Ack, I was getting this error "Netty-ip-10....", which made me really confused. I thoght it was a topology issue. Well, turned out it was a stupid mistake. Last minute change to the JVM setting, and I used "2048" instead of "2048m" (meg). Nothing in the log was popping out at me indicating that there wasn't enough memory assigned to java, until I ran the java command on the command line myself, from the ps output. Then I saw the out of memory error. Ah well. Only 3 hours wasted. :-O
On Fri, Oct 23, 2015 at 3:49 AM, John Yost <[email protected]> wrote: > Hi Dillian, what does the nimbus log indicate? > > --John > > On Thu, Oct 22, 2015 at 9:27 PM, Dillian Murphey <[email protected]> > wrote: > >> Tested this in my dev environment, but prod behaved differently. >> >> When I added a new worker node in prod, I assumed it would sit there with >> no running jobs until I ran rebalance. Instead, as soon as the server came >> up, jobs were sent to it, and then my entire topology stopped working. >> >> Any clues at all? Very frustrated. :( >> > >
