Couchbase failover cluster fault automatic transfer scheme research!

Recently fascinated Couchbase, now all the site trial couchbase to cache and persistent, so it seems that the risk is relatively large ah, the cache server is dead.

It's easy to see a cluster scheme that talks about Couchbase, and then follow the tutorial. Reference article (http://www.cnblogs.com/long-gengyun/p/3772504.html)

Elder brother used two servers, respectively deployed Couchbase service side, and simulated a server hangs, see is the cluster's failure automatic transfer function is effective, testing, found dead or alive, check all the information at home and abroad, Chinese is too little, English can not understand, hard to die brother.

At least the effort is not negative, finally found the problem lies. Share with you here!

1, in the blog Park accidentally saw a piece of article, found that there is a brother of the server individual port is not open, was blocked by the firewall. If you want to implement a cluster, you must turn on all ports

Reference article http://www.cnblogs.com/wucg/p/3450863.html

2, if you want to achieve automatic fault transfer, it seems to require at least 3 node server (English is not good, we can see it)

Related documents: http://docs.couchbase.com/admin/admin/Tasks/tasks-nodeFailover.html

3, the effect of testing, hehe

Finally said two, the elder brother spends nearly 3 days to ponder Couchbase's cluster and the failover plan, the middle difficulty, but fortunately did not give up, finally solves the problem.

Then the problem came, brother busy to Busy, you reckoned write useful words, give a praise Bai!

Wish you a happy New Year!

Couchbase failover cluster fault automatic transfer scheme research!