FE节点（leader）服务挂掉

[spihiker]

Version
doris版本：1.0.0（二进制包）
操作系统：centos7.9_x86_64

问题描述：
集群扩容时，添加be或fe节点时都遇到了follower(leader)节点服务挂掉。

错误日志：
2022-04-28 09:24:02,486 ERROR (heartbeat mgr|22) [BDBJEJournal.write():176] write bdb failed. will exit. journalId: 200667, bdb database Name: 183223
[2022-04-28 09:24:02] write bdb failed. will exit. journalId: 200667, bdb database Name: 183223
2022-04-28 09:24:02 INFO AbstractConnector:381 - Stopped ServerConnector@67eeb310{HTTP/1.1, (http/1.1)}{0.0.0.0:8030}
2022-04-28 09:24:02 INFO session:149 - node0 Stopped scavenging
2022-04-28 09:24:02 INFO application:2347 - Destroying Spring FrameworkServlet 'dispatcherServlet'
2022-04-28 09:24:02 INFO ContextHandler:1153 - Stopped o.s.b.w.e.j.JettyEmbeddedWebAppContext@4cc26df{application,/,[file:///tmp/jetty-docbase.8030.6754360883789190151/],STOPPED}
2022-04-28 09:24:02 INFO HikariDataSource:110 - HikariPool-1 - Starting...
2022-04-28 09:24:02 WARN DriverDataSource:70 - Registered driver with driverClassName=org.apache.derby.jdbc.EmbeddedDriver was not found, trying direct instantiation.
2022-04-28 09:24:02 INFO PoolBase:541 - HikariPool-1 - Driver does not support get/set network timeout for connections. (Feature not implemented: No details.)
2022-04-28 09:24:02 INFO HikariDataSource:123 - HikariPool-1 - Start completed.
2022-04-28 09:24:02,841 INFO (tablet scheduler|29) [BeLoadRebalancer.selectAlternativeTabletsForCluster():85] cluster is balance: default_cluster with medium: HDD. skip
2022-04-28 09:24:02,841 INFO (tablet scheduler|29) [BeLoadRebalancer.selectAlternativeTabletsForCluster():85] cluster is balance: default_cluster with medium: SSD. skip
2022-04-28 09:24:02,841 INFO (tablet scheduler|29) [BeLoadRebalancer.selectAlternativeTabletsForCluster():85] cluster is balance: default_cluster with medium: S3. skip
2022-04-28 09:24:03 INFO HikariDataSource:350 - HikariPool-1 - Shutdown initiated...
2022-04-28 09:24:03 INFO HikariDataSource:352 - HikariPool-1 - Shutdown completed.

[spihiker]

这里问题。一直没人回复吗？

[dataalive]

这个我们来复现下，就是普通的添加节点，没有任何其他的操作对吧。

[spihiker]

你好。是的。就是开始只有一个fe节点和be节点。然后增加了1个fe和be节点时，遇到的。我已经自己重新编译了一个包。我再试试，看看有没有问题？

[spihiker]

昨天又遇到了同样的问题。添加非leader follower时，leader follower节点服务宕掉。我启服务之前，把时钟同步了下。就没问题了。但同步时钟时，发现时差也就都在5秒内。不知道跟这个有没有关系？如果方便时，可以自己试下看看。谢谢！

[hf200012]

你加我微信：35926237，我来看看

[spihiker]

已添加。我自己编译了个包。我再试试看。要还有问题，再请教你。谢谢了！

[hf200012]

Did you modify the source code of fe yourself and add the Hikari connection pool?
doris does not use the Hikari connection pool, and there is no r of this connection pool in the startup log.