About Enterprise backup/restore/replicate on Innodb cluster environment(s)

[Liang Liang]

Hi, Team
We have one (community edition) MySQL innodb cluster, including one primary server and 2 slave servers. In the past one year, the system may hit IO issue about several times, and the IO issue cause innodb cluster down, sometimes 1 slave is down, while sometimes 2 slaves are down, or master is switched. The most serious situation we met is that, slave(s) could not recovered by restart mysql/group_replication process. To cover it, we spent several hours, that is stop the whole environment first, and then
1. Removing all binlogs from primary server
2. copy all left files from primary server to other slave servers.
3. Start previous primary server, and then run some sqls to let it run as primary again
4. start group_replication on slave nodes

In our environment, the total size of datafiles is more than 200G. So step #2 takes very long time, more than 1 hour, or 2 hours, it depends on one or two slave nodes need to recover.

In feture, we may setup another innodb cluster, lets say current one is env1, and the new one is called env2. We need to replicate these two environments. At first, we will copy the data from env2, and every one month, we need to sync the data from env1 to env2 again. The first direct solution is that we make a new full copy from env1, stop env2, remove all files from env2, and copy all data to env2, start env2. But it will take more than 2 hours everytime, is not a good solution.

We notice that MySQL enterprise edition supports incremental backup, and we wonder if it works with innodb cluster environment, and if yes, will it speed up the recover process on slave nodes, and the sync process between env1 and env2? Thanks