fix(carddav): don't read after write for address book creation

[miaulalala]

Summary

In clustered setups, a read after write will sporadically fail. We have all the data, so we don't need to query the backend again.

Checklist

• Code is properly formatted
• Sign-off message is added to all commits
• Tests (unit, integration, api and/or acceptance) are included
• Screenshots before/after for front-end changes
• Documentation (manuals or wiki) has been updated or is not required
• Backports requested where applicable (ex: critical bugfixes)

[ChristophWurst]

Does this address #41996?

[ChristophWurst]

👍 makes a lot of sense

[miaulalala]

Does this address #41996?

I'm not sure, but it might help.

[ChristophWurst]

On a second thought I'm not sure this does help. The INSERT+SELECT is already wrapped in a transaction. A read/write split should still execute both queries on a primary node. Read after write is only problematic outside of transactions.

[ChristophWurst]

See nextcloud/groupware#24. Mariadb experts have told me to either

1. Avoid the read
2. Wrap write+read in a transaction

The address book creation does the latter since #34412.

Did you see this error on a productive instance? Does it run 25?

[miaulalala]

right, this is superfluous then