Recently doing Galera cluster upgrading and encounter A problem, A, B, C three nodes are running, the data is consistent, close the B, C node, make A backup, upgrade B, C mairiadb version, shut down A node at this moment, at this point A and B, C have incremental gap, good backup of A database, complete A upgrade, open the first node A, then open B \ C nodes to join the cluster, why this time B, C do SST not IST, because the data is more, I am afraid that when the time comes A slow sync lead to influence business,
A node of the error is:
The 139985571683072 3:32:14 2019-05-16 [Note] WSREP: IST request: a4f45219 e9 - b715-8627-7704-11 b47985c6:448-458 | TCP://172.21.2.82:4568
The 139985571683072 3:32:14 2019-05-16 [Note] WSREP: IST first seqno 449 not found from the cache, falling back to SST

Has been said that the seqno not found in cache, set up a 1 g I cache, didn't also several incremental transactions, other nodes should do IST in theory,
The man has encountered such a problem? For help!