Hello everyone,

I have a cache server for compilation artifacts (
https://github.com/uber/buck-http-cache - the current OSS version) which
I'm currently upgrading to 2.4.0.

It's got 4 different caches setup, all partitioned and with 1 backup copy.
While using 1.7.0, the data is distributed among the cluster nodes as
expected.
But when I change to 2.4.0 and enable persistence on the default data
region, all data is only being directed to a single node. Checking the db
directory, I can see that only a single node will have binaries and WAL
growing.
Using the visor cmd, all hits are directed to the same cluster node,
although the topology correctly shows the caches partitioned among all
cluster nodes.

I've been trying to figure this out for a while now, any help would be
appreciated.

Thanks!

Jose Rizzatti

Reply via email to