fluvio
fluvio copied to clipboard
Lean and mean distributed stream processing system written in rust and web assembly.
Related to #1514 This should be captured in a test. The expectation is if we open a consumer stream, and then simulate a network failure (like if we use `iptables`...
Related #1514 Related #1954 This is the cluster-side portion of #1954 If a client has a consumer stream open, and an SPU fails (and restarts), the client should not fail....
Related #1955 If a client has an open consume stream, and the cluster is going through an upgrade, the client should resume once the cluster becomes available again w/o any...
When there's a new version available, we want to let users know (especially if there are breaking changes cli w/ server updates). This doesn't have to be complicated, but there...
We create a new tag for docker images on every merge to master. The only users of these images are probably just us. So we need to be a little...
The process for upgrading a local cluster is to stop the current cluster and start a new cluster at a new version. Data is lost when the initial cluster is...
Based on this comment https://github.com/infinyon/fluvio-connectors/issues/125#issuecomment-1024468529 The script we use for building and loading k8 images needs to be split up so we can re-use the loading script in other projects...
Cluster start and cluster checks use the locally installed `kubectl`. This is about our usage of `kubectl` in the Rust code. Not in shell scripts or CI * [crates/fluvio-cluster/src/check/mod.rs](https://github.com/infinyon/fluvio/blob/c9d694a6df687ff24bab1edf20d95c819dc3b987/crates/fluvio-cluster/src/check/mod.rs) *...
Support Kubernetes/ WSL2: We do not yet support the Windows Subsystem for Linux (WSL), but we have an interest in making at least the client crate and the CLI work...
With a records-based retention policy, records are being deleted individually based on the policy, typically by age. This allows the preservation of at least one record per key. Since deletions...