You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: CHANGELOG.md
+22-11Lines changed: 22 additions & 11 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -5,27 +5,38 @@ CHANGELOG
5
5
-----
6
6
7
7
**ENHANCEMENTS**
8
-
- Add support for updating `SharedStorage` configuration.
8
+
- Add possibility to specify multiple instance types for the same compute resource.
9
+
- Add support for adding and removing shared storages at cluster update by updating `SharedStorage` configuration.
9
10
- Add new configuration parameter `DeletionPolicy` for EFS and FSx for Lustre shared storage
10
11
to support storage retention on deletion.
11
-
- Enable server-side encryption for the EcrImageBuilder SNS topic created when deploying ParallelCluster API and used to notify on docker image build events.
12
-
- Add support for on-demand capacity reservations.
13
-
- Add support for specifying the supported IMDS version in cluster and build image configurations via the `Imds > ImdsSettings` property. A value of `v1.0` (the default) means that both IMDSv1 and IMDSv2 are supported, while a value of `v2.0` means that only IMDSv2 is supported.
12
+
- Add support for AWS Trainium instances.
14
13
- Add support for Slurm Accounting.
14
+
- Add support for on-demand capacity reservations.
15
+
- Add support for specifying the supported IMDS version in cluster and build image configurations via the `Imds/ImdsSettings` property.
16
+
- Add support for instances with multiple network interfaces that allows only one ENI per device.
15
17
- Improve validation of networking for external EFS file systems by checking the CIDR block in the attached security group.
16
-
- Improve support for EC2 instances with several NICs.
17
-
- Add validator for instances being launched in a cluster placement group when the instance type does not support it.
18
+
- Add validator for instances being launched in a cluster placement group when the instance type does not support it.
19
+
- Configure NFS threads to be `min(256, max(8, num_cores * 4))` to ensure better stability and performance.
20
+
- Move NFS installation at build time to reduce configuration time.
21
+
- Enable server-side encryption for the EcrImageBuilder SNS topic created when deploying ParallelCluster API and used to notify on docker image build events.
18
22
19
23
**CHANGES**
20
24
- Remove support for Python 3.6 in aws-parallelcluster-batch-cli.
21
-
- Upgrade Python and NodeJS versions in API infrastructure, API Docker container and cluster Lambda resources.
25
+
- Upgrade Python to version 3.9 and NodeJS to version 16 in API infrastructure, API Docker container and cluster Lambda resources.
22
26
- Move head node tags from Launch Template to instance definition to avoid head node replacement on tags updates.
23
27
- Disable Multithreading through script executed by cloud-init and not through CpuOptions set into Launch Template.
24
-
- Add support for multiple instance types in the same Compute Resource.
25
-
- Add support for a Name field in PlacementGroup as the preferred naming method.
26
-
- Add support for Networking.PlacementGroup in the SlurmQueues.ComputeResources section
28
+
- Add support for `PlacementGroup/Name` as the preferred naming method.
29
+
- Add support for `Networking/PlacementGroup` in the `SlurmQueues/ComputeResources` section.
30
+
- Change behaviour of `SlurmQueues/Networking/PlacementGroup/Enabled`: now it creates a unique managed placement
31
+
group for each compute resource instead of a single managed placement group for all compute resources.
32
+
- Upgrade Slurm to version 22.05.5.
33
+
- Upgrade NVIDIA driver to version 470.141.03.
34
+
- Upgrade NVIDIA Fabric Manager to version 470.141.03.
35
+
- Upgrade NVIDIA CUDA Toolkit to version 11.7.1.
36
+
- Upgrade Python used in ParallelCluster virtualenvs from 3.7.13 to 3.9.15.
27
37
- Upgrade Slurm to version 22.05.5.
28
-
- SlurmQueues.Networking.PlacementGroup.Enabled alone will now create a unique managed placement for each compute resource instead of a single managed placement group for all compute resources
38
+
- Upgrade EFA installer to version 1.18.0.
39
+
- Upgrade NICE DCV to version 2022.1-13300.
29
40
30
41
**BUG FIXES**
31
42
- Fix validation of `filters` parameter in `ListClusterLogStreams` command to fail when incorrect filters are passed.
0 commit comments