7f6b021cef
Add h1 and m5 AWS instance types
2017-12-04 13:16:48 -05:00
7942427a17
Set code types for syntax highlighting in README.md
2017-11-28 12:47:34 +00:00
78a04a81d8
Fix image name for azure
...
The official image name sould be used instead of a pernal one.
2017-11-24 16:22:57 +08:00
d31e65f037
Merge pull request #483 from miry/aws-c5-instances
...
Added aws c5 instance types.
2017-11-23 01:13:38 +01:00
fa0a66396d
Updated aws instance types.
2017-11-22 22:42:30 +01:00
8e1c712f39
Added aws c5 instance types.
2017-11-22 08:28:55 +01:00
9618cc61cb
Add deploy guide for Azure
2017-11-16 21:33:59 +08:00
d53b58ae5e
Fix unit testing
2017-11-16 21:31:49 +08:00
03f1b91c89
Convert instance ID to lower case.
...
This is because instance.ID is in different in different API calls (e.g. GET and LIST).
2017-11-16 21:31:49 +08:00
8f7d35b4e0
Enable azure options for autoscaler
2017-11-16 21:31:49 +08:00
4cc7ac56b0
Add unit tests for azure provider
2017-11-16 21:27:55 +08:00
3a72e97db9
Add azure cloud provider
2017-11-16 21:27:55 +08:00
6c8d3aa37d
Fix unit static autoscaler unit tests.
2017-11-15 16:13:18 +01:00
4c83330c29
Fix GCE pricing model to properly evaluate GPUs
2017-11-10 17:38:47 +01:00
2b28ac1a04
Add a workaround for scaling of VMs with GPUs
...
When a machine with GPU becomes ready it can take
up to 15 minutes before it reports that GPU is allocatable.
This can cause Cluster Autoscaler to trigger a second
unnecessary scale up.
The workaround sets allocatable to capacity for GPU so that
a node that waits for GPUs to become ready to use will be
considered as a place where pods requesting GPUs can be
scheduled.
2017-11-06 16:04:22 +01:00
4104a91991
more spelling fixes
2017-11-02 14:21:36 -07:00
3d043f73cb
Renaming the interface function to Cleanup() for CloudProvider type
2017-11-01 12:41:13 -04:00
77aa30a5c1
Fixing for issue 252 by implementing a channel to stop the go routine
2017-11-01 11:00:00 -04:00
5d599ca678
Merge pull request #429 from MaciekPytel/update_nap_api_again
...
update GKE cloudprovider to match changes in GKE API
2017-11-01 13:37:36 +08:00
b21a11cebe
Add a check against dereferencing nil in GKE cloudprovider
2017-10-31 21:38:50 +01:00
c25a0a6ec2
Call location API for regional clusters.
2017-10-31 10:24:00 +01:00
50587992ae
Update fetching ResourceLimits from GKE to use GB not MB
2017-10-30 19:02:41 +01:00
26b2de9b63
Cluster Autoscaler integration with regional clusters.
...
Make GKE cloud provider work with locations instead of zones.
Use beta GKE endpoint for regional clusters.
2017-10-30 14:48:59 +01:00
8b8599e5af
Merge pull request #423 from krzysztof-jastrzebski/resource_limit3
...
Gets resource limits from GKE API.
2017-10-30 12:21:35 +01:00
ea9aa6fe14
update aws instance types
2017-10-28 21:47:24 -04:00
aee8682cb7
Gets resource limits from GKE API.
2017-10-26 14:05:56 +02:00
e8e6ad1c7a
Adds gke-api-endpoint flag to GCE could provider.
2017-10-26 10:32:00 +02:00
07511f444a
Add Refresh method to cloud provider
...
This can be used to dynamically update cloud provider
config (in particular list of managed NodeGroups and their
min/max constraints).
Add GKE implementation.
2017-10-24 18:36:29 +02:00
596f478e63
Merge pull request #414 from krzysztof-jastrzebski/resource_limit
...
Adds resource limits to cloud provider.
2017-10-23 20:38:04 +02:00
56ac572666
Adds resource limits to cloud provider.
2017-10-23 16:06:56 +02:00
7b95e71315
Use GKE alpha client when autoprovisioning is enabled
2017-10-23 15:21:02 +02:00
f752fdfa8a
fix typo
2017-10-20 10:02:09 +02:00
68647fa22b
Fix GPU resource name.
2017-10-19 09:26:34 +02:00
2418dba233
Merge pull request #401 from bskiba/gpu
...
Support GPUs in scale from 0
2017-10-18 11:28:45 +02:00
70287e486f
Support GPUs in scale to/from 0
2017-10-17 16:11:00 +02:00
91759d41cf
GKE cloudprovider bugfixes
...
* Make autoprovisioned node pools names shorter (previously it
was exceeding max name length).
* Set autoscaling fields when creating node pool
2017-10-17 14:02:46 +02:00
0d791cda3c
Fix some bugs in GKE cloudprovider
...
- Mig.TargetSize tried to query non-existing mig if autoprovisioning
was enabled.
- GKE client was created using hosted master project instead of users
project.
2017-10-16 14:03:40 +02:00
fa7f71ac4b
Refresh mig cache if node pool config has changed
2017-10-03 16:30:47 +02:00
f7b7755bbd
Merge pull request #377 from bskiba/nap-allocatable
...
Compute allocatable for auto-provisioned migs
2017-09-27 12:51:50 +02:00
30ae3fa837
Compute allocatable for auto-provisioned migs
2017-09-27 12:07:23 +02:00
16e9106c07
Fix setting target size for group in core/static_autoscaler_test.go.
2017-09-26 10:58:00 +02:00
1d0c237adc
Cloudprovider/gce/gce_cloud_provider.go unit tests.
2017-09-20 07:57:43 +02:00
6b8b8b8fe1
Cloudprovider/gce/gce_manager.go unit tests.
2017-09-19 11:16:08 +02:00
1368f51d53
Fix calculating autoprovisioned property for node pool.
2017-09-15 17:12:56 +02:00
7cf36de42b
Merge pull request #343 from bskiba/allocatable
...
Compute allocatable for scale up from 0 based on kube-reserved
2017-09-15 16:33:42 +02:00
90389930b7
Compute allocatable based on kube-reserved
2017-09-14 16:21:51 +02:00
530ce16417
Cloudprovider unit tests.
2017-09-13 10:46:56 +02:00
8309630048
fix formatting
2017-09-10 12:39:09 -04:00
d87a3e8615
add gpu to node template
2017-09-10 12:25:51 -04:00
829ec9a0ef
Merge pull request #299 from mwielgus/update-test-provider
...
Expand test cloud provider with autoprovisioning options
2017-09-05 17:33:50 +05:30
Timothy Palpant
Gary Rennie
Pengfei Ni
Marcin Wielgus
Michael Nikitochkin
Krzysztof Jastrzebski
Maciej Pytel
Beata Skiba
Edward Tsang
mmerrill3
Seth Pollack