Clojurians Log v2

Thank you for answering. CockroachDB is certainly not my proposal, some colleagues at work were worried about scalability and would like to use a distributed DB while sticking to SQL storage, so it was a natural option. We will meet with CockroadDB lab tomorrow, what would be the critical questions to ask them regarding Datomic support?

Hmmm. Thats cool. I think we would just need to commit to supporting on our end and ensure we test it. But I will share this question with the development team and see if they can chime in.

i could blindly guess that cockroachdb is probably easier to setup as a cluster, than postgres. haiku agrees too: > CockroachDB vs PostgreSQL: what are the pros and cons? ... > TL;DR: PostgreSQL is the go-to for traditional applications; CockroachDB is built for distributed, globally-scaled systems that need built-in resilience.

disclaimer: i never tried cockroach, but i had the "misfortune" to setup a psql cluster and test its failover capability sometime around ~2010, when AWS RDS didn't have support for it. i did it all through a 10kbyte/s mobile EDGE connection, on a 10" asus eee pc 1000ha (because my 12" PPC iBook became unreliable), from the thai jungle, coordinating thru skype, which sometimes ate up all my bandwidth, because it promoted itself to be a local supernode or something... lolcry

Sorry I didnt get back to you @huahaiy, hope you had a good meeting. We believe we can already support CockroachDB, as it should just work given our generalized usage of all SQL backend storages. However, I am not familiar with all the knobs you can possibly set and perhaps there are edge cases where our K/V usage falls over (i.e. maybe replication). We simply havent had users report using it as a storage and thus haven't worked through any edge case issues. If they're aware of any Datomic usage or test with Datomic and find issues we would be happy to review, especially if more folks in the community start wanting to use CockroachDB.

We did some initial testing and it seems to work. We will continue observing and will report if there's any edge cases. Thank you for the response.