updated HBASE-20144: -- Assignee: Duo Zhang Fix Version/s: 2.0.0 Status: Patch Available (was: Open) Call HMaster.stop directly if no online servers in ServerManager.shutdownCluster. Also enable TestAssignmentManagerMetrics to see if it is fine after this fix. > The shutdown of master will hang if there are no live region server > --- > > Key: HBASE-20144 > URL: https://issues.apache.org/jira/browse/HBASE-20144 > Project: HBase > Issue Type: Bug >Reporter: Duo Zhang >Assignee: Duo Zhang >Priority: Major > Fix For: 2.0.0 > > Attachments: HBASE-20144-UT.patch, HBASE-20144.patch > > > In a graceful shutdown, we will set cluster down in ServerManager, and in > ServerManger.expireServer, if the cluster down flag is true and all rses are > gone, we will call master.stop and quit. > But if we have no live rs when shutting down, then expireServer will never be > triggered, so we are dead... -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Updated] (HBASE-20144) The shutdown of master will hang if there are no live region server
[ https://issues.apache.org/jira/browse/HBASE-20144?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Duo Zhang updated HBASE-20144: -- Attachment: HBASE-20144.patch > The shutdown of master will hang if there are no live region server > --- > > Key: HBASE-20144 > URL: https://issues.apache.org/jira/browse/HBASE-20144 > Project: HBase > Issue Type: Bug >Reporter: Duo Zhang >Priority: Major > Attachments: HBASE-20144-UT.patch, HBASE-20144.patch > > > In a graceful shutdown, we will set cluster down in ServerManager, and in > ServerManger.expireServer, if the cluster down flag is true and all rses are > gone, we will call master.stop and quit. > But if we have no live rs when shutting down, then expireServer will never be > triggered, so we are dead... -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Updated] (HBASE-20144) The shutdown of master will hang if there are no live region server
[ https://issues.apache.org/jira/browse/HBASE-20144?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Duo Zhang updated HBASE-20144: -- Attachment: HBASE-20144-UT.patch > The shutdown of master will hang if there are no live region server > --- > > Key: HBASE-20144 > URL: https://issues.apache.org/jira/browse/HBASE-20144 > Project: HBase > Issue Type: Bug >Reporter: Duo Zhang >Priority: Major > Attachments: HBASE-20144-UT.patch > > > In a graceful shutdown, we will set cluster down in ServerManager, and in > ServerManger.expireServer, if the cluster down flag is true and all rses are > gone, we will call master.stop and quit. > But if we have no live rs when shutting down, then expireServer will never be > triggered, so we are dead... -- This message was sent by Atlassian JIRA (v7.6.3#76005)
