Can you submit a bug in JIRA for this? If nothing else, it may prompt a developer who's got some time to go through all the places that counter is updated and see if there are any code paths that could result in this behavior. Make sure you describe whatever might have happened between the last time it was known to be good and when it was seen to be bad, as well as any errors in the logs between those times. On Jun 18, 2015 5:08 AM, "Christopher Shannon" < christopher.l.shan...@gmail.com> wrote:
> I've been doing a lot of work with the metrics and monitoring of the broker > recently so if this happens again in the future and you are able to find a > way to reproduce it I'd be happy to take a look at it. > > -Chris > > On Thu, Jun 18, 2015 at 5:12 AM, ALi <osat...@gmail.com> wrote: > > > no .... it has happen only once ... and after working a year without any > > problem. so it is a bit difficult to reproduce. we will continue > monitoring > > the metric but without an alert. > > > > thank you tim > > > > On Tue, Jun 9, 2015 at 2:34 PM, Tim Bain <tb...@alumni.duke.edu> wrote: > > > > > Does this happens reliably and quickly for you? Can you force it to > > happen > > > by taking a certain series of actions? > > > On Jun 8, 2015 11:45 PM, "ALi" <osat...@gmail.com> wrote: > > > > > > > It is 5.8.0. As it was working ok we didn't want to update it. The > new > > > > version 5.10 had problems with stomp and websocket. We need to test > > this > > > > new version to see if it works and if it doesn't have any problems > with > > > it. > > > > And we don't know if it worths the update. But as we are having this > > > > incident we need to think about it. > > > > El 09/06/2015 01:33, "Christopher Shannon" < > > > > christopher.l.shan...@gmail.com> > > > > escribió: > > > > > > > > > The count shouldn't be negative so this might be a bug. The broker > > > keeps > > > > > track of the consumer count using a counter and the counter should > be > > > > > incremented on a new subscription and decremented when a > subscription > > > is > > > > > removed. So if there are no consumers the count should be 0, not > > > > negative. > > > > > > > > > > What version of ActiveMQ are you using? If it is older than 5.11.1 > > you > > > > > could try a newer version to see if that helps. Would you be able > to > > > put > > > > > together a test case to reproduce this issue? > > > > > > > > > > On Mon, Jun 8, 2015 at 4:42 PM, ALi <osat...@gmail.com> wrote: > > > > > > > > > > > When can the total consumer count be negative? We are using this > > > metric > > > > > to > > > > > > know how many people is connected. > > > > > > > > > > > > We are getting this metric from the jmx perhaps a problem with > the > > > > query > > > > > in > > > > > > check jmx eval? > > > > > > > > > > > > > > > > > > > > >
