|
1 | 1 | # Changelog |
2 | 2 |
|
| 3 | +<!-- DO NOT EDIT. Generated from git commit history by scripts/gen-changelog.sh. |
| 4 | + Hand-authored behavior/upgrade notes live in RELEASE_NOTES.md (same directory). --> |
| 5 | + |
3 | 6 | All notable changes to this package will be documented in this file. |
4 | 7 |
|
5 | | -## [0.4.0] |
| 8 | +## [Unreleased] |
| 9 | + |
| 10 | + |
| 11 | +## [0.5.0] - 2026-07-09 |
6 | 12 |
|
7 | 13 | ### New Features |
8 | 14 |
|
9 | | -- *(nvidia-setup)* Bump EKS h100/gb200 defaults to kernel `6.17.0-1019-aws` and EFA `1.48.0` by [@ayuskauskas](https://github.com/ayuskauskas) |
| 15 | +- *(nvidia-setup)* Add aks-h100 combination for IB RDMA + memlock |
10 | 16 |
|
11 | | -### Bug Fixes |
| 17 | +## [0.4.0] - 2026-07-02 |
| 18 | + |
| 19 | +### New Features |
| 20 | + |
| 21 | +- *(changelog)* Tag-range generator, split CHANGELOG/RELEASE_NOTES, release helper |
| 22 | +- *(nvidia-setup)* EKS h100/gb200 → 6.17.0-1019-aws kernel + EFA 1.48.0 |
| 23 | + |
| 24 | +### Other Tasks |
12 | 25 |
|
13 | | -- *(nvidia-setup)* Resolve the 64k-page kernel for arm64: `resolve_full_kernel` now appends `-64k` on aarch64 (e.g. GB200 installs `6.17.0-1019-aws-64k`) instead of the non-64k flavor by [@ayuskauskas](https://github.com/ayuskauskas) |
14 | | -- *(nvidia-setup)* Self-heal the kernel install when dpkg is left half-configured: on a dpkg error the apt-get install now runs `dpkg --configure -a` and retries once instead of failing the step by [@ayuskauskas](https://github.com/ayuskauskas) |
| 26 | +- *(nvidia-setup)* Update config.json version to 0.4.0 |
15 | 27 |
|
16 | 28 | ## [0.3.0] - 2026-05-29 |
17 | 29 |
|
| 30 | +### Bug Fixes |
| 31 | + |
| 32 | +- *(nvidia-setup)* Warn but exit 0 when /usr/src/linux-$(uname -r) is a real dir |
| 33 | + |
18 | 34 | ### New Features |
19 | 35 |
|
20 | | -- *(nvidia-setup)* Add `bcm` service that aliases `/usr/src/linux-$(uname -r)` to Ubuntu's `linux-headers-$(uname -r)` tree so `aicr validate` finds `.config` ([AICR #1093](https://github.com/NVIDIA/aicr/issues/1093)) by [@ayuskauskas](https://github.com/ayuskauskas) |
| 36 | +- *(nvidia-setup)* Add bcm service that aliases kernel headers |
| 37 | + |
| 38 | +### Other Tasks |
| 39 | + |
| 40 | +- *(nvidia-setup)* Downgrade refuse-to-replace log from ERROR to WARNING |
| 41 | +- Merge pull request #48 from NVIDIA/feat/nvidia-setup-bcm-service |
| 42 | + |
| 43 | +feat(nvidia-setup): add bcm service that aliases kernel headers |
21 | 44 |
|
22 | 45 | ## [0.2.2] - 2026-04-27 |
23 | 46 |
|
24 | 47 | ### Bug Fixes |
25 | 48 |
|
26 | | -- *(nvidia-setup)* Remove sudo as we are root already by [@ayuskauskas](https://github.com/ayuskauskas) |
27 | | -- *(nvidia-setup)* Pass dpkg force-conf{def,old} to silence conffile prompts by [@ayuskauskas](https://github.com/ayuskauskas) |
| 49 | +- *(nvidia-setup)* Remove sudo as we are root already |
| 50 | +- *(nvidia-setup)* Pass dpkg force-conf{def,old} to silence conffile prompts |
28 | 51 |
|
29 | 52 | ### Other Tasks |
30 | 53 |
|
31 | 54 | - Merge pull request #33 from NVIDIA/fix/no-sudo |
32 | 55 |
|
33 | | -fix(nvidia-setup): remove sudo as we are root already by [@ayuskauskas](https://github.com/ayuskauskas) |
34 | | -- Update project to follow the template by [@lockwobr](https://github.com/lockwobr) |
| 56 | +fix(nvidia-setup): remove sudo as we are root already |
| 57 | +- Update project to follow the template |
35 | 58 | - Merge pull request #40 from NVIDIA/fix/tuned |
36 | 59 |
|
37 | | -Fix nvidia-tuned and nvidia-setup for eks-gb200 by [@ayuskauskas](https://github.com/ayuskauskas) |
| 60 | +Fix nvidia-tuned and nvidia-setup for eks-gb200 |
| 61 | +- *(nvidia-setup)* Bump changelog and version for release |
38 | 62 |
|
39 | 63 | ## [0.2.1] - 2026-03-09 |
40 | 64 |
|
41 | 65 | ### New Features |
42 | 66 |
|
43 | | -- *(nvidia-setup)* Change expected service to aws from eks to align with nvidia-tuned by [@ayuskauskas](https://github.com/ayuskauskas) |
44 | | -- *(nvidia-setup)* Update service to be eks to match aicr by [@ayuskauskas](https://github.com/ayuskauskas) |
| 67 | +- *(nvidia-setup)* Change expected service to aws from eks to align with nvidia-tuned |
| 68 | +- *(nvidia-setup)* Update service to be eks to match aicr |
45 | 69 |
|
46 | 70 | ### Other Tasks |
47 | 71 |
|
48 | 72 | - Merge pull request #27 from NVIDIA/feat/nvidia-tuned/fix-containerd |
49 | 73 |
|
50 | | -feat(nvidia-tuned): change containerd drop to be a script by [@ayuskauskas](https://github.com/ayuskauskas) |
| 74 | +feat(nvidia-tuned): change containerd drop to be a script |
51 | 75 | - Merge pull request #29 from NVIDIA/fix/nvidia- |
52 | 76 |
|
53 | | -Fix/nvidia by [@ayuskauskas](https://github.com/ayuskauskas) |
| 77 | +Fix/nvidia |
54 | 78 |
|
55 | 79 | ## [0.2.0] - 2026-03-04 |
56 | 80 |
|
57 | 81 | ### Bug Fixes |
58 | 82 |
|
59 | | -- *(nvidia-setup)* Remove the running in a temp dir for efa install by [@ayuskauskas](https://github.com/ayuskauskas) |
| 83 | +- *(nvidia-setup)* Remove the running in a temp dir for efa install |
60 | 84 |
|
61 | 85 | ### New Features |
62 | 86 |
|
63 | | -- *(nvidia-setup)* Add ofi setup by [@ayuskauskas](https://github.com/ayuskauskas) |
| 87 | +- *(nvidia-setup)* Add ofi setup |
64 | 88 |
|
65 | 89 | ### Other Tasks |
66 | 90 |
|
67 | 91 | - Merge pull request #26 from NVIDIA/feat/nvidia-setup-ofi |
68 | 92 |
|
69 | | -feat(nvidia-setup): add ofi setup by [@ayuskauskas](https://github.com/ayuskauskas) |
| 93 | +feat(nvidia-setup): add ofi setup |
70 | 94 |
|
71 | | -## [0.1.1] - 2026-03-03 |
| 95 | +## [0.1.1] - 2026-03-02 |
72 | 96 |
|
73 | 97 | ### New Features |
74 | 98 |
|
75 | | -- *(nvidia-setup)* Can use NVIDIA_SETUP_KERNEL_ALLOW_NEWER to set if… by [@ayuskauskas](https://github.com/ayuskauskas) |
| 99 | +- *(nvidia-setup)* Can use NVIDIA_SETUP_KERNEL_ALLOW_NEWER to set if… |
76 | 100 |
|
77 | 101 | ## [0.1.0] - 2026-03-02 |
78 | 102 |
|
79 | 103 | ### Bug Fixes |
80 | 104 |
|
81 | | -- *(nvidia-setup)* Renable the checks for chrony and raid by [@ayuskauskas](https://github.com/ayuskauskas) |
82 | | -- *(nvidia-setup)* Remove kernel check from apply_check as this hasn't restarted yet so won't pass by [@ayuskauskas](https://github.com/ayuskauskas) |
83 | | -- *(nvidia-setup)* Move tests by [@ayuskauskas](https://github.com/ayuskauskas) |
84 | | -- *(nvidia-setup)* Bad call to the memory check by [@ayuskauskas](https://github.com/ayuskauskas) |
85 | | -- Add install of pkgs to support setup_local_disks by [@ayuskauskas](https://github.com/ayuskauskas) |
| 105 | +- *(nvidia-setup)* Renable the checks for chrony and raid |
| 106 | +- *(nvidia-setup)* Remove kernel check from apply_check as this hasn't restarted yet so won't pass |
| 107 | +- *(nvidia-setup)* Move tests |
| 108 | +- *(nvidia-setup)* Bad call to the memory check |
| 109 | +- Add install of pkgs to support setup_local_disks |
86 | 110 |
|
87 | 111 | ### New Features |
88 | 112 |
|
89 | | -- Add nvidia-setup package for bringing a node up to ai runtime spec by [@ayuskauskas](https://github.com/ayuskauskas) |
90 | | -- Add basic testing of packages via containers by [@ayuskauskas](https://github.com/ayuskauskas) |
91 | | -- *(nvidia-setup)* Kernel install flow, shared defaults, EFA resilience by [@ayuskauskas](https://github.com/ayuskauskas) |
92 | | -- *(nvidia-setup)* Turn chrony and raid string change back on by [@ayuskauskas](https://github.com/ayuskauskas) |
93 | | -- *(nvidia-setup)* Change that pinning is optional by [@ayuskauskas](https://github.com/ayuskauskas) |
94 | | -- *(nvidia-setup)* Tests for kernel checks by [@ayuskauskas](https://github.com/ayuskauskas) |
| 113 | +- Add nvidia-setup package for bringing a node up to ai runtime spec |
| 114 | +- Add basic testing of packages via containers |
| 115 | +- *(nvidia-setup)* Kernel install flow, shared defaults, EFA resilience |
| 116 | +- *(nvidia-setup)* Turn chrony and raid string change back on |
| 117 | +- *(nvidia-setup)* Change that pinning is optional |
| 118 | +- *(nvidia-setup)* Tests for kernel checks |
95 | 119 |
|
96 | 120 | ### Other Tasks |
97 | 121 |
|
98 | 122 | - Merge pull request #21 from NVIDIA/testing |
99 | 123 |
|
100 | | -Add nvidia-tuned and unit-like integration tests for packages by [@ayuskauskas](https://github.com/ayuskauskas) |
| 124 | +Add nvidia-tuned and unit-like integration tests for packages |
101 | 125 | - Merge pull request #23 from NVIDIA/nvidia-tuned-fix |
102 | 126 |
|
103 | | -feat(nvidia-setup): kernel install flow, shared defaults, EFA resilience by [@ayuskauskas](https://github.com/ayuskauskas) |
| 127 | +feat(nvidia-setup): kernel install flow, shared defaults, EFA resilience |
104 | 128 |
|
105 | 129 | <!-- Generated by git-cliff --> |
0 commit comments