Interestingly, if I increase the sleep period between random http messages I still get the null pointer exception. I'm at 3000 milliseconds now and I'm still getting the Oops.
On Friday, June 5, 2015 at 12:03:29 PM UTC-7, Mark Moorcroft wrote: > > > So the problem was that the only "local" input on our alternate server was > internal metrics. The only reason I even have a second server is because > you don't allow searches for a non-admin. I added the Random HTTP input and > the error disappeared. I don't want to have that either, but it seems I > have no choice but to have some sort of local input now? So I guess the > question is, what is the best throw-away input to have, since there is no > reason for it to exist? > > On Thursday, June 4, 2015 at 8:42:08 PM UTC-7, Mark Moorcroft wrote: >> >> I yum updated both of my CentOS6 graylog servers to 1.1. The primary >> server where all the ES indexes reside seemed to have worked no problem. >> The second one that connects to the 1st seems to work perfectly in every >> way, BUT any attempt to Search results in the Oops message. I see no errors >> in the logs or the System Overview. Even my Dashboard with statistics on >> source message qty values works. >> > -- You received this message because you are subscribed to the Google Groups "graylog2" group. To unsubscribe from this group and stop receiving emails from it, send an email to graylog2+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.