-
The shard config data should be a (strict or not) subset of actual shards. That is, it should not contain any treeIDs that don't exist in trillian in reality. We should add a check for this at some po…
-
### Is your feature request related to a problem? Please describe
Sometimes, In a cluster where there are indices of varied sizes, it unbalances the disks of the nodes to a great extent as currently …
-
Any version of `crystal >= 1.12.0` fails with this
```
$ make
diff -w -c
-
### Description
### Description
Blindly contacting nodes and shards to run a plan or query (even if cheap) has a detrimental effect especially in large clusters due to the associated cost:
- no…
-
When not specifying the version of shrine.cr I get the following error:
```
Error in shrine:shard.yml: duplicate attribute "development_dependencies" at line 22, column 1
20. gitlab: arctic…
-
Apply this patch:
```patch
diff --git a/service/topology_coordinator.cc b/service/topology_coordinator.cc
index 591f1d341c..6e94895bad 100644
--- a/service/topology_coordinator.cc
+++ b/service…
-
Today we have a single big database for storage, which is probably fine, given that with sharding phase 1 all nodes need to track all shards all the time. This is, however non-ideal, for multiple reas…
-
Right now sharding is done in round-robin manner per folder, see https://github.com/pytorch/tutorials/blob/master/.jenkins/build.sh
Which creates a very unbalanced experience as some folders have 20+…
-
The service_level_controller applies changes to service levels. The controller stores the information about the existing in a `_service_levels_db` map on each shard. The updates are always initiated o…
-
It might be possible to use [Beaver](https://github.com/beaver-lodge/beaver) to parse StableHLO modules into an Elixir data structure, and convert that into the equivalent Nx.Defn.Expr.
This should…