HBASE-26088 Fix thread leaks in conn#getBufferedMutator(tableName) method call#3506
Conversation
|
@saintstack @anoopsjohn @apurtell This PR is just for branch-2. The behavior is different in master branch. In master brach, there is only AsyncConnection and I don't see it creating a new ThreadPoolExecutor but still studying master code base. Please review. |
|
🎊 +1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
anoopsjohn
left a comment
There was a problem hiding this comment.
Just a nit.
This issue is not there in branch-1 anyways right?
So for all branch-2 branches, we are sticking with the existing behaviour to have the compatibility and correcting the Javadoc. Pls mark this issue accordingly and we can address master branch (3.x also?) in another task. There we can even change to use the current javadoc way of using Connection's pool (If bug is there still)
| * use the Connection's ExecutorService. This object can be used for long lived operations. | ||
| * {@link BufferedMutator} returned by this method is thread-safe. | ||
| * This accessor will create a new ThreadPoolExecutor and will be shutdown once we close the | ||
| * BufferedMutator. This object can be used for long lived operations. |
There was a problem hiding this comment.
Nit: May be for API getBufferedMutator(BufferedMutatorParams params) we can say clearly that if user passes a ThreadPool in params, we will use that or else same as this method (new one and handled by us on close)
There was a problem hiding this comment.
@anoopsjohn Done ! Please review again !
branch-1 seems no such issue, because there's no this flag |
|
🎊 +1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
saintstack
left a comment
There was a problem hiding this comment.
Great. Needs release note.
…thod call (#3506) Co-authored-by: Rushabh Shah <rushabh.shah@rushabh-ltmflld.internal.salesforce.com> Signed-off-by: zhangduo <zhangduo@apache.org> Signed-off-by: Anoop <anoopsamjohn@apache.org> Signed-off-by: stack <stack@duboce.net> Signed-off-by: Reid Chan <reidchan@apache.org>
…thod call (#3506) Co-authored-by: Rushabh Shah <rushabh.shah@rushabh-ltmflld.internal.salesforce.com> Signed-off-by: zhangduo <zhangduo@apache.org> Signed-off-by: Anoop <anoopsamjohn@apache.org> Signed-off-by: stack <stack@duboce.net> Signed-off-by: Reid Chan <reidchan@apache.org>
No description provided.