-
with the following code:
```
instance VerifiedJoinSemilattice Nat where
joinSemilatticeJoinIsIdempotent = maximumIdempotent
joinSemilatticeJoinIsCommutative = maximumCommutative
joinSemilattice…
-
I started two servers, then switched the number of replicas between 2 and 1 to test backfilling.
Then I killed a server, declared it dead, lowered the number of required replicas.
At that moment, I wa…
-
Splitting #2663 into multiple issues. This issue covers the low-level cluster config API.
Server status:
``` py
# Documents in this table are readable but not writeable by the user;
# table gets upd…
-
We should give users a way to see a list of current issues through the ReQL admin API.
Here's a proposal to get the ball rolling: Give the user a function `r.get_cluster_issues()`. It returns an arra…
-
We had a discussion on Thursday about adjusting clustering primitives to account for everything we've learned. Here is a concrete proposal based on that discussion. My goal is to pick the simplest, mo…
-
I can add new servers and datacenters ... but it seems I have to go into every individual table and tell it to use the new servers as shards/replicas .... this seems wrong, it I have 200 customers, e…
-
Recently RethinkDB changed storage format twice 1.4->1.5 and 1.5->1.6
Version bump requires manual migration as described here https://github.com/rethinkdb/rethinkdb/tree/next/scripts/migration
This…
-
We can't have the OOM killer wreaking havoc on our users' processes willy-nilly. We need to have a smarter cache memory manager that limits the total cache size per machine to some reasonable number,…
-
There are a few things I'd like to add to indexes to make them a bit nicer:
- a way to get rows that errored
- a way to wait for an index to be ready
- a way to check if an index is ready
- a way to g…
-
We need to notify users about software updates because RethinkDB is evolving very quickly, and because of potentially critical integrity and security updates. Here's how I propose to do it:
1. When th…