chore: update dbcopyall8 command

[mumarkhan999]

The previous dbcoopyall8 command causing operational errors

The occurrence of the operational errors: table already exists issue can be traced to the following:

• During the provisioning process, the mysql80 container is populated with the most recent data, (new migrations/tables).
• When data is transferred from mysql57 using dbcopyall8 command, the newly created tables persist in the mysql80 container. But, the django_migrations table becomes outdated in this process.
• Consequently, the records for any new migration are removed from the django_migrations table. This leads to a failure when attempting to execute dev.migrate.

This PR is to fix the above-mentioned issue.
It will wipe out all the data of mysql8 before actually dumping mysql57 data inside mysql80 container.

[timmc-edx]

The copying is now failing for me, with the following error:

docker compose exec mysql57 mysqldump "credentials" > .dev/credentials.sql
docker compose exec -T mysql80 mysql "credentials" < .dev/credentials.sql
ERROR 2013 (HY000) at line 621: Lost connection to MySQL server during query
make[1]: *** [Makefile:481: dev.dbcopy8.credentials] Error 1

If I dump the logs just before this statement runs, I get a line about "Temporary server started." This temporary server is started and then stopped before the regular mysql server runs. I don't see this in the mysql57 logs. This might mean that mysql80 needs a different availability check.

[timmc-edx]

If I move the sleep 10 to below the wait-for line, it works. Something is still very strange here, because I can run various checks against MySQL immediately between the provision-mysql80.sql and _db_copy8_targets lines and they always pass, even when run repeatedly in short succession.

[timmc-edx]

OK, I think I've got it! The problem is that the mysql command is talking to the temporary server over a UNIX socket. The solution seems to be to tell mysql to use TCP. I'll submit a separate PR for that, though.

[mumarkhan999]

Closing this to re-run openedx/cla. It's stuck for an unknown reason.