new restore parallelization settings

[mwang1026]

Michael Wang (mwang1026) commented:

https://cockroachlabs.atlassian.net/browse/CRDB-6406

Jira Issue: DOC-1500

[katmayb]

new restore parallelization settings
Release note (performance improvement): This change adds a knob
(kv.bulk_io_write.restore_node_concurrency) to control the concurrency
per restore worker. This enables more concurrent calls to AddSSTable,
which will be bottlenecked by the cluster setting
kv.bulk_io_write.concurrent_addsstable_requests.

Per Liv's spreadsheet, this setting isn't being documented for 21.2, but for 22.1 dev docs.

[katmayb]

I'm moving this into the 22.1 milestone and label. Feel free to change if that's not correct @livlobo