Error converting anonymous replica to regular

Bug description

We added an anonymous replica to the replica set as a workaround for a known issue where adding regular replicas fails in a cluster with synchronous transactions enabled. When attempting to convert this anonymous replica to a regular replica, the operation failed with errors.

tarantool version | 3.4.1-0-g096322fad

Steps to reproduce

Start anon replicas with next config
replication:
anon: true
election_mode: off

Reconfigure replicas
replication:
anon: false
election_mode: voter

WA and Example:
Procedure for removing old replica and adding anonymous replica

1. Enable automatic expulsion in the global configuration
   Set the following parameters under replication at the global level:

replication:
  autoexpel:
    by: prefix
    enabled: true
    prefix: 'your_prefix'

2. Remove old replica

3. Remove old replica from config

4. Deploying new replica with next configuration

replication:
  anon: true
  election_mode: off

5. Explicitly specify replication peers for all existing nodes
   Example for the first candidate node:

          tdb_storage-1-2:
            iproto:
              advertise:
                client: 10.181.64.13:3404
              listen:
              - uri: 10.181.64.13:3404
            memtx:
              memory: 536870912
            replication:
              election_mode: candidate
              peers:
                - replicator:i-changed-a-password-here@10.181.8.1:3403 (replica candidate)
                - replicator:i-changed-a-password-here@10.181.64.13:3404 (master candidate)
            roles_cfg:
              roles.metrics-export:
                http:
                - endpoints:
                  - format: prometheus
                    path: /metrics
                  listen: 8404

Example for the second candidate node

          tdb_storage-1-1:
            iproto:
              advertise:
                client: 10.181.8.1:3403
              listen:
              - uri: 10.181.8.1:3403
            memtx:
              memory: 536870912
            replication:
              election_mode: candidate
              peers:
               - replicator:i-changed-a-password-here@10.181.8.1:3403
               - replicator:i-changed-a-password-here@10.181.64.13:3404 
            roles_cfg:
              roles.metrics-export:
                http:
                - endpoints:
                  - format: prometheus
                    path: /metrics
                  listen: 8403

6. Explicitly define the UUID for the new replica in its configuration
   Example for the new replica:

          tdb_storage-1-3:
            iproto:
              advertise:
                client: 10.181.64.13:3405
              listen:
              - uri: 10.181.64.13:3405
            database:
              instance_uuid: 1c32c157-7bb7-4fd6-94ce-de5a97cab588
            memtx:
              memory: 536870912
            replication:
              anon: true
              election_mode: off
            roles_cfg:
              roles.metrics-export:
                http:
                - endpoints:
                  - format: prometheus
                    path: /metrics
                  listen: 8405

7. Change the parameters of the added replica
   Update configuration to:

replication:
  anon: false
  election_mode: voter

8. Fix the missing instance name warning box_cfg.apply: name tdb_storage-1-3 for b7e7e85a-e2cc-40d0-8cb1-a8454ec5919d uuid is missing from the snapshot. It will be automatically set when possible.

10.181.64.13:3404> box.space._cluster:fselect()
---
- |-
  +-----+--------------------------------------+-----------------+
  | id  |                 uuid                 |      name       |
  +-----+--------------------------------------+-----------------+
  |  1  |"8a07b2ee-04af-4d67-bddd-b797bf1f1693"|"tdb_storage-1-2"|
  |  2  |"b7e7e85a-e2cc-40d0-8cb1-a8454ec5919d"|                 |
  |  3  |"dd33fa06-d06f-4146-a5cf-226a201d4f01"|"tdb_storage-1-1"|
  +-----+--------------------------------------+-----------------+
...

10.181.64.13:3404> box.space._cluster:update(2, {{'=', 'name', 'tdb_storage-1-
3'}})
---
- [2, 'b7e7e85a-e2cc-40d0-8cb1-a8454ec5919d', 'tdb_storage-1-3']
...

No source file or test is identified. Start by reproducing the anonymous-to-regular replica transition with the configurations and cluster setup described, then inspect the replication reconfiguration path and the _cluster name update shown in the report. Done means the replica becomes a voter without errors or a missing-name warning.