Thanks for the response, Tim. This had been troubling us for a while, but I
think we finally figured it out. It was indeed a port conflict, but not
with another ZMQ process. It turns out our ephemeral port range was messed
up on the machines:

$ cat /proc/sys/net/ipv4/ip_local_port_range
1024    65000

​
We fixed that, and now everything's great.  Hopefully this will help
someone else down the road, because other threads on this error were dead
when I looked.

Thanks again!

Keith.

On Wed Jan 14 2015 at 12:21:40 PM Tim Molter <[email protected]> wrote:

> Hi Keith,
>
> Some ZeroMQ process is trying to start up, but some other ZeroMQ process
> is already using the port, and it could have entered into an unrecoverable
> state. To fix it, I always restarted everything.
>
> ~Tim
>
> On 2015_01_14 5:41 PM, Keith Bourgoin wrote:
> > Hey Everyone,
> >
> > This will be a blast from the past, but some problems have prevented us
> > from upgrading to 0.9.x at the moment.
> >
> > The problem we're seeing is ZMQ trying to use an address already in use:
> >
> > |2015-01-14 04:39:38 util [ERROR] Async loop died!
> > org.zeromq.ZMQException: Address already in use(0x62)
> >     at org.zeromq.ZMQ$Socket.bind(Native Method)
> >     at zilch.mq <http://zilch.mq/>$bind.invoke(mq.clj:69)
> >     at backtype.storm.messaging.zmq.ZMQContext.bind(zmq.clj:57)
> >     at backtype.storm.messaging.loader$launch_receive_thread_
> BANG_$fn__1629.invoke(loader.clj:26)
> >     at backtype.storm.util$async_loop$fn__465.invoke(util.clj:375)
> >     at clojure.lang.AFn.run(AFn.java:24)
> >     at java.lang.Thread.run(Thread.java:745)
> > |
> >
> >
> > I googled around and couldn't find anything. It looks like this tends to
> > affect some workers on a machine, but not all. Those workers are
> > functionally dead, and eventually Storm will switch to using a different
> > workers. However, as our machines get more full, these get hit more
> > often, and the time to rebalance usually means failed tuples.
> >
> > I know it's an old version, but any help would be greatly appreciated.
> >
> > Thanks,
> >
> > Keith.
>
>

Reply via email to