Timeline
2010-12-13:
- 15:40 Ticket #198 (Repository.create() on already existing record does not throw ...) created by
- Seems like RecordExistsException? is never thrown.
2010-12-10:
- 14:47 Ticket #150 (Excessive storage of delete markers in records) closed by
- fixed: Fixed in r4501
- 14:47 Changeset [4501] by
- Avoid excessive storage of delete markers (#150)
- 13:27 Ticket #157 (Consider using less column families for storing a record.) closed by
- fixed: Implemented in r4500 The column families have been merged into two column …
- 13:26 Ticket #197 (Consider merging the system and data column families) created by
- By fixing ticket #157 (r4500) the number of column families that are used …
- 13:16 Changeset [4500] by
- Reduced the record column families to two column families : system and …
2010-12-09:
- 15:29 Changeset [4499] by
- Some changes were missing in previous commit (r4498) (#152)
- 14:39 Ticket #152 (Large hbase column keys for storing fields) closed by
- fixed: Implemented in r4498
- 14:38 Changeset [4498] by
- Using the 16bytes instead of 36 bytes version to store column keys.
2010-12-08:
- 18:20 Changeset [4497] by
- Exclude netcdf library (tika dependency) as it includes an slf4j …
- 16:36 Changeset [4496] by
- Upgrade to Tika 0.8
- 15:45 Ticket #196 (Deal gently with unreachable remote listeners) created by
- (This came up while thinking about handling io errors in the lily client.) …
- 14:34 Changeset [4495] by
- * Adding extra logging to ValidationMessageListener? to enable …
- 14:29 Ticket #195 (Keep netty connection to remote listeners open) created by
- Currently for each message that is sent to a listener, the netty …
- 14:16 Ticket #194 (Avoid that all leader roles are claimed by the first started Lily node) created by
- Currently there is a good chance that one Lily node will take all leader …
- 14:12 Ticket #193 (Conditional record updates) created by
- A 'check and save' functionality for Lily: when doing an update, allow to …
- 14:08 Ticket #192 (Schema: allow specifying target record type of link fields) created by
- When making use of link fields, a schema would be better understandable …
- 14:04 Ticket #191 (Make records searchable by their record type) created by
- There should be some way to index the record type of a record, either by …
- 13:46 Ticket #163 (RowLocker: use different columns for different purposes) closed by
- invalid: This was a misunderstanding. The RowLocker? does provide the possibility …
- 08:14 Ticket #122 (Update subscriptions on the RowLogConfigurationManager) closed by
- fixed: Implemented in r4494 * Extended the RowLogConfigurationManager? to add and …
- 08:13 Ticket #121 (Extend RowLogConfigurationManager for RowLog) closed by
- fixed: Implemented in r4494 * Extended the RowLogConfigurationManager? to add and …
- 08:11 Changeset [4494] by
- * Extended the RowLogConfigurationManager? to add and update configuration …
- 08:10 Ticket #190 (Move message lock timeout from rowlog to subscription level) created by
- The LockTimeout? for messages is configured at the RowLog? level. It seems …
- 08:08 Ticket #189 (CLI commands for the RowLogConfigurationManager) created by
- Add CLI commands to manage the rowlog and susbcriptions information on the …
2010-12-06:
- 15:45 Changeset [4493] by
- Adjust LinkIndex? for hbaseindex API changes + make use of its new ability …
- 15:44 Changeset [4492] by
- hbaseindex: add methods to add or remove multiple index entries in one …
- 14:57 Changeset [4491] by
- Add an indexerconf & SOLR schema that can be used with the default tester …
2010-12-02:
- 15:07 Changeset [4490] by
- Add a tool to print the detected host name / ip address, that is used by …
- 14:24 Ticket #187 (RowLogMessage should explicitly contain a timestamp) closed by
- fixed
- 14:16 Ticket #167 (RowLog: optimize queue scanning) closed by
- fixed: This has been implemented in r4489 .
- 14:15 Changeset [4489] by
- In the RowLogShard? only scan messages starting from a certain timestamp if …
2010-12-01:
- 15:25 Ticket #184 (Add minimal message age for processing) closed by
- fixed: This is implemented in r4488 This removes again the performance impact …
- 15:22 Changeset [4488] by
- The rowlog processor only processes messages that have a minimal age …
- 13:40 Ticket #188 (Make RowLogProcessor delays configurable) created by
- The RowLogProcessor? has 3 delays : notifyDelay, minimalProcessDelay and …
- 13:29 Ticket #187 (RowLogMessage should explicitly contain a timestamp) created by
- The RowLogMessage? needs an explicit getTimestamp() call. Currently the …
2010-11-30:
- 15:35 Ticket #186 (Notification system for important events) created by
- A notification system is needed to make it possible to inform an …
- 14:53 Ticket #172 (exists check not needed in read path) closed by
- fixed: Fixed in r4487 Also avoided unnecessary exists checks in create path, …
- 14:51 Ticket #90 (WAL processor handles open WAL messages) closed by
- fixed: Implemented in r4487
- 14:50 Changeset [4487] by
- * Created a WAL processor which handles open WAL messages. (#90) * Added …
- 14:44 Ticket #185 (Regression tests) created by
- Define and create a suite of tests that can be used performance regression …
- 14:25 Ticket #184 (Add minimal message age for processing) created by
- Consider to make it possible to configure the minimal age of a message …
- 14:21 Ticket #183 (Optimize Rowlog processor) created by
- Profile the rowlog processor thread and also the processMessage call. …
- 14:20 Ticket #182 (Optimize CRUD paths) created by
- Profile the CRUD paths of the repository and optimize where possible.
- 14:14 Ticket #181 (Investigate effect of LZO compression on performance) created by
- Investigate the influence of using LZO compression at HBase level on the …
- 13:53 Ticket #180 (Consistent logging) created by
- Review all logging and make the log-levels and message style consistent.
- 13:48 Ticket #179 (Make size-based blobstore selection configurable) created by
- Lily currently only provides a size based selection for the blobstore …
- 13:33 Ticket #178 (Optimize uploading large blobs) created by
- Uploading very large blobs (>> 100MB) should be profiled and optimized …
- 13:23 Ticket #177 (Provide a tool to clean up orphaned blobs) created by
- Blobs that are uploaded but never coupled to a record become orphaned …
- 13:07 Ticket #176 (Access to freshly uploaded blobs should be limited) created by
- This ticket has been split off of ticket #95 , taking the access control …
2010-11-29:
- 09:03 Ticket #159 (lily-developers list is not receiving notifications anymore) closed by
- fixed: It works!
2010-11-26:
- 16:58 Ticket #175 (Optimize the expensive call calculateRecordChanges) created by
- The call calculateRecordChanges on HBaseRepository is expensive. It should …
- 16:55 Ticket #174 (Change record type of a record without changing a field) created by
- This issue has been split off from ticket #25 It is not possible to …
- 14:03 Ticket #37 (Recovery of clients talking to dissapearing Lily repo nodes) closed by
- duplicate: Duplicate of #142
- 10:44 Ticket #173 (Create multiple regions at first cluster start) created by
- When starting a new Lily cluster a new table is created to store the …
2010-11-24:
- 15:28 Ticket #172 (exists check not needed in read path) created by
- The exists check that happens in the read path (with a fail-fast purpose) …
- 15:01 Ticket #171 (Reuse put to take custom rowlock : create path) created by
- In the create path, a put is used for both the actual record data as the …
- 14:07 Ticket #170 (Administration UI : for schema configuration , indexer) created by
- An administration UI should make it easier to manage lily's schema and as …
- 13:46 Ticket #169 (Access control : misc thoughts) created by
- This is a placeholder ticket to collect ideas around access control. …
- 13:42 Ticket #168 (Performance : misc thoughts) created by
- This is a placeholder ticket to collect ideas around performance and …
- 13:07 Ticket #68 (Assure RowLog stays usbale outside of Lily) closed by
- fixed: Updated technical documentation : …
2010-11-23:
- 17:43 Changeset [4486] by
- Updated some javadocs
- 17:09 Changeset [4485] by
- Register a listener on the configuration manager. Shutdown properly. …
2010-11-22:
- 13:27 Ticket #167 (RowLog: optimize queue scanning) created by
- When scanning the rowlog table, to avoid scanning over a lot of puts and …
2010-11-21:
- 17:16 Ticket #166 (ZKPublisher: blobconfig in zookeeper not updated after initial creation) created by
- Updates done to the blobconfig are not reflected in zookeeper, causing …
2010-11-16:
- 16:54 Ticket #165 (hbaseindex: allow to set scanner cache size) created by
- Querying hbase indexes can probably be sped up by setting a cache size …
- 16:49 Ticket #164 (hbaseindex: allow adding multiple entries in one call) created by
- Currently each index entry needs to be added one by one, consequently each …
- 16:46 Ticket #163 (RowLocker: use different columns for different purposes) created by
- Currently all usages of the RowLocker? (custom row lock mechanism) all use …
- 16:41 Ticket #162 (Do not store unchanged versioned system fields for each version) created by
- Using the record-row-visualizer tool it can be noticed that the record …
- 09:06 Changeset [4484] by
- Merge from trunk: Fix hardcoded time waited for zk connection to be …
- 09:06 Changeset [4483] by
- Fix hardcoded time waited for zk connection to be established.
2010-11-15:
- 12:36 Ticket #161 (Use of semicolon in record ID: impact on URIs) created by
- As came up in the mailing list thread: …
Note: See TracTimeline
for information about the timeline view.