UniqAI/nvidia-container-toolkit
1
0
Fork 0
mirror of https://github.com/NVIDIA/nvidia-container-toolkit synced 2024-11-22 16:29:18 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
1,152 Commits 20 Branches 73 Tags 28 MiB
1e5a6e1fa3
Commit Graph

1152 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Evan Lezar
46de426cc4 Merge branch 'CNT-1330/jenkins-ci' into 'master' 
Add Jenkins file for CI build steps

See merge request nvidia/container-toolkit/container-toolkit!28
 2021-03-18 10:06:44 +00:00
Evan Lezar
1c7d6a233a Add golang check targets 
This change adds check targets for Golang to the make file. These are also
added as stages to the to the Jenkinsfile definition and the GitLab CI
is modified to use them too.

Signed-off-by: Evan Lezar <elezar@nvidia.com>
 2021-03-17 16:58:39 +01:00
Evan Lezar
635aeb8343 Add Jenkinsfile definition for build targets 
Signed-off-by: Evan Lezar <elezar@nvidia.com>
 2021-03-17 13:52:19 +01:00
Evan Lezar
ec9d296afe Move docker.mk to docker folder 
Signed-off-by: Evan Lezar <elezar@nvidia.com>
 2021-03-17 13:52:14 +01:00
Evan Lezar
ff44395b31 Merge branch 'upstream-bump-v1.4.2' into 'master' 
Bump version to 1.4.2

See merge request nvidia/container-toolkit/container-toolkit!27
 2021-02-05 12:47:01 +00:00
Kevin Klues
8571e5ac5d Bump version to 1.4.2 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2021-02-05 10:26:10 +00:00
Kevin Klues
108c99bb9b Merge branch 'upstream-bump-v1.4.1' into 'master' 
Bump version to 1.4.1

See merge request nvidia/container-toolkit/container-toolkit!26
 2021-01-25 13:35:42 +00:00
Kevin Klues
dfb5daf200 Bump version to 1.4.1 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2021-01-25 10:42:32 +00:00
Kevin Klues
e8aa3cc8c3 Merge branch 'ignore-nvidia-visible-devices' into 'master' 
Ignore NVIDIA_VISIBLE_DEVICES for containers with insufficent privileges

See merge request nvidia/container-toolkit/container-toolkit!25
 2021-01-25 10:25:00 +00:00
Evan Lezar
fc408a32c7 Add utility function to get config name from struct 
Signed-off-by: Evan Lezar <elezar@nvidia.com>
 2021-01-22 16:08:45 +01:00
Evan Lezar
f6b1b1afad Ignore NVIDIA_VISIBLE_DEVICES for containers with insufficent privileges 
This change ignores the value of NVIDIA_VISIBLE_DEVICES instead of
raising an error when launching a container with insufficient permissions.

This changes the behaviour under the following conditions:

NVIDIA_VISIBLE_DEVICES is set
and

accept-nvidia-visible-devices-envvar-when-unprivileged = false (default: true)

or

privileged = false (default: false)

This means that a user need not explicitly clear the NVIDIA_VISIBLE_DEVICES
environment variable if no GPUs are to be used in unprivileged containers.
Note that this envvar is set to 'all' by default in many CUDA images that
are used as base images.

Signed-off-by: Evan Lezar <elezar@nvidia.com>
 2021-01-22 15:34:52 +01:00
Kevin Klues
97516467c0 Merge branch 'upstream-bump-v1.4.0' into 'master' 
Bump version to 1.4.0

See merge request nvidia/container-toolkit/container-toolkit!24
 2020-12-14 14:41:02 +00:00
Kevin Klues
01063c0433 Bump version to 1.4.0 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-12-11 18:05:49 +00:00
Kevin Klues
119f75dcf8 Merge branch 'upstream-add-compute-to-default-capabilities' into 'master' 
Add 'compute' capability to list of defaults.

See merge request nvidia/container-toolkit/container-toolkit!23
 2020-12-08 11:31:27 +00:00
Kevin Klues
20604621e4 Add 'compute' capability to list of defaults. 
For most practical purposes, it should be fine to set
NVIDIA_DRIVER_CAPABILITIES=all nowadays.

Historically, these different capabilities exist because they were added
incrementally, with varying degrees of stability. It's fairly common to
run with GPUs in containers today, but a few years ago the driver didn't
support them very well, and it was important to make sure the libraries
being injected into the container actually worked in a containerized
environment. When they didn't, it was common to get information leaks,
crashes, or even silent failures.

In the past, whenever a new set of libraries was being vetted for
injected, a new capability was added to make sure that users had control
to explicitly include only those libraries they were comfortable having
injected into their containers.

The idea being that whoever puts together a container image for use with
GPUs should have the knowledge of what capabilities the software in that
container image requires, and can set the NVIDIA_DRIVER_CAPABILITIES
envvar in that image appropriately.

After some back and forth, we've decided it doesn't quite make sense to
set it to "all" just yet, but we should set it to "utility, compute"
instead of just "utility", so that at least the core CUDA libraries work
by default (once installed in the container).

Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-12-07 12:10:23 +00:00
Kevin Klues
8cfb3c29f6 Merge branch 'upstream-bump-v1.3.0' into 'master' 
Bump to version 1.3.0

See merge request nvidia/container-toolkit/container-toolkit!22
 2020-09-16 13:34:37 +00:00
Kevin Klues
98e202d0d8 Bump to version 1.3.0 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-09-16 11:45:31 +00:00
Kevin Klues
26668097c4 Merge branch 'upstream-bump-1.3.0-rc.2' into 'master' 
Bump to version 1.3.0 rc.2

See merge request nvidia/container-toolkit/container-toolkit!21
 2020-08-10 15:33:25 +00:00
Kevin Klues
caf2792463 Update changelogs for 1.3.0-rc.2 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-08-10 13:08:17 +00:00
Kevin Klues
b2be0b08ac Bump version to 1.3.0-rc.2 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-08-10 13:03:00 +00:00
Kevin Klues
edc5041636 Merge branch 'upstream-update-devices-from-volume-mounts-semantics' into 'master' 
Refactor accepting device lists from volume mounts as a boolean

See merge request nvidia/container-toolkit/container-toolkit!20
 2020-08-07 18:40:56 +00:00
Kevin Klues
2c1809475c Add more tests for new semantics with device list from volume mounts 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-08-07 16:30:31 +00:00
Kevin Klues
7c00385797 Refactor accepting device lists from volume mounts as a boolean 
Also hard code the "root" path where these volume mounts will be looked
for rather than making it configurable.

Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-08-07 16:30:19 +00:00
Kevin Klues
322006c361 Merge branch 'upstream-bump-1.3.0-rc.1' into 'master' 
Bump version to 1.3.0-rc.1

See merge request nvidia/container-toolkit/container-toolkit!19
 2020-07-24 20:36:38 +00:00
Kevin Klues
a25017fb8a Merge branch 'upstream-build-prerelease' into 'master' 
Update build system to accept a TAG variable for things like rc.x

See merge request nvidia/container-toolkit/container-toolkit!18
 2020-07-24 20:22:00 +00:00
Kevin Klues
928905ce94 Update changelogs for 1.3.0-rc.1 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 20:10:42 +00:00
Kevin Klues
7ed17bb9ca Bump version to 1.3.0-rc.1 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 20:03:48 +00:00
Kevin Klues
b50d86c174 Update build system to accept a TAG variable for things like rc.x 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 19:54:29 +00:00
Kevin Klues
bf342fb4c9 Merge branch 'upstream-fix-ci' into 'master' 
Generalize CI variables

See merge request nvidia/container-toolkit/container-toolkit!17
 2020-07-24 14:28:49 +00:00
Kevin Klues
1791372f22 Generalize CI variables 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 14:01:39 +00:00
Kevin Klues
4448319605 Merge branch 'upstream-add-alternate-device-list' into 'master' 
Add the ability to pull the device list from mounted files instead of just Envvars

See merge request nvidia/container-toolkit/container-toolkit!15
 2020-07-24 13:18:53 +00:00
Kevin Klues
2ea3150b60 Merge branch 'upstream-simplify-nvidia-config-generation' into 'master' 
Simplify logic for `nvidiaConfig` generation

See merge request nvidia/container-toolkit/container-toolkit!14
 2020-07-24 13:18:35 +00:00
Kevin Klues
32b4b09bc9 Add tests to verify priority of device list from mounts vs. envvar 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
cc0a22a6d9 Consolidate logic for building nvidiaConfig into a single function 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
e48d23d107 Add test for getDevicesFromMounts() 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
430dda41e9 Remove getNvidiaConfigLegacy() function 
A subsequent commit will add equivalent functionality back in

Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
8bcd02ee5d Add logic implementing getDevicesFromMounts() 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
4791fab747 Simplify getMigConfigDevices() and getMigMonitorDevices() 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
7313069d4c Update getDevices() to account for getting the devices list from mounts 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
a24b0c8b4e Split isLegacyCUDAImage() into its own helper function 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
f46d1861d3 Add stub implementation for getDevicesFromMounts() 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
0a9dc3c653 Add test to make sure that getNvidiaConfig() operates as expected 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
889ebae1fe Pull logic to get the device list from ENVVARs out to its own function 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
e4b9318de3 Only run gofmt over go files under pkg/ in CI 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
aec9a28bc3 Push HookConfig and privileged flags down to getDevices() call 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
2ae7cb07cf Add ability to consider container mounts to generate nvidiaConfig 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
da36874e91 Add new config options to pull device list from mounted files not ENVVAR 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
b9ef2db205 Remove unnecessary files from version control 
Signed-off-by: Kevin Klues <kklues@nvidia.com>
 2020-07-24 12:50:05 +00:00
Kevin Klues
da6fbb343a Revert "Add ability to merge envars of the form NVIDIA_VISIBLE_DEVICES_*" 
This reverts commit 01b4381282.
 2020-07-24 12:50:05 +00:00
Kevin Klues
647a805341 Merge branch 'upstream-add-ci-tests' into 'master' 
Add common CI tests for things like golint, gofmt, unit tests, etc.

See merge request nvidia/container-toolkit/container-toolkit!16
 2020-07-24 12:39:45 +00:00