[BucketDurability]: Updating bucket_level_durability with diag_eval turns bucket into stale state

Build: 6.6.0 - 7687 **

Scenario:

1. Two node cluster, turn off auto_failover
2. Create Couchbase bucket(replica=1, bucket_durability=MAJORITY)
3. Load 10K docs into the bucket
4. Update bucket_durability=majorityAndPersistActive (ns_bucket:update_bucket_props)
5. Restart kv_servers after diag_eval update (/bin/systemctl restart couchbase-server.service)

Observation:

• After restarting servers, bucket turns into "amber" color stale state.
• Also seeing compactor error from UI logs continuously.
• Bucket_min_durability has not been updated.

Cb server logs:
https://cb-jira.s3.us-east-2.amazonaws.com/logs/bucket_d_update_issue/collectinfo-2020-05-28T115720-ns_1%4010.112.191.101.zip
https://cb-jira.s3.us-east-2.amazonaws.com/logs/bucket_d_update_issue/collectinfo-2020-05-28T115720-ns_1%4010.112.191.102.zip