add truncate support for a index

[easyice]

when i want to change type for some filed in mapping, i should reindex or bulk write to a new index，but，in my case，create an index can be costly，because：

1. i have more than 100 index need to reindex from remote every mouth
2. my cluster has 50 nodes and 10000 shards, when create a new index, the shards-stared RPC need much more time for master node processing, the rpc has URGENT priority，He caused the other rpc to have no chance to execute, like put-mapping.

i want to reuse the old index,Don't delete it.if the truncate api can be support， i can truncate then old index, including then all index segments, and mappings, then put new mapping, and write index to the old index

thanks

[elasticmachine]

Pinging @elastic/es-core-features (Team:Core/Features)

[DaveCTurner]

@easyice what version are you using? In particular if it's before 7.4.0 then you are missing #44433 which considerably speeds up the processing of shard-started events in high-shard-count clusters.

[easyice]

thanks for replay，my cluster is 6.8.0, in this version the one shard-started event need over 50 seconds for processing done, the cluster has 50000 shards.

let me test how much fast on 7.4.0, thanks a lot

[easyice]

@DaveCTurner I tested on 7.6.0, #44433 and #47817 can solve the problem，so this issues can be closed