Always on troubleshoot

 Scenarios where the failover not happening in Always on.

1. Port number is not allowed 1433,5022 in firewall settings.

2. Event viewer no erros found.

3. Cluadmin- Check preffered nodes in node properties.

4. check if connect permission is not enabled  for service account.

5. check alteravailability group and serverstate permission for NTAuthority/system account  on both the nodes.  (why Ntauthority/system- coz healthcheck is authorised to this account)

Failover disconnected .

1. Check Endpoint algorithm on both the nodes which must be compatible i.e AES from 2016 version or rc4 from 2014 version.

2. Check nodes are up in cluadmin.

Failover is initiated and state is in resolving state on the secondary.

Investigate the sql error log and find if any errors, this may be due to login failed 

The account Ntauthority system must be available on  both the nodes , if not then create the account with grant alter any availability group to NTauthority system service account.

what happens when secondary server is offline and comes back online in synchrnous commit mode.

The state will be shown as not synchronizing in secondary and  the transactions are not hardened due to disconnected state. the primary will not wait for the acknowledgement as it changes to asynchronous commit mode . in this case there may be chances to log file size gets increased. 

As soon as the secondary comes online it establises conneciotn to primary replica sends the lsn details to primary and primary starts sends  log block to secondary while here the state changes to synchronzing in secondary until all the lsn are matched then the state changes to synchronised and commit mode is changed to synchronous commit mode.