You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hbase.apache.org by "Andrew Purtell (JIRA)" <ji...@apache.org> on 2009/07/31 21:40:15 UTC
[jira] Created: (HBASE-1730) online schema updates
online schema updates
---------------------
Key: HBASE-1730
URL: https://issues.apache.org/jira/browse/HBASE-1730
Project: Hadoop HBase
Issue Type: Improvement
Reporter: Andrew Purtell
Fix For: 0.21.0
We should not need to take a table offline to update HCD or HTD.
One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
{{/hbase/tables/<table-name>/schema}}
Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-1730) Near-instantaneous online schema and
table state updates
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12771760#action_12771760 ]
stack commented on HBASE-1730:
------------------------------
@apurtell I tried putting jackson 1.0.1 up in hbase/lib dir but doesn't look like stargate picks it up from there (though looking at stargate build.xml, it should -- maybe it dont like jackson-1.0.1).
> Near-instantaneous online schema and table state updates
> --------------------------------------------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Assignee: stack
> Fix For: 0.21.0
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-1730) online schema and tableAdd updates
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12771706#action_12771706 ]
stack commented on HBASE-1730:
------------------------------
For sure @ryan. Thats what I'm about: schema and state changes that are near-instantaneous.
> online schema and tableAdd updates
> ----------------------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Assignee: stack
> Fix For: 0.21.0
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HBASE-1730) Near-instantaneous online schema and
table state updates
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
stack updated HBASE-1730:
-------------------------
Attachment: 1730-v2.patch
More work on HCD. Removes UnmodifyableHCD. Since HCD now implements map, can just do a Collections.unmodifyableMap wrapper on delegate Map. Jackson is sweet in that for new HCD.toString, I ask jackson to serialize to a StringReader. It produces something like a ruby Map -- e..g. {"NAME": "cf1", "IN_MEMORY": true, "COMPRESSION", "LZO"} -- only I can then take this 'String' and deserialize to get me back the original HCD. Nice for shell, UI, and content of a znode (I don't think we need to get toBinaryString in there since I don't think we have binary data in schema and table state). Working on HTD next.
> Near-instantaneous online schema and table state updates
> --------------------------------------------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Assignee: stack
> Fix For: 0.21.0
>
> Attachments: 1730-v2.patch, 1730.patch
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Assigned: (HBASE-1730) online schema updates
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
stack reassigned HBASE-1730:
----------------------------
Assignee: stack
> online schema updates
> ---------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Assignee: stack
> Fix For: 0.21.0
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-1730) online schema updates
Posted by "ryan rawson (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12737688#action_12737688 ]
ryan rawson commented on HBASE-1730:
------------------------------------
lets decide on a serialization format to use within ZK, that is NOT hadoop writables. I would vote for JSON since the data is small, and it would be nice for it to be human readable without programmatic intervention.
> online schema updates
> ---------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Fix For: 0.21.0
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-1730) online schema updates
Posted by "Jim Kellerman (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12738152#action_12738152 ]
Jim Kellerman commented on HBASE-1730:
--------------------------------------
+1 for putting descriptions both on disk for cold start and in ZK for on-line.
As for format, how about Avro (which is json based)?
> online schema updates
> ---------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Fix For: 0.21.0
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HBASE-1730) Near-instantaneous online schema and
table state updates
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
stack updated HBASE-1730:
-------------------------
Summary: Near-instantaneous online schema and table state updates (was: online schema and tableAdd updates)
> Near-instantaneous online schema and table state updates
> --------------------------------------------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Assignee: stack
> Fix For: 0.21.0
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-1730) online schema and tableAdd updates
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12771705#action_12771705 ]
stack commented on HBASE-1730:
------------------------------
Going to serialize HTD and HCD to zk as JSON; this seems easiest way to go for schema. Will remove notion of state change from HTD and HCD though. For instance if a region is offline, you won't know by reading the .META.; master will know because will have read state from zk.
@apurtell I should use jackson for pojo 2 json because thats what you and avro are using. One concern is versions. Jackson is up to 1.2. avro is jackson-core-asl-1.0.1.jar. Stargate is 0.94. Let me start with the version in stargate. OK I pull it up to trunk/lib? You can get it when its there? Can I update to 1.0.1?
> online schema and tableAdd updates
> ----------------------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Assignee: stack
> Fix For: 0.21.0
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-1730) Near-instantaneous online schema and
table state updates
Posted by "Jean-Daniel Cryans (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12772368#action_12772368 ]
Jean-Daniel Cryans commented on HBASE-1730:
-------------------------------------------
Looking good! I like how we get rid of all the doubled stuff because we had to maintain a string and a writable.
> Near-instantaneous online schema and table state updates
> --------------------------------------------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Assignee: stack
> Fix For: 0.21.0
>
> Attachments: 1730-v2.patch, 1730.patch
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HBASE-1730) online schema and tableAdd updates
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
stack updated HBASE-1730:
-------------------------
Summary: online schema and tableAdd updates (was: online schema updates)
Assigned myself and added online state changes to title.
> online schema and tableAdd updates
> ----------------------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Assignee: stack
> Fix For: 0.21.0
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-1730) Near-instantaneous online schema and
table state updates
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12771770#action_12771770 ]
stack commented on HBASE-1730:
------------------------------
@apurtell nm... it seems like jackson is split in two in 1.0.1... if I have both jars in hbase/lib, then stargate tests all pass.
> Near-instantaneous online schema and table state updates
> --------------------------------------------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Assignee: stack
> Fix For: 0.21.0
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-1730) online schema and tableAdd updates
Posted by "ryan rawson (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12771703#action_12771703 ]
ryan rawson commented on HBASE-1730:
------------------------------------
while we are thinking about this, i'd like to be able to do:
- add new CF
- remove existing CF
- change compression
- change other things
without taking a table offline. that would be some killer features right there!
> online schema and tableAdd updates
> ----------------------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Assignee: stack
> Fix For: 0.21.0
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-1730) online schema updates
Posted by "Andrew Purtell (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12738155#action_12738155 ]
Andrew Purtell commented on HBASE-1730:
---------------------------------------
@Jim: If AVRO is JSON based, then why not just use JSON? Pulls in less external dependencies...
> online schema updates
> ---------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Fix For: 0.21.0
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-1730) online schema updates
Posted by "Andrew Purtell (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12737689#action_12737689 ]
Andrew Purtell commented on HBASE-1730:
---------------------------------------
+1 on all of Ryan's points.
> online schema updates
> ---------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Fix For: 0.21.0
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-1730) Near-instantaneous online schema and
table state updates
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12794317#action_12794317 ]
stack commented on HBASE-1730:
------------------------------
Started a branch over in github for this work: http://github.com/saintstack/hbase/tree/1730
> Near-instantaneous online schema and table state updates
> --------------------------------------------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Assignee: stack
> Fix For: 0.21.0
>
> Attachments: 1730-v2.patch, 1730.patch
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-1730) Near-instantaneous online schema and
table state updates
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12771827#action_12771827 ]
stack commented on HBASE-1730:
------------------------------
Playing w/ Jackson, making it understand our HTD and HCD would take a bunch of ugly work. I'm thinking that we just make HTD and HCD be maps. Then Jackson can do it w/o customization. HTD and HCD don't need to be Writables anymore now they are to be serialized instead as JSON into znode.
> Near-instantaneous online schema and table state updates
> --------------------------------------------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Assignee: stack
> Fix For: 0.21.0
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HBASE-1730) Near-instantaneous online schema and
table state updates
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1730?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
stack updated HBASE-1730:
-------------------------
Attachment: 1730.patch
Started in on HCD. Added test for new serialization. Depends on jackson 1.0.1. Long way from finished.
This work might have us move from ruby map representations to JSON in UI, shell and in znode. Will see. Shouldn't be prob. really since HRegionInfo will no longer be shown in shell as much since it will no longer be in .META. when this work is done.
> Near-instantaneous online schema and table state updates
> --------------------------------------------------------
>
> Key: HBASE-1730
> URL: https://issues.apache.org/jira/browse/HBASE-1730
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: Andrew Purtell
> Assignee: stack
> Fix For: 0.21.0
>
> Attachments: 1730.patch
>
>
> We should not need to take a table offline to update HCD or HTD.
> One option for that is putting HTDs and HCDs up into ZK, with mirror on disk catalog tables to be used only for cold init scenarios, as discussed on IRC. In this scheme, regionservers hosting regions of a table would watch permanent nodes in ZK associated with that table for schema updates and take appropriate actions out of the watcher. In effect, schema updates become another item in the ToDo list.
> {{/hbase/tables/<table-name>/schema}}
> Must be associated with a write locking scheme also handled with ZK primitives to avoid situations where one concurrent update clobbers another.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.