Menu

#1528 Daemon becomes non-responsive upon comms loss

open
nobody
5
2010-06-22
2010-06-22
No

If you lose connectivity to the monitored shelf(ves) after discovery, either because of a fault of the switches connecting the machine the daemon resides on to the monitored systems, or if the shelf manager or similar is powered off or fails, the daemon quits responding to callers and causes them to hang if there is a query pending against the daemon. This state continues until the connectivity is restored to the IPMI server(s) in question. ipmidirect appears to be trying to re-establish communication by the ping/pong process, followed by a re-authentication against the server. Not sure if this is a plugin problem or a daemon problem because of that. High-Availability middleware shouldn't hang like this.

Discussion

  • Lars Wetzel

    Lars Wetzel - 2010-06-22

    Please don't assign issues with the impi direct plugin to me.
    I don't have any hardware for verification.

     
  • Lars Wetzel

    Lars Wetzel - 2010-06-22
    • assigned_to: larswetzel --> nobody