sdk
sdk copied to clipboard
networkservicemesh
## Issue Currently dial timeout in `connect` client is a configurable but static value: * Setting it too high can lead us to delays if there are some still registered...
## Current Behavior This behavior is related to healing. For example, if local `icmp-responder` restarted - we start healing from `nsmgr`. New `icmp-responder` receives new request and pulls IP addresses....
## Issue link Closes https://github.com/networkservicemesh/sdk/issues/997 ## How Has This Been Tested? - [x] Added unit testing to cover - [x] Tested manually - [x] Tested by integration testing - [...
## Expected Behavior The NSC and NSE are running in different worker nodes which are connected through `veth pair vxlan tunnel veth pair `vWire connection with this [example ](https://github.com/Nordix/nsm-deployments-k8s/tree/ovs-example/examples/use-cases/Kernel2Kernel) using...
## Overview It can be helpful for a user to run `iperf3` on existing `kernel-...-kernel` connection to get some network performance details for it. So what exactly we want: 1....
## Problem Current logic for round robin do not consider priority when picking next candidate https://github.com/networkservicemesh/sdk/blob/main/pkg/networkservice/common/roundrobin/round_robin_selector.go#L46-L47 ## Expected behaviour For scenarios where we have entry point with >1 possible destinations...
## Description Currently many and many cmd-apps doing the same setup for registry memory or for emulating manager in each test that making reading and supporting tests difficult. Examples: 1....
## Overview Currently we set log level only per app, so everything going through this app is logged at this level. This may be inconvenient. For example, we may have...
# Situation 1 There is a client that wants services `S1` and `S2`, and an endpoint `E1` that implement these services. Client makes a request for `S1` and gets a...