connection timeout on a previously established connection

  • I have been seeing this in the error log file but can't seem to know/find out the root cause of this. This occurs very frequently but seem to resolve on its own within 10 seconds. This is all the info I have to go by but maybe I am looking in the wrong place. Any advise is highly appreciated.

    A connection timeout has occurred on a previously established connection to availability replica with id. Either a networking or a firewall issue exists or the availability replica has transitioned to the resolving role.

    "He who learns for the sake of haughtiness, dies ignorant. He who learns only to talk, rather than to act, dies a hyprocite. He who learns for the mere sake of debating, dies irreligious. He who learns only to accumulate wealth, dies an atheist. And he who learns for the sake of action, dies a mystic."[/i]

  • There's a blog post made by Will Assaf regarding this issue.

    take a look.

    https://www.sqlservercentral.com/blogs/a-connection-timeout-has-occurred-on-a-previously-established-connection-to-availability-replica

     

  • Could be a bug. Check this blog post on SSC and make sure you are on the correct patch:

    "A connection timeout has occurred on a previously established connection to availability replica"

    Sue

     

  • This is the patch level we are on.

    Microsoft SQL Server 2016 (SP2-CU10) (KB4524334) - 13.0.5492.2 (X64) 

    "He who learns for the sake of haughtiness, dies ignorant. He who learns only to talk, rather than to act, dies a hyprocite. He who learns for the mere sake of debating, dies irreligious. He who learns only to accumulate wealth, dies an atheist. And he who learns for the sake of action, dies a mystic."[/i]

  • Multisubnet had some issues with that error so you can search for that if you are using that.

    Virtual machines had some issues with that error and backups, I think it may have been specifically for snapshots. Also with some monitoring tools on VMs. But something to consider if this is virtual instead of physical.

    In general, try checking the always on extended events session when you get the error to see if you can find any additional information.

    Sue

  • I believe we have 4 servers (Node1, Node2, node3 and node4) in one subnet.  All physical servers. Environment details.

    Node1 act as a primary for 1st AG. Node 2, Node 3 and node 4 are secondary

    Node2 act as a primary for 2nd AG. Node 3 and node 4 are secondary

    Node3 act as a primary for 3rd AG. Node 2 and node 4 are secondary

    We also have storage issues on Node 4 because of the IO issue I see on a daily basis. I always see log send rate is greater than redo rate. We also run backups on Node 4 so this server is constantly busy.

    • This reply was modified 4 years, 10 months ago by  LearningDBA.
    • This reply was modified 4 years, 10 months ago by  LearningDBA.
    • This reply was modified 4 years, 10 months ago by  LearningDBA.
    Attachments:
    You must be logged in to view attached files.

    "He who learns for the sake of haughtiness, dies ignorant. He who learns only to talk, rather than to act, dies a hyprocite. He who learns for the mere sake of debating, dies irreligious. He who learns only to accumulate wealth, dies an atheist. And he who learns for the sake of action, dies a mystic."[/i]

  • Having disk issues isn't going to make it easy for you to address so you likely want to correct that. Have you checked the extended events session for AlwaysOn health? That's likely to give you more information.

    Sue

  • Sue_H, screenshots which I shared last week were from the Extended events session for the Always on health.

    "He who learns for the sake of haughtiness, dies ignorant. He who learns only to talk, rather than to act, dies a hyprocite. He who learns for the mere sake of debating, dies irreligious. He who learns only to accumulate wealth, dies an atheist. And he who learns for the sake of action, dies a mystic."[/i]

  • LearningDBA wrote:

    Sue_H, screenshots which I shared last week were from the Extended events session for the Always on health.

    Looking at one second of a log, extended events session is often not enough to find potential issues. I don't think there is enough info here for anyone to say whatever the issues are you may be having other than the guesses we already provided. I'd suggest spending more time going through all the logs and extended events pretty thoroughly.

    Sue

Viewing 9 posts - 1 through 8 (of 8 total)

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