问题描述
使用EFK(Elasticsearch, Fluentd and Kibana)在收集日志的解决方案中, 可以先把日志发送到EventHub中,然后通过Logstash消费EventHub中的事件并分发出去。但是在使用Logstash的过程中,遇见了连接不上EventHub的错误,和Receiver一直不停关闭的问题。 错误的信息分别为:
Exception while initializing stores, not starting partition manager com.microsoft.azure.eventhubs.IllegalEntityException: Failure getting partition ids for event hub ... ... Caused by: com.microsoft.azure.eventhubs.TimeoutException: Opening MessagingFactory timed out. |
[WARN ][com.microsoft.azure.eventprocessorhost.PartitionPump][main][cbc2dac224225cd02511820a8ee314e73f1c0800809c9c534154188acb14fbac] host logstash-fe4f6e2e-e260-4522-a3f8-f292a8902dad: 3: Receiver disconnected on create, bad epoch? |
问题原因
- 对于TimeoutException,需要判断是否是当前环境连接不上EventHub服务器,所以可以通过排查网络连接的方式来解决。(Link)
- 对于ReceiverDisconnectedException,错误是Receiver在不停通过同一个消费组,同一个分区建立连接,当新连接建立时,会导致旧的连接关闭。所以需要检查客户端是不是又多个进程在建立连接或者时多个客户端在消费同一个分区数据
具体的解释可以参考:https://github.com/Azure/azure-event-hubs-spark/blob/master/FAQ.md
Why am I getting a
ReceiverDisconnectedException
?In version 2.3.2 and above, the connector uses epoch receivers from the Event Hubs Java client. This only allows one receiver to be open per consumer group-partition combo. To be crystal clear, let's say we have
receiverA
with an epoch of0
which is open within consumer groupfoo
on partition0
. Now, if we open a new receiver,receiverB
, for the same consumer group and partition with an epoch of0
(or higher), thenreceiverA
will be disconnected and get theReceiverDisconnectedException
.In order to avoid this issue, please have one consumer group per Spark application being run. In general, you should have a unique consumer group for each consuming application being run.
Note that this error could happen if the same structured stream is accessed by multiple queries (writers).
Spark will read from the input source and process the dataframe separately for each defined sink. This results in having multiple readers on the same consumer group-partition combo. In order to prevent this, you can create a separate reader for each writer using a separate consumer group or use an intermediate delta table if you are using Databricks.
解决方案
对于TimeoutException问题,只要解决另外客户端环境问题后,问题会得到解决。但是对于ReceiverDisconnectedException则如何解决呢? 由于都是在Logstash中配置,并没有代码可以修改。所以解决这个问题就是要设置Logstash的工作进程,不能让进程数大于分区数。 并且为Logstash在EventHub中单独建立一个消费组。以下是为一个成功通过Logstash消费EventHub的配置
input {
azure_event_hubs {
event_hub_connections => ["Endpoint=sb://xxxx.servicebus.chinacloudapi.cn/;SharedAccessKeyName=test;SharedAccessKey=xxxxxxxx=;EntityPath=logstest"]
threads => 8
decorate_events => true
consumer_group => "logs"
storage_connection => "DefaultEndpointsProtocol=https;AccountName=xxx;AccountKey=xxxxxxx=;EndpointSuffix=core.chinacloudapi.cn"
}
} output { stdout {
}
}
启动命令为:
./bin/logstash -f config/ehtest.conf -w 1
- -f 后面的参数为logstash的配置文件
- -w 后面的参数1表示只启动一个worker,默认情况为当前CPU的核数。(https://www.elastic.co/guide/en/logstash/master/logstash-settings-file.html#logstash-settings-file)
启动后成功结果如:(成功捕获到EventHub中进入的事件)
在获取连接字符串的过程中,可以参考Logstash中关于EventHub插件的说明文档:https://www.elastic.co/guide/en/logstash/current/plugins-inputs-azure_event_hubs.html(如要了解全面的EventHub中的配置参数,也可以参考该文档说明)
The plugin uses the connection string to access Azure Events Hubs. Find the connection string here: Azure Portal
-> Event Hub -> Shared access polices
. The event_hub_connections option passes the Event Hub connection strings for the basic configuration.
注:不要使用默认的消费组或者是与其他应用公用消费组,这样会导致Logstash连接不上。