Lucas Kent
Lucas Kent
We need some way to mark inaccessible nodes as inaccessible to all transform instances. And then if every node is marked as inaccessible we restart our initialization process from contact_points....
Just publishing this to demonstrate the new serialization format of serde_yaml. We will have to decide if we want to actually upgrade to this new format, find an alternative crate...
Cassandra already has a timeout when waiting for a response from the server, we should add the same for redis (and make it configurable length) The field should be named...
## Motivation: Unlike redis, a queried cassandra node will become a coordinator node and will fetch the results of your query from another node if needed. So we dont technically...
Tried writing this bench and hit a shotover failure that appears to be caused by opening lots of connections really fast. Opening a PR to put visibility on the problem...
closes https://github.com/shotover/shotover-proxy/issues/702 In order to catch invalid commands I inserted a validation check both just after the command is decoded and just before it is encoded. This seems like a...
Pulled out of https://github.com/shotover/shotover-proxy/pull/645 which uncovered the issue: By dumping the requests/responses in RedisSinkSingle and running `cargo test cassandra_int_tests::test_cassandra_redis_cache` we get: ``` ERROR request{id=5 source="CassandraSource"}: shotover_proxy::transforms::redis::sink_single: request: [Message { inner:...
The transform message invariants allow for less messages to be returned than were sent, as long as every message eventually receives a reply. The reason for this is to allow...
The following unsafe code is used to uninit the metrics recorder so it can be later reinited. `unsafe { metrics::clear_recorder() };` There is no safe way to perform the same...
## Describe the bug We should not be vendoring random library code, instead lets upstream our custom FuturesOrdered::prepend method to https://github.com/rust-lang/futures-rs