[jira] [Updated] (HBASE-18103) [AMv2] If Master gives OPEN to another, if original eventually succeeds, Master will kill it

2017-08-21 Thread stack (JIRA)

 [ 
https://issues.apache.org/jira/browse/HBASE-18103?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

stack updated HBASE-18103:
--
  Resolution: Fixed
Hadoop Flags: Reviewed
  Status: Resolved  (was: Patch Available)

Pushed to master and branch-2. Thanks for the patch [~uagashe]

> [AMv2] If Master gives OPEN to another, if original eventually succeeds, 
> Master will kill it
> 
>
> Key: HBASE-18103
> URL: https://issues.apache.org/jira/browse/HBASE-18103
> Project: HBase
>  Issue Type: Sub-task
>  Components: master, proc-v2
>Reporter: stack
>Assignee: Umesh Agashe
>Priority: Critical
> Fix For: 2.0.0
>
> Attachments: hbase-18103.master.001.patch, 
> hbase-18103.master.001.patch
>
>
> If a RS is slow to open a Region, the Master will give the Region to another 
> to open it (In this case, was a massive set of edits to process and a load of 
> StoreFiles to open...). Should the original RS succeed with its open 
> eventually, on reporting the master the successful open, the Master currently 
> kills the RS because the region is supposed to be elsewhere.
> This is an easy fix.
> The RS does not fully open a Region until Master gives it the go so just 
> close the region if master rejects the open
> See '6.1.1 If Master gives Region to another to Open, old RS will be kill 
> itself on reject by Master; easy fix!' in 
> https://docs.google.com/document/d/1eVKa7FHdeoJ1-9o8yZcOTAQbv0u0bblBlCCzVSIn69g/edit#heading=h.qtfojp9774h



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)


[jira] [Updated] (HBASE-18103) [AMv2] If Master gives OPEN to another, if original eventually succeeds, Master will kill it

2017-08-21 Thread Umesh Agashe (JIRA)

 [ 
https://issues.apache.org/jira/browse/HBASE-18103?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Umesh Agashe updated HBASE-18103:
-
Attachment: hbase-18103.master.001.patch

retry

> [AMv2] If Master gives OPEN to another, if original eventually succeeds, 
> Master will kill it
> 
>
> Key: HBASE-18103
> URL: https://issues.apache.org/jira/browse/HBASE-18103
> Project: HBase
>  Issue Type: Sub-task
>  Components: master, proc-v2
>Reporter: stack
>Assignee: Umesh Agashe
>Priority: Critical
> Fix For: 2.0.0
>
> Attachments: hbase-18103.master.001.patch, 
> hbase-18103.master.001.patch
>
>
> If a RS is slow to open a Region, the Master will give the Region to another 
> to open it (In this case, was a massive set of edits to process and a load of 
> StoreFiles to open...). Should the original RS succeed with its open 
> eventually, on reporting the master the successful open, the Master currently 
> kills the RS because the region is supposed to be elsewhere.
> This is an easy fix.
> The RS does not fully open a Region until Master gives it the go so just 
> close the region if master rejects the open
> See '6.1.1 If Master gives Region to another to Open, old RS will be kill 
> itself on reject by Master; easy fix!' in 
> https://docs.google.com/document/d/1eVKa7FHdeoJ1-9o8yZcOTAQbv0u0bblBlCCzVSIn69g/edit#heading=h.qtfojp9774h



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)


[jira] [Updated] (HBASE-18103) [AMv2] If Master gives OPEN to another, if original eventually succeeds, Master will kill it

2017-08-21 Thread Umesh Agashe (JIRA)

 [ 
https://issues.apache.org/jira/browse/HBASE-18103?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Umesh Agashe updated HBASE-18103:
-
Attachment: hbase-18103.master.001.patch

> [AMv2] If Master gives OPEN to another, if original eventually succeeds, 
> Master will kill it
> 
>
> Key: HBASE-18103
> URL: https://issues.apache.org/jira/browse/HBASE-18103
> Project: HBase
>  Issue Type: Sub-task
>  Components: master, proc-v2
>Reporter: stack
>Assignee: Umesh Agashe
>Priority: Critical
> Fix For: 2.0.0
>
> Attachments: hbase-18103.master.001.patch
>
>
> If a RS is slow to open a Region, the Master will give the Region to another 
> to open it (In this case, was a massive set of edits to process and a load of 
> StoreFiles to open...). Should the original RS succeed with its open 
> eventually, on reporting the master the successful open, the Master currently 
> kills the RS because the region is supposed to be elsewhere.
> This is an easy fix.
> The RS does not fully open a Region until Master gives it the go so just 
> close the region if master rejects the open
> See '6.1.1 If Master gives Region to another to Open, old RS will be kill 
> itself on reject by Master; easy fix!' in 
> https://docs.google.com/document/d/1eVKa7FHdeoJ1-9o8yZcOTAQbv0u0bblBlCCzVSIn69g/edit#heading=h.qtfojp9774h



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)


[jira] [Updated] (HBASE-18103) [AMv2] If Master gives OPEN to another, if original eventually succeeds, Master will kill it

2017-08-21 Thread Umesh Agashe (JIRA)

 [ 
https://issues.apache.org/jira/browse/HBASE-18103?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Umesh Agashe updated HBASE-18103:
-
Status: Patch Available  (was: In Progress)

> [AMv2] If Master gives OPEN to another, if original eventually succeeds, 
> Master will kill it
> 
>
> Key: HBASE-18103
> URL: https://issues.apache.org/jira/browse/HBASE-18103
> Project: HBase
>  Issue Type: Sub-task
>  Components: master, proc-v2
>Reporter: stack
>Assignee: Umesh Agashe
>Priority: Critical
> Fix For: 2.0.0
>
> Attachments: hbase-18103.master.001.patch
>
>
> If a RS is slow to open a Region, the Master will give the Region to another 
> to open it (In this case, was a massive set of edits to process and a load of 
> StoreFiles to open...). Should the original RS succeed with its open 
> eventually, on reporting the master the successful open, the Master currently 
> kills the RS because the region is supposed to be elsewhere.
> This is an easy fix.
> The RS does not fully open a Region until Master gives it the go so just 
> close the region if master rejects the open
> See '6.1.1 If Master gives Region to another to Open, old RS will be kill 
> itself on reject by Master; easy fix!' in 
> https://docs.google.com/document/d/1eVKa7FHdeoJ1-9o8yZcOTAQbv0u0bblBlCCzVSIn69g/edit#heading=h.qtfojp9774h



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)


[jira] [Updated] (HBASE-18103) [AMv2] If Master gives OPEN to another, if original eventually succeeds, Master will kill it

2017-05-26 Thread stack (JIRA)

 [ 
https://issues.apache.org/jira/browse/HBASE-18103?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

stack updated HBASE-18103:
--
Priority: Critical  (was: Major)

> [AMv2] If Master gives OPEN to another, if original eventually succeeds, 
> Master will kill it
> 
>
> Key: HBASE-18103
> URL: https://issues.apache.org/jira/browse/HBASE-18103
> Project: HBase
>  Issue Type: Sub-task
>  Components: master, proc-v2
>Reporter: stack
>Priority: Critical
> Fix For: 2.0.0
>
>
> If a RS is slow to open a Region, the Master will give the Region to another 
> to open it (In this case, was a massive set of edits to process and a load of 
> StoreFiles to open...). Should the original RS succeed with its open 
> eventually, on reporting the master the successful open, the Master currently 
> kills the RS because the region is supposed to be elsewhere.
> This is an easy fix.
> The RS does not fully open a Region until Master gives it the go so just 
> close the region if master rejects the open
> See '6.1.1 If Master gives Region to another to Open, old RS will be kill 
> itself on reject by Master; easy fix!' in 
> https://docs.google.com/document/d/1eVKa7FHdeoJ1-9o8yZcOTAQbv0u0bblBlCCzVSIn69g/edit#heading=h.qtfojp9774h



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)