|
欢迎加入运维网交流群:263444886
Storm 是一个类似Hadoop MapReduce的系统, 用户按照指定的接口实现一个任务,然后将这个任务递交给JStorm系统,Jstorm将这个任务跑起来,并且按7 * 24小时运行起来,一旦中间一个worker 发生意外故障, 调度器立即分配一个新的worker替换这个失效的worker。因此,从应用的角度,JStorm 应用是一种遵守某种编程规范的分布式应用。从系统角度, JStorm一套类似MapReduce的调度系统。 从数据的角度, 是一套基于流水线的消息处理机制。实时计算现在是大数据领域中最火爆的一个方向,因为人们对数据的要求越来越高,实时性要求也越来越快,传统的Hadoop Map Reduce,逐渐满足不了需求,因此在这个领域需求不断。
新特性:
- 1.5~6x performance boost from worst to best scenarios compared to JStorm-2.1.0
- Add application-level auto-batch
- Add independent control channel to separate control msgs from biz msgs to guarantee high priority for control msgs
- Dramatic performance boost in metrics, see "Improvements" section
- Support jdk1.8
- Add Nimbus hook and topology hook
- Metrics system:
- Support disable/enable metrics on the fly
- Add jstorm metrics design docs, see JSTORM-METRICS.md
- JStorm web UI:
- Add zookeeper viewer in web UI, thanks to @dingjun84
- Add log search and deep log search, support both backward search and forward search
- Support log file download
- Support changing log level on the fly
- Change error structure in zk, add errorLevel, errorCode and duration.
- Add supervisor health check
- Add -Dexclude.jars option to enable filtering jars manually
Bug 修复:
- Fix deadlock when starting workers
- Fix the bug that when localstate file is empty, supervisor can't start
- Fix kryo serialization for HeapByteBuffer in metrics
- Fix total memory usage calculation
- Fix the bug that empty worker is assigned when configured worker number is bigger than the actual number for user defined scheduler
- Fix UI log home directory
- Fix XSS security bug in web UI
- Don't start TopologyMetricsRunnable thread in local mode, thanks to @L-Donne
- Fix JSTORM-141, JSTORM-188 that TopologyMetricsRunnable consumes too much CPU
- Remove MaxTenuringThreshold JVM option support jdk1.8, thanks to @249550148
- Fix possible NPE in MkLocalShuffer
下载地址:
- jstorm-2.1.1.tgz
- Source code (zip)
- Source code (tar.gz)
|
|
|