Great, I've actually verified that for my case the approach worked as expected.
Thanks, Santi On Mon, Nov 4, 2013 at 8:27 PM, Zhen Zhang <[email protected]> wrote: > Hi Santi, as long as the new state model definition is compatible with the > old one, your approach should be fine. I've added a test for this. > > > On Mon, Nov 4, 2013 at 1:34 PM, Santiago Perez <[email protected]> > wrote: > > > Participants are already running with code that supports new state > > transitions. > > > > Also all states currently in CURRENT_STATES are in both models. > > > > > > On Mon, Nov 4, 2013 at 6:29 PM, Shi Lu <[email protected]> wrote: > > > >> Hi Santi: > >> > >> For your updated statemodel, does it involve adding new state > transitions? > >> > >> If yes then the helix participant will need to be updated as well, since > >> they need to implement the new added state transitions. If not then the > >> participant will not be able to process the updated state transition > >> messages and most likely will go into error state. > >> > >> > >> > >> > >> On Mon, Nov 4, 2013 at 1:14 PM, Santiago Perez <[email protected]> > >> wrote: > >> > >> > Hi, > >> > > >> > I need to redefine a state model on a live cluster and would like to > >> know > >> > the recommended (safe) way to do so. > >> > > >> > My current plan was to do the fololwing: > >> > > >> > 1) Stop all controllers > >> > 2) Remove the node from STATEMODELDEFS > >> > 3) Run code that will rewrite the state model > >> > 4) Start controllers again > >> > > >> > I'm assuming that only controllers load the state models, but I may be > >> > mistaken. > >> > > >> > Please advice. > >> > > >> > Thanks, > >> > Santiago > >> > > >> > > > > >
