TPC-C Large Doc Bug & Implementation Bug

Richard Lewis commented:

If this should be separated into two Issues/tickets, please advise.

Doc Bug:
https://www.cockroachlabs.com/docs/v21.2/performance-benchmarking-with-tpcc-large.html
is missing Step 6.
There is no documentation on how best to start CockroachDB in an environment where more than one cloud service provider(CSP) is used with a single database.
For example, if each of the three CSPs(AWS, GCP, Azure) are defined region flag/list of the "workload fixtures import tpcc" command, then no more than 3 nodes can be deployed because the value of the partition would fail/be incorrect.

Implementation Bug:
The below commands apply to a 3 CSP(AWS, GCP, Azure) database where each CSP has two nodes in a single region/zone.
e.g. cockroach start \
--certs-dir=certs \
--advertise-addr=34.202.166.149 \
--join=35.193.167.5,34.68.89.219,54.221.40.0,34.202.166.149,20.124.24.12,13.68.148.116 \
--cache=.25 \
--max-sql-memory=.25 \
--background \
--locality=region=aws,zone=us-east-1c

nohup cockroach workload fixtures import tpcc --partitions=3 --warehouses=10000 --replicate-static-columns --partition-strategy=leases 'postgres://root@54.221.40.0:26257?sslmode=verify-full&sslrootcert=certs/ca.crt&sslcert=certs/client.root.crt&sslkey=certs/client.root.key' --survival-goal=region --regions=aws,gcp,azure

I211210 17:49:52.810033 1 ccl/workloadccl/fixture.go:345 [-] 1 starting import of 9 tables
I211210 17:49:53.736660 75 ccl/workloadccl/fixture.go:478 [-] 2 imported 7.8 MiB in item table (100000 rows, 0 index entries, took 924.52693ms, 8.41 MiB/s)
Error: importing fixture: importing table warehouse: pq: at or near "regional": syntax error

Using this workload statement:
nohup cockroach workload fixtures import tpcc --partitions=6 --warehouses=10000 --replicate-static-columns --partition-strategy=leases ‘postgres://root@54.221.40.0:26257?sslmode=verify-full&sslrootcert=certs/ca.crt&sslcert=certs/client.root.crt&sslkey=certs/client.root.key’

I received this error:
Error: Could not postload: could not partition tables: Couldn’t exec “ALTER PARTITION p0_0 OF TABLE warehouse CONFIGURE ZONE USING num_replicas = COPY FROM PARENT, constraints = ‘{\“+rack=0\“:1}‘, lease_preferences = ‘[[+rack=0]]’“: pq: constraint “+rack=0” matches no existing nodes within the cluster - did you enter it correctly?

Jira Issue: DOC-2141