Best architecture

Ask Bjørn Hansen ask at
Thu Jul 12 17:53:27 UTC 2007

On Jul 12, 2007, at 10:51 AM, Matt Sergeant wrote:

> What are the other potential failure conditions - we're getting  
> on_fail
> called, but immediately sometimes rather than after the timeout.

And you are sure the workers don't return a failure immediately  
sometimes?   :-/   It sounds like there are multiple failures going  

IIRC if it can't connect or loses the connection to the gearmand  
it'll do that.

Would it be possible to run
    tcpdump -nnXSs 0 "port $gearmand"
on some host that might capture the problem?

  - ask

-- -

More information about the Gearman mailing list