I have restored maxUserTime. Sometimes we must hold the connection for a long time. For example, in the loop retrieve PLC data.
If request timeout the base drive will throw a timeoutException and cacheManage will close this real connection and set the isConnected of the leased connection as false. So we can determine whether the connection is closed in the loop and automatically connect if it is closed. If you can merge my New PR. We should remove maxUserTime again. Christofer Dutz <[email protected]> 于2023年2月22日周三 18:40写道： > Could you please summarize which changes these were exactly? Removing the > maxUserTime? Because if this is the case, we probably need to have a look > at how you are using the ConnectionCache. > > Chris > > > From: Unai Leria <[email protected]> > Date: Wednesday, 22. February 2023 at 10:34 > To: dev <[email protected]> > Subject: Re: Issue with CachedPlcConnectionManager in nifi integration > For the nifi integration to work I did only use the changes on > plc4x/plc4j/tools/connection-cache from spnettec/plc4x/tree/heyoulin. > Plus some minor changes in the nifi integration to add a timeout to all > processors. > > Unai > > ----- Mensaje original ----- > De: "youlin he" <[email protected]> > Para: "dev" <[email protected]>, "Christofer Dutz" < > [email protected]> > Enviados: Miércoles, 22 de Febrero 2023 10:21:51 > Asunto: Re: Issue with CachedPlcConnectionManager in nifi integration > > But it should be based on driver base revision. Otherwise network > connection breaks would still happen. > @Christofer Dutz <[email protected]> > > youlin he <[email protected]> 于2023年2月22日周三 17:16写道： > > > Ok. I will create a PR > > > > Christofer Dutz <[email protected]> 于2023年2月22日周三 16:44写道： > > > >> Unfortunately, this is not a branch, but a fork … Would you be able to > >> pull the changes that made it work in a separate PR? Then we can more > >> quickly adopt the changes, because we definitely are not going to merge > >> that PR (At least I’m not going to do it) > >> > >> Chris > >> > >> > >> From: Unai Leria <[email protected]> > >> Date: Wednesday, 22. February 2023 at 09:28 > >> To: dev <[email protected]> > >> Subject: Re: Issue with CachedPlcConnectionManager in nifi integration > >> It does work in the heyoulin branch. Thanks you. > >> > >> Unai > >> > >> ----- Mensaje original ----- > >> De: "youlin he" <[email protected]> > >> Para: "dev" <[email protected]> > >> Enviados: Martes, 21 de Febrero 2023 14:54:14 > >> Asunto: Re: Issue with CachedPlcConnectionManager in nifi integration > >> > >> I removed the maxUseTime. I think it is unnecessary to add > request-timeout > >> feature in the next version. > >> > >> youlin he <[email protected]> 于2023年2月21日周二 21:48写道： > >> > >> > Can you test use https://github.com/spnettec/plc4x heyoulin branch. > I > >> > fixed this problem. > >> > > >> > Unai Leria <[email protected]> 于2023年2月21日周二 21:17写道： > >> > > >> >> Hi, > >> >> > >> >> I've been working with the CachedPlcConnectionManager on the NiFi > >> >> integration and I have encountered a problem while looking at issue [ > >> >> https://github.com/apache/plc4x/issues/623 | #623 ] : > >> >> When a successful connection is already stored in the cache and the > >> >> network connection breaks the connections in the cache are no longer > >> >> usable, but I have not been able to remove them. > >> >> This makes the processor not work until it is manually disabled and > >> >> enabled. > >> >> > >> >> For the NiFi integration to work properly there should be a way of > >> >> removing a connection from the cache if it is invalid. > >> >> > >> >> I would appreciate some guidance. > >> >> > >> >> Unai > >> >> > >> > > >> > > >
