Must Create Maxwell Database ? about maxwell HOT 5 CLOSED

hi @darionyaphet, yes, Maxwell will create a database named maxwell on each instance, to store schema snapshots as well as its replication position. Currently we don't have plans to implement any alternative storage for this state, although I haven't ruled out someday putting all the state in kafka itself.

The reasoning is that I wanted maxwell (a) recoverable -- any process of maxwell requires no local state, and so may pick up where another maxwell left off, and (b) low ops-overhead -- I didn't want to require more bits of architecture than necessary.

from maxwell.

Thank you for you relay @osheroff , I know what you mean keep schema metadata and replication point in reading MySQL instance would reduce another storage , it's very useful . Just a small question , our product environment or DBA maybe not allow create another database instance on each replication slave machine . I try to put this in a config to point where to storage metadata . Thank you for you great work and useful help :)

from maxwell.

I see, so you just want to be able to configure a different server to store schemas and positions in?

from maxwell.

I'm not sure . I wish to design a common metadata schema . Store in MySQL maybe not a goods idea for example MySQL has single point problem .

from maxwell.

ok, not quite sure what you're after, but you can check out https://github.com/zendesk/maxwell/blob/master/src/main/java/com/zendesk/maxwell/schema/Schema.java as a starting point -- that's the in-memory representation of the schema. I'll close this issue out, but feel free to re-open it if you have something more specific.

from maxwell.