Releases: truefoundry/terraform-azure-truefoundry-cluster
Releases · truefoundry/terraform-azure-truefoundry-cluster
v0.6.2
What's Changed
- Making node count as null to avoid downscaling by @dunefro in #21
- This change ensure that nodepools don't downscale, so if you get the node count from n -> 0 in the terraform plan please switch to this version for avoiding downscaling
Full Changelog: v0.6.1...v0.6.2
v0.6.1
v0.6.0
What's Changed
- chore(terraform): update azurerm provider version constraint to allow newer versions by @sachincool in #16
- Moving to common workflows by @dunefro in #17
- hl/update version by @sachincool in #18
- Added support for existing cluster by @dunefro in #19
New Contributors
- @sachincool made their first contribution in #16
Full Changelog: v0.5.2...v0.6.0
Changelog Description
- Added support for existing cluster using
var.use_existing_cluster
- Added data block to fetch existing cluster details
- Added upgrade guide to have migration steps.
- Added a
scripts/
directory to have migration scripts and other uses.
v0.5.2
v0.5.1
v0.5.0
v0.4.3
v0.4.2
v0.4.1
v0.4.0
What's Changed
Description
Below are the list of the changes
- Added multiple node pools for GPU
- Added spot and on-demand node pool
- Add support for adding additional cpu/gpu spot/on-demand node pools
- AKS module ==
8.0.0
Important
Node pool change
This change will create lot of node pools by default when the cluster gets created. If you want to avoid that just make sure to keep create_default_cpu_pools
to false
and create_default_gpu_pools
to false
. However, these changes wouldn't impact your existing nodepools.
Max pods per node
By Default to avoid pod cluttering and IP distribution max_pods_per_node
is by default set to 32
. In the default scenario it will be 110
. Upon drift detection in this make sure to keep max_pods_per_node
to 110
. This can cause your cluster to re-create.
Full Changelog: v0.3.2...v0.4.0