We use NiFi 1.5.0. Ok we will try upgrading NiFi and revert. Thanks for your help.
On Fri, Apr 12, 2019 at 6:30 PM Bryan Bende <bbe...@gmail.com> wrote: > What version of NiFI? There have been some significant improvements > to clustering performance since 1.7.0 forward I believe. > > On Fri, Apr 12, 2019 at 3:00 AM saloni udani <saloniudani...@gmail.com> > wrote: > > > > We do not have any custom processors.Although we do use lot of > InvokeScriptedProcessor with Groovy scripts. Our topology is huge (1200+ > processors). When we try to stop manually whole topology, it does stop > cleanly taking lot of time. > > > > When we try to update flow version from UI, it timeouts with default > config. But if we update cluster connection and read timeout to 5min then > the update completes successfully. > > > > On Thu, Apr 11, 2019 at 7:25 PM Bryan Bende <bbe...@gmail.com> wrote: > >> > >> The underlying issue is that there are components in your versioned > >> flow that can't be stopped for some reason. > >> > >> Do you have any custom processors? If you try to stop everything > >> manually do all processors stop cleanly? > >> > >> On Thu, Apr 11, 2019 at 7:22 AM saloni udani <saloniudani...@gmail.com> > wrote: > >> > > >> > Ref: https://issues.apache.org/jira/browse/NIFI-5340 > >> > > >> > Any updates on this? We are facing same issue. For time being we have > increased connection and read timeout. > >> > > >> > > >> > > >> > > >> > Thanks >