Cause: The company has a strategic alliance with Huawei, Huawei needs to migrate applications to the cloud, the migration process application container is connected life and death Rom mongo, mongo, what are the security group is not added, the phenomenon as follows
Mongo standby node configuration information exists for 10.125.80.1:8635,10.125.80.2:8635 where zk, zk program reads the configuration to connect when there are mistakes mongo
Problem Description:
As shown above, the correct read zk in the two segments configured to 10.125 correct ip, then look at the log found Nani? 192.168 mongo master how to become a segment from the hosts? how is it? Hell? Then I checked all the configuration could not find any place to configure 192.168, you can only go to the company's operation and maintenance colleagues,
First find the operation and maintenance A: "I do not know how it was so strange is not your configuration wrong???" So ... I went back again checked the configuration, the code did not find the problem.
Then find O & M B: "is not your configuration wrong?" I firmly believe I have to say that the investigation did wrong before. Then a moment to check the operation and maintenance of information security group, also did not find the Editor's Note. Well, then, can only rely on their own track about the mongo log information according to the source code it so a piece of code found in the driver's jar mongo
After the connection is created when initializing mongo daemon thread then connects execute ismaster command to get mongo service information in the present, to see this code is feeling rush to see the hope of Huawei console to perform a bit this command returns the following results
192.168 mongo original drive will return here to get this result is returned to resolve whether the hosts and then compare initialized to address if it is not then remove the log in information remove information in the hosts inside.
然后我拿着我的"证据",就是上面这个截图去找运维了 "这么诡异？ 这个ip是啥 ？ 没配置这个ip啊"运维表示也没头绪... so 然后就联系了华为的运维人员，然后在群里双方的运维balabala了一大堆之后，华为表示 之前确实也遇到过这个问题，192.168的网段是华为的同步ip,我们配置的是业务ip ,容器里面想要连上 需要华为同事手动在后台操作才可以。
然后就提供我们的服务器信息给华为同事后台配置完之后果然连上了！！！ 坑已填上..
这里要夸一下华为，华为的服务态度和响应速度还是挺棒的，华为表示这个问题后续会优化。 当然也得谢谢我们的运维同事的跟进。
ok！