You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@bookkeeper.apache.org by "Rakesh R (Commented) (JIRA)" <ji...@apache.org> on 2012/03/16 11:53:39 UTC
[jira] [Commented] (BOOKKEEPER-186) Bookkeeper throttling - permits
is not released when read has failed from all replicas
[ https://issues.apache.org/jira/browse/BOOKKEEPER-186?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13231050#comment-13231050 ]
Rakesh R commented on BOOKKEEPER-186:
-------------------------------------
+Suggestion:+
Since permits is acquired for each entries, we couldn't keep the release logic inside submitCallback. IMO, it would be fine to place 'lh.opCounterSem.release();' before submitCallback(lastErrorCode);
> Bookkeeper throttling - permits is not released when read has failed from all replicas
> --------------------------------------------------------------------------------------
>
> Key: BOOKKEEPER-186
> URL: https://issues.apache.org/jira/browse/BOOKKEEPER-186
> Project: Bookkeeper
> Issue Type: Bug
> Components: bookkeeper-client
> Affects Versions: 4.0.0
> Reporter: Rakesh R
> Fix For: 4.1.0
>
>
> Permit is not releasing in the case when there is no quorum available and the readEntries operation failed from all replicas.
> Following is the condition where it checks the failure in quorum and invoking ReadCallback:
> {noformat}
> void sendRead(ArrayList<InetSocketAddress> ensemble, LedgerEntry entry, int lastErrorCode) {
> if (entry.nextReplicaIndexToReadFrom >= lh.metadata.quorumSize) {
> // we are done, the read has failed from all replicas, just fail the
> // read
> submitCallback(lastErrorCode);
> return;
> }
> {noformat}
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira