Endopoint stop alwayson during X time

  • Hello everyone ,

    I just discovered that the databases in my Alwayson cluster are in NON SYNCHRONIZED status I tested Endpoint I found it in status stoped

    select name, state_desc, port FROM sys.tcp_endpoints where name='hadr_endpoint'

    go

    I launched Endpoint to start it again

    alter endpoint [Hadr_endpoint] state = started

    it has gone well and the endpoint has started well i would like to have now during the period where the endpoint was one stop can I have lost data? thank you for your explanations

    thank you

     

  • during the period where the endpoint was one stop can I have lost data?

    You will not lose data when the Primary Replica cannot communicate with a Secondary Replica. The Primary continues normal operations. Transactions that commit on the Primary remain in the transaction log until they can be replicated to all Secondaries. It works just like when a Secondary is unavailable during a reboot.

    In the worst case, a Secondary may fall behind to the point where you would prefer to drop it and re-create from a fresh backup of the Primary, but the Primary will hold all data captured while the Secondary was unavailable.

     

    Eddie Wuerch
    MCM: SQL

  • Eddie Wuerch wrote:

    during the period where the endpoint was one stop can I have lost data?

    You will not lose data when the Primary Replica cannot communicate with a Secondary Replica. The Primary continues normal operations. Transactions that commit on the Primary remain in the transaction log until they can be replicated to all Secondaries. It works just like when a Secondary is unavailable during a reboot. In the worst case, a Secondary may fall behind to the point where you would prefer to drop it and re-create from a fresh backup of the Primary, but the Primary will hold all data captured while the Secondary was unavailable.  

    Adding something

    If im not mistaken, your primary database transaction log file will not truncate the VLFS (even with T-log backups) till the transactions are replicated, this could make your transaction log grow in size.

     

Viewing 3 posts - 1 through 2 (of 2 total)

You must be logged in to reply to this topic. Login to reply