2023-04-07 bugfix release

[renatolfc]

Yesterday we updated FarmVibes.AI with fixes to the both the backend and the client library. With these fixes, workflow management will make more sense now. We also added a new feature that will improve quality of life when running FarmVibes.AI: the ability to resubmit workflows for execution! Below you will find a summary of all changes.

As usual, if you already have cloned the repository:

git pull

should fetch the new code. We recommend an update of the library and images with:

bash farmvibes-ai.sh update

Please let us know if you have any issues.

---

Notebooks

• [⚒️ FIX] We've updated the baseline and scenario files for the Carbon sequestration example notebook, adding a more comprehensive set of examples, including fertilization and organic amendment examples. In addition to that, we have included a more concise set of dates to comply with the current crop year.

Client

• [⚒️ FIX] We resolved the issue with the monitor showing negative times by correctly adjusting displayed times to the time zone of the client running the monitor.
• [🎉 NEW] We now display warning messages in the monitor, such as when the system is running low on space, or when we can't determine the user's time zone.
• [🎉 NEW] We added support to resubmitting runs with the same input parameters via client (FarmvibesAiClient().resubmit_run(run_id)) or directly through the run object (VibeWorkflowRun().resubmit()).

CLI

• [📈 IMPROVEMENT] The dapr sidecar components are now created by a TerraForm script.

Backend

• [⚒️ FIX] We fixed a bug in the worker in which failed tasks would be marked as completed successfully, but without any output. Now these tasks are correctly marked as failed.
• [🎉 NEW] The backend now supports more general types (BaseVibe) that don't require a geometry and a time range to be processed by the system.
• [⚒️ FIX] We've improved how we download Sentinel data from the Planetary Computer, failing download operations that would otherwise hang in previous versions.
• [⚒️ FIX] We've fixed a resource leakage in our orchestrator that would make it consume unbounded amounts of memory.
• [📈 IMPROVEMENT] We've changed our default logging level from DEBUG to INFO, potentially reducing the amount of disk space used in user deployments.

Known Issues

• [🐛 BUG] Workflows that are submitted soon after the cluster is started (after a call to farmvibes-ai.sh stop followed by a call to farmvibes-ai.sh start) might fail, and may require a call to farmvibes-ai.sh restart to restore cluster state. This happens because the kubernetes implementation we use reports deployments being in the Ready state without that being the case. We are investigating how that can be fixed.

[richstep]

running bash farmvibes-ai.sh update gives me this error

Plan: 16 to add, 0 to change, 0 to destroy.
module.kubernetes.kubernetes_namespace.kubernetesdaprnamespace: Creating...
module.kubernetes.kubernetes_persistent_volume.user_storage_pv: Creating...
module.kubernetes.helm_release.redis: Creating...
module.kubernetes.helm_release.rabbitmq: Creating...
╷
│ Error: namespaces "dapr-system" already exists
│
│ with module.kubernetes.kubernetes_namespace.kubernetesdaprnamespace,
│ on modules/kubernetes/dapr.tf line 1, in resource "kubernetes_namespace" "kubernetesdaprnamespace":
│ 1: resource "kubernetes_namespace" "kubernetesdaprnamespace" {
│
╵
╷
│ Error: persistentvolumes "user-storage-pv" already exists
│
│ with module.kubernetes.kubernetes_persistent_volume.user_storage_pv,
│ on modules/kubernetes/persistentvolume.tf line 1, in resource "kubernetes_persistent_volume" "user_storage_pv":
│ 1: resource "kubernetes_persistent_volume" "user_storage_pv" {
│
╵
╷
│ Error: cannot re-use a name that is still in use
│
│ with module.kubernetes.helm_release.rabbitmq,
│ on modules/kubernetes/rabbitmq.tf line 1, in resource "helm_release" "rabbitmq":
│ 1: resource "helm_release" "rabbitmq" {
│
╵
╷
│ Error: cannot re-use a name that is still in use
│
│ with module.kubernetes.helm_release.redis,
│ on modules/kubernetes/redis.tf line 1, in resource "helm_release" "redis":
│ 1: resource "helm_release" "redis" {