Hello! Why not have different caches for these? You can control which nodes hold data for a given cache by mean of node filters.
It is not recommended to distribute data on this basis, since load will still be uneven between nodes (much more posts than users, for example). It is recommended to partition data on ID or some other evenly-distributed value. Regards, -- Ilya Kasnacheev сб, 2 мая 2020 г. в 12:16, Sudhir Patil <[email protected]>: > Hello, > > Use case is in given cluster of machines (e.g. 3 to 4), I want to store > all cache item with key name users on one machine, another cache item with > key name posts on different machine etc. Thus is to distribute the load of > cache storage hoping to get performance benefit, better cache management, > etc.. > > Any recommendations i.e. pros cons on this approach is welcome for me... > > On Thursday, April 30, 2020, Ilya Kasnacheev <[email protected]> > wrote: > >> Hello! >> >> I think it is doable to some extent, but I would like to hear why you >> want that. >> >> Regards, >> -- >> Ilya Kasnacheev >> >> >> чт, 30 апр. 2020 г. в 13:21, Sudhir Patil <[email protected]>: >> >>> Hi All, >>> >>> Use case is to store a cache item each on specific machine within >>> cluster. >>> >>> What are recommendations for this? >>> Is it beneficial / performant one? >>> Is it getting away from distributed caching fundamentals & it benefits? >>> If yes, what are reasons behind it ??? >>> >>> Regards, >>> SP >>> >>
