Thank you Pavel! Having raft and split brain protection in 3.0 is attractive so we will likely migrate to 3.0 once distributed data structures are available. Do you think they'll be available within 6-9mo after general 3.0 release?
On Fri, Nov 8, 2024 at 5:27 AM Pavel Tupitsyn <ptupit...@apache.org> wrote: > > 1. Is it really coming this year? > > Yes, late this year / early next year. > > > 2. When distributed data structures (locks, semaphores, queues etc) are > planned? > > Later, not in the upcoming 3.0 > > > 3. What is your take on embedding? Are you moving away from embedding > > Embedding is implemented, it will be available in 3.0 > > > 4. Should we wait for 3.0 to come out to start our development or start > with 2.x in thin client mode given what appears to be a move away from easy > embedding? Our current architecture is embedded hazelcast > > Don't wait for 3.0. If you like how 2.16 works - go with it. > 2.x and 3.x are developed separately and will co-exist in the foreseeable > future. > > On Fri, Nov 8, 2024 at 1:03 AM Alex Roytman <roytm...@gmail.com> wrote: > >> Hello, >> >> We are considering Ignite for a new project (actually for getting off >> Hazelcast) >> I went over 2.16 docs and played with it and all key APIs that I need are >> there (Queues, Locks, Semaphores, Atomic numbers and references and id >> generators). It also is easily embeddable with its relatively compact and >> repackaged jars. Easy embedding is very attractive for us because we use it >> primarily to coordinate and orchestrate some heavy data data processing not >> stored in ignite (multi stage files processing, text extraction indexing, >> storage distribution etc) hence most interest is in queues, locks, >> semaphores vs tables/caches >> >> Then I found a 3.0 roadmap which seems to have a release scheduled very >> soon in a matter of a month or two. >> >> So I have a few questions >> >> 1. Is it really coming this year? There was no anything beyond beta 2 >> from more than a year ago, little to no docs etc. Is the roadmap obsolete >> or 2024 deployment still on the books? >> 2. When distributed data structures (locks, semaphores, queues etc) are >> planned? Same release timeline? >> 3. I tried to include API and Runner artifacts into my project and it >> brought a significant set of dependencies including MIcronaut. Making >> embedding into applications that already have CDI, Web container etc. >> potentially challenging. What is your take on embedding? Are you moving >> away from embedding as a first class integration to a closed >> "server bundle" distribution and no native user facing Java APIs? >> 4. Should we wait for 3.0 to come out to start our development or start >> with 2.x in thin client mode given what appears to be a move away from easy >> embedding? Our current architecture is embedded hazelcast >> >> Thank you very much >> Alex Roytman >> >
