diff --git a/docs/jstorm-doc/.jekyll-metadata b/docs/jstorm-doc/.jekyll-metadata deleted file mode 100644 index 11916c584..000000000 Binary files a/docs/jstorm-doc/.jekyll-metadata and /dev/null differ diff --git a/docs/jstorm-doc/ProgrammingGuide_cn/AdvancedUsage/Theory/Acker.md b/docs/jstorm-doc/ProgrammingGuide_cn/AdvancedUsage/Theory/Acker.md index 5a2907da1..cc2a205d1 100644 --- a/docs/jstorm-doc/ProgrammingGuide_cn/AdvancedUsage/Theory/Acker.md +++ b/docs/jstorm-doc/ProgrammingGuide_cn/AdvancedUsage/Theory/Acker.md @@ -27,7 +27,7 @@ JStorm的acker机制,能够保证消息至少被处理一次(at least once spout emit了一条消息,发送至bolt。bolt作为最后一个处理者,没有再向下游emit消息。
- +
从上图可以看到,所有的ack消息都会发送到acker,acker会根据算法计算从特定spout发射出来的tuple tree是否被完全处理。如果成功处理,则发送\__acker_ack消息给spout,否则发送\__acker_fail消息给spout。然后spout中可以做相应的逻辑如重发消息等。 @@ -67,7 +67,7 @@ Map _anchorsToIds; 即spout同时向bolt1和bolt2发送消息,它们处理完后,都向bolt3发送消息。bolt3没有后续处理节点。
- +
1). spout发射一条消息,生成root_id,由于这个值不变,我们就用root_id来标识。 diff --git a/docs/jstorm-doc/img/programguide/AdvancedUsage/acker_1.png b/docs/jstorm-doc/img/programguide/AdvancedUsage/acker_1.png new file mode 100644 index 000000000..19299637b Binary files /dev/null and b/docs/jstorm-doc/img/programguide/AdvancedUsage/acker_1.png differ diff --git a/docs/jstorm-doc/img/programguide/AdvancedUsage/acker_2.png b/docs/jstorm-doc/img/programguide/AdvancedUsage/acker_2.png new file mode 100644 index 000000000..60ba13d36 Binary files /dev/null and b/docs/jstorm-doc/img/programguide/AdvancedUsage/acker_2.png differ