commits@curator.apache.org, 2013-09

You are viewing a plain text version of this content. The canonical link for it is here.

- git commit: ADD tickTime AND maxClientCnxns SUPPORT TO CURATOR TESTS - posted by ra...@apache.org on 2013/09/04 06:15:43 UTC, 1 replies.
- git commit: Callbacks were not called on retry-able errors - this change adds it in - posted by ra...@apache.org on 2013/09/05 07:19:18 UTC, 1 replies.
- git commit: InterProcessSemaphoreMutex changed to use InterProcessSemaphoreV2. In the process, a bug in InterProcessSemaphoreV2 was noticed that was been long fixed in the older InterProcessSemaphore. If the session drops in the process of creating the l - posted by ra...@apache.org on 2013/09/06 08:06:41 UTC, 1 replies.
- git commit: ensurePath might throw an exception if connection couldn't succeed. The exception was being swallowed by createNode causing the recipe to fail if there wasn't a good connection when the object was started. Switched to use creatingParentsIfNee - posted by ra...@apache.org on 2013/09/06 19:49:49 UTC, 1 replies.
- git commit: Taking a custom Executor and ThreadFactory wasn't thought through. It's actually useless. This constructor is now deprecated in favor of passing in an ExecutorService. The old version is wrapped internally so that it works as it used to. But, - posted by ra...@apache.org on 2013/09/07 04:40:38 UTC, 1 replies.
- git commit: wrapExecutor was wrong - there's no way to use the ThreadFactory. Just ignore it now - posted by ra...@apache.org on 2013/09/07 05:00:47 UTC, 1 replies.
- git commit: Added missing license: - posted by ra...@apache.org on 2013/09/07 06:11:55 UTC, 1 replies.
- [1/2] git commit: Added test from user Michael Morello that exposes major retry hole in background operations. The getZooKeeper() method is not retried - posted by ra...@apache.org on 2013/09/10 22:45:50 UTC, 0 replies.
- [2/2] git commit: Fix edge case reported as CURATOR-52. ConnectionState.checkTimeouts() throws KeeperException.ConnectionLossException when the initial (or previously failed) connection cannot be re-established. This needs to be run through the retry pol - posted by ra...@apache.org on 2013/09/10 22:45:51 UTC, 0 replies.
- git commit: It was not possible to correctly handle connection state issues with the LeaderSelectorListener. There was an edge case where if the connection was lost before takeLeadership() was called, the thread would not be known and the existing sample - posted by ra...@apache.org on 2013/09/11 00:58:22 UTC, 1 replies.
- git commit: Added LeaderSelectorListenerAdapter to enforce/help with correct connection state handling. Also, updated docs and added another test, - posted by ra...@apache.org on 2013/09/11 17:58:50 UTC, 1 replies.
- git commit: Removed a Java 7ism: - posted by ra...@apache.org on 2013/09/11 18:11:42 UTC, 1 replies.
- git commit: CURATOR-55: EnsurePath should have a getPath() function - posted by ra...@apache.org on 2013/09/17 17:34:43 UTC, 1 replies.
- git commit: When performBackgroundOperation gets a retryable CuratorConnectionLossException, don't do it inline. Queue the background operation as usual - posted by ra...@apache.org on 2013/09/23 02:52:33 UTC, 0 replies.
- git commit: When the message was requeued, the wrong node name was being used causing a node name length leak - increasingly long node names - posted by ra...@apache.org on 2013/09/23 03:28:26 UTC, 1 replies.
- [1/4] git commit: Added test from user Michael Morello that exposes major retry hole in background operations. The getZooKeeper() method is not retried - posted by ra...@apache.org on 2013/09/24 03:54:37 UTC, 0 replies.
- [2/4] git commit: Fix edge case reported as CURATOR-52. ConnectionState.checkTimeouts() throws KeeperException.ConnectionLossException when the initial (or previously failed) connection cannot be re-established. This needs to be run through the retry pol - posted by ra...@apache.org on 2013/09/24 03:54:38 UTC, 0 replies.
- [3/4] git commit: When performBackgroundOperation gets a retryable CuratorConnectionLossException, don't do it inline. Queue the background operation as usual - posted by ra...@apache.org on 2013/09/24 03:54:39 UTC, 0 replies.
- [4/4] git commit: Merge branch 'CURATOR-52' - posted by ra...@apache.org on 2013/09/24 03:54:40 UTC, 0 replies.
- git commit: added license - posted by ra...@apache.org on 2013/09/24 04:16:24 UTC, 0 replies.
- [1/2] git commit: added license - posted by ra...@apache.org on 2013/09/24 04:16:39 UTC, 0 replies.
- [2/2] git commit: Merge branch 'CURATOR-52' - posted by ra...@apache.org on 2013/09/24 04:16:40 UTC, 0 replies.
- git commit: CuratorEventType.SET_ACL didn't exist - wrong type was being set - posted by ra...@apache.org on 2013/09/25 05:22:58 UTC, 1 replies.
- git commit: Warn if session timeout is less than connection timeout - posted by ra...@apache.org on 2013/09/25 05:28:34 UTC, 1 replies.
- git commit: Use isTraceEnabled to prevent unnecessary string concat - posted by ra...@apache.org on 2013/09/25 05:31:36 UTC, 1 replies.
- git commit: Wrong path was being usedin getParticipantNodes() - posted by ra...@apache.org on 2013/09/25 05:49:03 UTC, 1 replies.
- git commit: Work around edge case with protected mode. Make sure created node gets deleted if retries fail during connection loss - posted by ra...@apache.org on 2013/09/25 17:59:47 UTC, 0 replies.
- [1/2] git commit: Work around edge case with protected mode. Make sure created node gets deleted if retries fail during connection loss: see CURATOR-45. The issue is that retries fail due to connection timeout but the node was actually created on the ser - posted by ra...@apache.org on 2013/09/25 18:07:22 UTC, 0 replies.
- [2/2] git commit: Merge branch 'CURATOR-45' of https://git-wip-us.apache.org/repos/asf/incubator-curator into CURATOR-45 - posted by ra...@apache.org on 2013/09/25 18:07:23 UTC, 0 replies.
- [1/3] git commit: Work around edge case with protected mode. Make sure created node gets deleted if retries fail during connection loss - posted by ra...@apache.org on 2013/09/25 18:09:38 UTC, 0 replies.
- [2/3] git commit: Work around edge case with protected mode. Make sure created node gets deleted if retries fail during connection loss: see CURATOR-45. The issue is that retries fail due to connection timeout but the node was actually created on the ser - posted by ra...@apache.org on 2013/09/25 18:09:39 UTC, 0 replies.
- [3/3] git commit: Merge branch 'CURATOR-45' of https://git-wip-us.apache.org/repos/asf/incubator-curator into CURATOR-45 - posted by ra...@apache.org on 2013/09/25 18:09:40 UTC, 0 replies.
- git commit: Major bug! If the initial connection never succeeded, ConnectionState.CONNECTED would get sent when ConnectionState.LOST was intended - posted by ra...@apache.org on 2013/09/26 19:54:28 UTC, 1 replies.
- git commit: Include log4j so that LeaderSelectorExample runs - posted by ra...@apache.org on 2013/09/30 16:54:44 UTC, 0 replies.