| 7899314 | Chinmay Soman | 19 October 2013, 01:06:03 UTC | Bumping up version for release 1.5.6 | 19 October 2013, 01:06:03 UTC |
| c883ae2 | Chinmay Soman | 19 October 2013, 01:02:38 UTC | Removing getStoreClientFactoryStats from the StoreClientFactory interface, since it broke other shit | 19 October 2013, 01:03:21 UTC |
| 4deff3c | Siddharth Singh | 18 October 2013, 21:05:26 UTC | Bump release number to 1.5.5 | 18 October 2013, 21:05:26 UTC |
| 2ad302f | Siddharth Singh | 18 October 2013, 20:48:17 UTC | Release notes for release-1.5.5 | 18 October 2013, 20:53:23 UTC |
| c35751e | Zhongjie Wu | 18 October 2013, 01:10:11 UTC | revision based on review | 18 October 2013, 01:11:36 UTC |
| c453ec6 | Zhongjie Wu | 10 October 2013, 02:07:44 UTC | fix parallel put action 1. recheck quorum and zone satisfaction after draining queue 2. Fix a logging typo | 18 October 2013, 01:11:36 UTC |
| d695d1b | Zhongjie Wu | 10 October 2013, 02:07:24 UTC | added zone count write test | 18 October 2013, 01:11:36 UTC |
| dbce935 | Chinmay Soman | 15 October 2013, 23:16:23 UTC | Initializing the AtomicInteger numConsecutiveCatastrophicErrors properly in the constructor of NodeStatus | 15 October 2013, 23:16:23 UTC |
| 6d5d4bf | Chinmay Soman | 15 October 2013, 03:33:42 UTC | Improvements to the failure detector: Adding a threhold for fatal failures while marking a node down | 15 October 2013, 03:33:42 UTC |
| 728b63a | Siddharth Singh | 11 October 2013, 04:03:00 UTC | Refactor build once push multi code path | 14 October 2013, 22:57:38 UTC |
| 851b781 | Vinoth Chandar | 08 October 2013, 02:07:11 UTC | Disabling the gossiper test | 08 October 2013, 02:07:11 UTC |
| 93c9161 | Vinoth Chandar | 08 October 2013, 00:43:34 UTC | Fixes to rebalancing tests -- Fix bdb misconfig to speed up ~10x -- Disable the BlockingIO testing | 08 October 2013, 00:43:34 UTC |
| e7022bc | Siddharth Singh | 01 October 2013, 21:58:04 UTC | Exclude HadoopStoreBuilder tests | 03 October 2013, 22:09:15 UTC |
| 63888bb | Siddharth Singh | 25 September 2013, 21:56:51 UTC | Add a new "multi" option:build once,push multi | 03 October 2013, 22:08:18 UTC |
| 15ed9b9 | Chinmay Soman | 02 October 2013, 18:55:11 UTC | Cleaning up changes to TestUtils.java meant for creating temp directories in a standard way | 03 October 2013, 20:54:08 UTC |
| afaab5e | Chinmay Soman | 26 September 2013, 21:39:45 UTC | Creating all the unit test temp directories under a separate vold-test-data directory | 03 October 2013, 20:54:08 UTC |
| 90e947d | Vinoth Chandar | 02 October 2013, 21:27:38 UTC | Checking in examples for shell usage | 02 October 2013, 21:29:24 UTC |
| b4c20cf | Vinoth Chandar | 25 September 2013, 22:36:00 UTC | Rewriting ClientShell with Avro support | 02 October 2013, 21:29:24 UTC |
| 8e6d766 | vinoth chandar | 02 October 2013, 18:41:53 UTC | Merge pull request #152 from oldtopos/master Fix doctest error | 02 October 2013, 18:41:53 UTC |
| e79f9c0 | John Novak | 02 October 2013, 05:55:14 UTC | Fix doctest error Replace expected return value with proper string | 02 October 2013, 05:55:14 UTC |
| 8178845 | Zhongjie Wu | 30 September 2013, 19:19:19 UTC | fix on bind exceptions in tests | 01 October 2013, 23:27:46 UTC |
| fa12ee8 | Vinoth Chandar | 27 September 2013, 17:47:11 UTC | Removing vector clock sort code from KeyVersionFetcher | 01 October 2013, 21:00:19 UTC |
| 0a97728 | Vinoth Chandar | 27 September 2013, 17:12:23 UTC | Fixing throttling in repair job and prune job | 27 September 2013, 17:12:23 UTC |
| 66fefb6 | Vinoth Chandar | 27 September 2013, 01:46:45 UTC | Fixing bug to release lock when an record was not pruned | 27 September 2013, 01:46:45 UTC |
| 2666dca | Zhongjie Wu | 16 September 2013, 23:18:20 UTC | prevent NPE in worker thread when shutdown | 23 September 2013, 18:49:22 UTC |
| f3e6627 | Chinmay Soman | 20 September 2013, 19:36:55 UTC | Re-enabling the auto format on save option | 20 September 2013, 19:36:55 UTC |
| 427d2b2 | Siddharth Singh | 20 September 2013, 00:48:12 UTC | Check stats object to be null to avoid NPE | 20 September 2013, 04:29:55 UTC |
| 2f05b8d | Chinmay Soman | 20 September 2013, 01:02:34 UTC | Updating build.properties and release notes for Voldemort jar release 1.5.3 | 20 September 2013, 01:02:34 UTC |
| 07e3392 | Chinmay Soman | 19 September 2013, 22:54:59 UTC | Naming convention cleanup of R2Store | 19 September 2013, 22:55:32 UTC |
| 6d08137 | Chinmay Soman | 19 September 2013, 00:47:45 UTC | Refactoring RESTClientFactory and R2Store to accept a D2 client | 19 September 2013, 22:55:32 UTC |
| 0de0e46 | Vinoth Chandar | 18 September 2013, 21:21:07 UTC | Fixing NPE in during client pool shutdown | 19 September 2013, 22:03:01 UTC |
| ef28f0a | Siddharth Singh | 19 September 2013, 00:16:44 UTC | Track connection exception as a separate metric | 19 September 2013, 00:21:17 UTC |
| 294b95e | Siddharth Singh | 17 September 2013, 23:50:20 UTC | Separate sync and async operation times | 17 September 2013, 23:51:07 UTC |
| 4f9d970 | Siddharth Singh | 16 September 2013, 22:29:06 UTC | Check monitoring interval before getting histogram Make sure that that if the interval is expired you return a reset histogram and not the stale one. | 16 September 2013, 22:29:06 UTC |
| 7272eba | Siddharth Singh | 16 September 2013, 22:24:40 UTC | Record op time on the async code path | 16 September 2013, 22:24:40 UTC |
| 9a7faba | Siddharth Singh | 13 September 2013, 21:07:09 UTC | Test for monitoring interval Ensure checking the interval before returning the histogram | 13 September 2013, 21:10:11 UTC |
| 7ec27f8 | Siddharth Singh | 13 September 2013, 19:46:13 UTC | Minor changes as per the review | 13 September 2013, 21:10:01 UTC |
| 4a46321 | Siddharth Singh | 11 September 2013, 22:44:16 UTC | Fix time convertion (Nano to micro) bug Make the method accept nanoseconds to avoid unnecessary convertion. | 13 September 2013, 18:32:05 UTC |
| 19fbd2b | Siddharth Singh | 11 September 2013, 22:20:09 UTC | Use requestcounter in clientsocket stats | 13 September 2013, 18:30:56 UTC |
| 40f7d54 | Siddharth Singh | 09 September 2013, 23:04:56 UTC | Bootstrap stats Track bootstrap events, reboostrap events and failed bootstrap events | 13 September 2013, 18:29:26 UTC |
| 5bacb97 | Vinoth Chandar | 12 September 2013, 22:01:14 UTC | Relocating RepairJob and PruneJob | 13 September 2013, 01:10:52 UTC |
| 722386f | Vinoth Chandar | 09 September 2013, 23:28:57 UTC | Expose timebased resolving from Streaming Client. | 13 September 2013, 01:10:51 UTC |
| 8cb8439 | Siddharth Singh | 06 September 2013, 00:05:12 UTC | Lower down the reset value to 60 seconds | 09 September 2013, 20:07:56 UTC |
| 0cfe553 | Siddharth Singh | 05 September 2013, 23:45:58 UTC | Use time based counter instead of sample based Use RequestCounter for connection establishment metrics | 09 September 2013, 20:07:56 UTC |
| 5f3c4f9 | Siddharth Singh | 05 September 2013, 22:11:18 UTC | Add comment explaining potential race and fix | 09 September 2013, 20:07:56 UTC |
| 1b257c2 | Siddharth Singh | 05 September 2013, 21:54:35 UTC | Make sure bean is registered only once A potential race can cause the bean to be registered more than once. This commit uses the return value of putIfAbsent to short ciruit the registration code. | 09 September 2013, 20:07:56 UTC |
| 3be1ae2 | Siddharth Singh | 05 September 2013, 20:49:28 UTC | Jmx changes for conn establishment and op time Jmx changes related to connection establishment time and operation time. | 09 September 2013, 20:07:55 UTC |
| 478005a | Siddharth Singh | 05 September 2013, 00:49:51 UTC | Report accurate op time by excluding checkout time Fix operation time to not include the checkout time and expose it is a bean. | 09 September 2013, 20:07:55 UTC |
| 7583d0b | Siddharth Singh | 05 September 2013, 00:20:02 UTC | Report connection establishment time | 09 September 2013, 20:07:55 UTC |
| 662ce35 | Vinoth Chandar | 04 September 2013, 21:55:44 UTC | Adding ability to inject any je.properties param into server * Added cleanerFetchObsoleteSize=true as default * Removed unused proactive migration param | 05 September 2013, 18:34:52 UTC |
| 000c79b | Chinmay Soman | 28 August 2013, 20:28:12 UTC | Catching Bootstrap exception while starting CoordinatorService. This is done to ensure Coordinator comes up even without a valid cluster | 05 September 2013, 01:54:47 UTC |
| 7c61537 | Chinmay Soman | 19 August 2013, 22:44:28 UTC | Modified RestServiceR2StoreTest to use Rest server port | 05 September 2013, 01:54:47 UTC |
| 187340f | Chinmay Soman | 19 August 2013, 22:25:19 UTC | Bunch of cleanup changes: * Better shutdown of Rest server * Fixed bug in Get, GetAll parsing in R2Store * Added GetVersions in R2Store * Modified key encoding to use encodeBase64URLSafeString instead of standard base64 encoding * Additional tests for R2Store and RestClient | 05 September 2013, 01:54:47 UTC |
| 6380ecf | Chinmay Soman | 15 August 2013, 20:10:40 UTC | Adding a RestClient testcase | 05 September 2013, 01:54:47 UTC |
| 06b450d | Vinoth Chandar | 29 August 2013, 21:15:53 UTC | Addressing code review comments. | 29 August 2013, 21:23:51 UTC |
| af695ed | Vinoth Chandar | 28 August 2013, 20:32:40 UTC | Preventing DataCleanup Job from running during maintenance | 29 August 2013, 21:19:07 UTC |
| 17ea6e6 | Vinoth Chandar | 28 August 2013, 02:29:02 UTC | Adding the missed file | 29 August 2013, 21:19:07 UTC |
| 5aaa854 | Vinoth Chandar | 27 August 2013, 23:45:21 UTC | Timestamp based resolving for UpdateEntries * Refactored all the streaming write classes to reuse network IO code * Tests for Timebased resolving * Tests for getAndLock and putAndLock * Fix broken MySQL repair job from previous changes * Reimplemented multiVersionPut using getAndLock and putAndLock | 29 August 2013, 21:19:07 UTC |
| 2150275 | Vinoth Chandar | 22 August 2013, 22:42:25 UTC | Adding test for Prune job | 29 August 2013, 21:17:37 UTC |
| 155c98b | Siddharth Singh | 22 August 2013, 17:48:21 UTC | Address review comments | 29 August 2013, 21:17:37 UTC |
| c6d9720 | Siddharth Singh | 21 August 2013, 20:48:03 UTC | Add throttle option to versioned put prune job | 29 August 2013, 21:17:37 UTC |
| 4642417 | Siddharth Singh | 21 August 2013, 20:04:33 UTC | Unit test for repair job | 29 August 2013, 21:17:37 UTC |
| 9aa3b0d | Siddharth Singh | 20 August 2013, 19:53:55 UTC | Add throttle option to repair job | 29 August 2013, 21:17:37 UTC |
| 175ab04 | Vinoth Chandar | 20 August 2013, 02:40:16 UTC | Adding releaseLock(handle) for error handling and cleanup | 29 August 2013, 21:17:37 UTC |
| 027c51c | Vinoth Chandar | 20 August 2013, 02:22:59 UTC | Adding getAndLock and putAndUnlock storage calls -- Intended for the VersionedPutPruneJob -- Enables custom read-modify-write logic | 29 August 2013, 21:17:37 UTC |
| 8256518 | Vinoth Chandar | 20 August 2013, 00:59:36 UTC | Adding prune job to fix versioned puts post rebalance | 29 August 2013, 21:17:37 UTC |
| ed144a8 | Zhongjie Wu | 19 August 2013, 18:45:26 UTC | Change store tracking for 95th and 99th improvement 1. Use a smaller range histogram with higher accuracy 2. Changed update window from 5 minutes to 2 minutes 3. Changed update method for q95th and q99th values 4. Changed value type of q95th and q99th from long to double | 27 August 2013, 20:44:39 UTC |
| d2c2ded | Chinmay Soman | 13 August 2013, 23:58:01 UTC | Added ability to track key size in StoreStats | 26 August 2013, 21:36:39 UTC |
| 81e7d3e | Abhinay Nagpal | 26 August 2013, 18:23:16 UTC | The regular expression did not work for reords and unions | 26 August 2013, 18:24:54 UTC |
| 0a699eb | Abhinay Nagpal | 20 August 2013, 01:48:56 UTC | -Added a voldemort client shell with suport for avro stores - testing done: for a store with schema <key-serializer> <type>avro-generic-versioned</type> <schema-info version="0">"string"</schema-info> </key-serializer> <value-serializer> <type>avro-generic-versioned</type> <schema-info version="0">"string"</schema-info> </value-serializer> operations would be: put "a" "b" get "a" | 26 August 2013, 18:24:54 UTC |
| d247d98 | Siddharth Singh | 26 August 2013, 17:38:47 UTC | Disable actions on save in eclipse preference | 26 August 2013, 17:44:15 UTC |
| 0c34ff9 | Vinoth Chandar | 14 August 2013, 00:10:01 UTC | Accounting rebalancing puts in streaming stats | 23 August 2013, 23:02:52 UTC |
| 48b8f1c | Vinoth Chandar | 09 August 2013, 17:56:23 UTC | Adding aggregated BDB stats MBean | 23 August 2013, 23:00:40 UTC |
| e25966b | Bhavani Sudha Saktheeswaran | 16 August 2013, 17:49:14 UTC | Fixing zoneId in Node.java | 16 August 2013, 21:18:49 UTC |
| ba3ad46 | Bhavani Sudha Saktheeswaran | 13 August 2013, 19:03:52 UTC | Rest Service: Minor fixes - adding log messages - Editing java doc | 16 August 2013, 21:18:49 UTC |
| 4e9ec16 | Bhavani Sudha Saktheeswaran | 13 August 2013, 02:16:51 UTC | Adding per store latencies Defining new port for Rest Service in cluster.xml | 16 August 2013, 21:18:49 UTC |
| c59e2c8 | Bhavani Sudha Saktheeswaran | 09 August 2013, 23:51:37 UTC | Made Netty server backlog parameter configurable Made Http Aggregated content length parameter configurable Added few validation checks in VoldemortConfig Deleted Missing Routing type parameter test case from RestServerProtocolTests.java since Routing type is not mandatory anymore. | 16 August 2013, 21:18:49 UTC |
| b53f1c9 | Chinmay Soman | 13 August 2013, 00:18:01 UTC | * Added delete entries metric for DataCleanupJob, RepairJob * Fixed a critical bug in ScanPermitWrapper * Added slop scan rate in StreamingStats | 15 August 2013, 22:24:52 UTC |
| 9a0e9ec | Chinmay Soman | 09 August 2013, 00:24:54 UTC | Adding monitoring for Get Versions | 15 August 2013, 22:24:52 UTC |
| b5ee2d0 | Siddharth Singh | 25 July 2013, 21:20:45 UTC | Fix path in zone expansion | 15 August 2013, 21:18:45 UTC |
| 1075db7 | Jay J Wylie | 14 August 2013, 15:54:35 UTC | Fix typo. | 15 August 2013, 00:04:25 UTC |
| 0ac65db | Jay J Wylie | 08 August 2013, 21:54:34 UTC | Dropped dead code files. | 15 August 2013, 00:04:24 UTC |
| a84b669 | Jay J Wylie | 08 August 2013, 21:52:25 UTC | Added some more package.html files. | 15 August 2013, 00:04:24 UTC |
| e6c680e | Jay J Wylie | 08 August 2013, 21:52:00 UTC | Fixed all javadoc warnings. | 15 August 2013, 00:04:24 UTC |
| 2cb4b3e | Siddharth Singh | 06 August 2013, 20:05:03 UTC | Move replicatype methods to ROTestUtils | 14 August 2013, 23:39:59 UTC |
| b530366 | Siddharth Singh | 06 August 2013, 17:21:25 UTC | Drop unnecessary test parameterization. | 14 August 2013, 23:39:59 UTC |
| 6a8453c | Siddharth Singh | 06 August 2013, 01:32:28 UTC | Get rid of donor based rebalancing code | 14 August 2013, 23:39:59 UTC |
| d2d9b66 | Siddharth Singh | 02 August 2013, 20:48:24 UTC | Use storeroutingplan to get master partition id | 14 August 2013, 23:39:59 UTC |
| 1e588c5 | Siddharth Singh | 02 August 2013, 18:00:37 UTC | Use rebalanceutils instead of clusterutils | 14 August 2013, 23:39:59 UTC |
| 31a5f11 | Siddharth Singh | 02 August 2013, 02:24:18 UTC | Tweaks to Rebalance Controller and Scheduler Changed '==' to '>=' comparison to control parallelism permitted by the scheduler. This is defensive programming. Under failure, the '==' was insufficient. We need to debug/fix that separately. For now, this change adds some more defense. Changed some HashSet's to TreeSet's so that they print out in order. (No perf impact since sets are all small.) Additional print outs from scheduler so that current set of node (ids) with work is in the log from the controller/scheduler. Include donor node in logger.info from StealerBasedRebalanceTask constructor. This output is now sufficient to translate a task id to a specific stealer/donor pair in a single line (and all the partition-stores in follow on lines). | 14 August 2013, 23:39:59 UTC |
| 6154e7d | Siddharth Singh | 02 August 2013, 02:22:16 UTC | Clean up RebalanceUtils and AdminClient Moved a number of static methods that took an AdminClient argument from RebalanctUtils to become members of AdminClient.RebalancingOperations (invoked via adminClient.rebalanceOps). Also moved createTempAdminClient to be a static method of AdminClient. Moved some of AdminClient.HelperOps into new inner class AdminClient.ReplicaTypeOperations. | 14 August 2013, 23:39:59 UTC |
| e73db18 | Siddharth Singh | 02 August 2013, 00:10:30 UTC | Use .cloneCluster instead of Mapper.read/.write. | 14 August 2013, 23:39:59 UTC |
| b0bdc21 | Jay J Wylie | 15 July 2013, 19:48:09 UTC | Add pointer to proxy-put-metadata-listener branch. | 14 August 2013, 23:39:59 UTC |
| 0d18516 | Siddharth Singh | 01 August 2013, 23:55:53 UTC | Drop ClusterUtils and NodeUtils | 14 August 2013, 23:39:58 UTC |
| d4028e3 | Siddharth Singh | 01 August 2013, 23:54:47 UTC | Sundry clean up in voldemort/utils UpdateClusterUtils - added new utils class - moved all cluster & node manipulation methods to this class. (E.g., methods that tweak partition IDs on a node, and methods that copy all nodes in a cluster, etc.) Cluster - replaced implementation of .cloneCluster() with another helper method's implementation. PartitionBalanceUtils - new utils class - moved all helper methods that analyze cluster/storedef to this class RebalanceUtils - many helpers moved out of here to more other, new util classes ClusterUtils & NodeUtils - found new homes for all helper methods. Will drop these files in next commit. | 14 August 2013, 23:39:58 UTC |
| eba3646 | Jay J Wylie | 10 July 2013, 17:10:30 UTC | Removed maxTries option from Rebalancing maxTries was a hack to work around problems with the scheduling of rebalancing tasks. The new scheduler (is supposed to) ensure that no node is doing more than one rebalancing thing (stealing or donating) at time. Therefore, not being able to schedule an asynchronous rebalancing task (stealer-based) on a node is a big problem and should not be retried. | 14 August 2013, 23:39:58 UTC |
| 3920c81 | Jay J Wylie | 09 July 2013, 21:04:59 UTC | Drop server-side rebalancing optimization, fix proxy put config The RebalancePlan object now produces "optimized" plans. I.e., data is not moved to a server that already hosts that data. This means that server-side optimizations to not fetch data the server already has are not needed. The voldemort config had a parameter to turn proxy puts on/off. Now that they are tested, they should always be on for the sake of making sure rebalances are abortable. Also, the proxy put default number of threads was one. It is now set in the same manner as nio selectors. Sorry for all the unnecessary white space changes. I need to reset my eclipse settings. | 14 August 2013, 23:39:58 UTC |
| 67802f5 | Jay J Wylie | 09 July 2013, 17:54:53 UTC | Complete dropping KeyDistributionGenerator | 14 August 2013, 23:39:58 UTC |
| 573d1bf | Jay J Wylie | 09 July 2013, 17:51:46 UTC | Drop KeyDistributionGenerator and Entropy The intended usage of KeyDistributionGenerator and Entropy has been replaced. The PartitionAnalysisCLI and PartitionBalance tools replace the 'skew' estimation functionality of these tools. The KeySampler and KeyVerionFetcher replace the verification functionality of the EntropyChecker. The helper method KeyDistributionGenerator.getUniqueStoreDefinitionsWithCounts has moved to StoreDefinitionUtils.getUniqueStoreDefinitionsWithCounts. | 14 August 2013, 23:39:58 UTC |
| 7f2e757 | Jay J Wylie | 12 July 2013, 20:03:56 UTC | Sort versions displayed by KeyVersionFetcherCLI | 14 August 2013, 23:39:58 UTC |
