- 27 7月, 2023 5 次提交
-
-
由 Haonan 提交于
-
由 Haonan 提交于
-
由 shuwenwei 提交于
-
由 YuFengLiu 提交于
-
由 William Song 提交于
-
- 26 7月, 2023 3 次提交
-
-
由 Liao Lanyu 提交于
-
由 Itami Sho 提交于
Co-authored-by: NSteve Yurong Su <rong@apache.org>
-
由 Haonan 提交于
-
- 25 7月, 2023 8 次提交
-
-
由 Haonan 提交于
-
由 Haonan 提交于
-
由 Zhijia Cao 提交于
-
由 Haonan 提交于
-
由 Haonan 提交于
-
由 Xiangpeng Hu 提交于
-
由 William Song 提交于
-
由 Li Yu Heng 提交于
-
- 24 7月, 2023 5 次提交
-
-
由 William Song 提交于
-
由 马子坤 提交于
[IOTDB-6081] Pipe: use HybridExtractor instead of LogExtractor when realtime mode is set to log to avoid OOM under heavy insertion load (#10620) * When user specifies 'extractor.realtime.mode'='log', we use hybrid extractors which can automatically degrade to file mode when load is too heavy. * Add a parameter: 'extractor.realtime.mode'='forced-log', the pipe will always use log mode and will never degrade to file mode. --------- Co-authored-by: NSteve Yurong Su <rong@apache.org>
-
由 Haonan 提交于
-
由 Chen YZ 提交于
-
由 Chen YZ 提交于
-
- 22 7月, 2023 1 次提交
-
-
由 YangCaiyin 提交于
-
- 21 7月, 2023 7 次提交
-
-
由 Liao Lanyu 提交于
-
由 Xiangpeng Hu 提交于
-
由 CritasWang 提交于
-
由 ZhangHongYin 提交于
-
由 Weihao Li 提交于
-
由 Chen YZ 提交于
-
由 Itami Sho 提交于
## Description After loading tsfile, it causes 100% load on one CPU core. ## Reason Root cause: high processor load for ScheduledThreadPoolExecutor with 0 core threads. It was a bug in JDK1.8. See the following issues for more details: [JDK-8129861](https://bugs.openjdk.java.net/browse/JDK-8129861) [JDK-8022642](https://bugs.openjdk.java.net/browse/JDK-8022642) ## Solution 1. Set corePoolSize to 1 to avoid high processor load (a workaround) 2. Lazy initialization: init loadTsFileManager on first call (to avoid unnecessary threads' creation)
-
- 20 7月, 2023 5 次提交
-
-
由 Steve Yurong Su 提交于
[IOTDB-6075] Pipe: File resource races when different tsfile load operations concurrently modify the same tsfile at receiver (#10629) * Add a parameter in iotdb-core/node-commons/src/assembly/resources/conf/iotdb-common.properties to control the connection timeout. * Set the pipe connection timeout between sender and receiver to 15 mins to allow long time-cost load operation. * Redesign the pipe receiver's dir to avoid file resource races when different tsfile load operations concurrently modify the same tsfile.
-
由 Zhang.Jinrui 提交于
-
由 Weihao Li 提交于
-
由 Chen YZ 提交于
-
由 Zhang.Jinrui 提交于
Make the candidate size of seq tsfile in cross space compaction will be smaller than target_compaction_file_size * 1.5. (#10602)
-
- 19 7月, 2023 4 次提交
-
-
由 Weihao Li 提交于
-
由 Chen YZ 提交于
-
由 Beyyes 提交于
-
由 William Song 提交于
-
- 18 7月, 2023 2 次提交
-
-
由 YangCaiyin 提交于
-
由 马子坤 提交于
-