FreshnessBasedConsumptionStatusChecker never pass when a table got deleted during the restart.

Below are the log reference:
```
2024/02/08 01:19:09.149 INFO [BaseServerStarter] [Start a Pinot [SERVER]] Sleep for 10000ms as service status has not turned GOOD: MultipleCallbackServiceStatusCallback:IdealStateAndCurrentStateMatchServiceStatusCallback:None;IdealStateAndExternalViewMatchServiceStatusCallback:None;RealtimeConsumptionCatchupServiceStatusCallback:Waiting for consuming segments to catchup: numConsumingSegmentsNotCaughtUp=1, timeRemaining=13928231ms;;PinotServiceManagerStatusCallback:Started;

2024/02/08 01:19:09.802 INFO [FreshnessBasedConsumptionStatusChecker] [grizzly-http-server-11] TableDataManager is not yet setup for segment mytable__19__224__20240207T1805Z. Will check consumption status later
```

Start with FreshnessBasedConsumptionStatusChecker and the provided startup log, focusing on the segment whose table is deleted during restart. Trace how the checker handles a missing TableDataManager and verify the status check no longer prevents startup or waits indefinitely when the table has been removed.

Written by the indexing model from the issue text.