a849ead286
Precompute inter pod equivalence groups in checkPodsSchedulableOnNode
2019-05-29 18:05:52 +02:00
92ea680f1a
Implement an --ignore-taint flag
...
This change adds support for a user to specify taints to ignore when
considering a node as a template for a node group.
2019-05-14 10:22:59 -07:00
90666881d3
Move GPULabel and GPUTypes to cloud provider
2019-03-25 13:03:01 -07:00
5ae76ea66e
UPSTREAM: <carry>: fix max cluster size calculation on scale up
...
When scaling up the calculation for computing the maximum cluster size
does not take into account the number of any upcoming nodes and it is
possible to grow the cluster beyond the cluster
size (--max-nodes-total).
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1670695
2019-03-08 13:28:58 +00:00
128729bae9
Move schedulercache to package nodeinfo
2019-02-21 12:41:08 +08:00
79ef3a6940
unexporting methods in utils.go
2019-01-25 00:06:03 +05:30
85a83b62bd
Pass nodeGroup->NodeInfo map to ClusterStateRegistry
...
Change-Id: Ie2a51694b5731b39c8a4135355a3b4c832c26801
2019-01-08 15:52:00 +01:00
4002559a4c
Merge pull request #1516 from frobware/fix-max-nodes-total-upstream
...
fix calculation of max cluster size
2019-01-03 10:02:38 -08:00
3f0da8947a
Use listers in scale-up
2019-01-02 15:56:01 +01:00
ab7f1e69be
Merge pull request #1464 from losipiuk/lo/stockouts2
...
Better quota-exceeded/stockout handling
2018-12-31 05:28:08 -08:00
9689b30ee4
Do not use time.Now() in RegisterFailedScaleUp
2018-12-28 17:17:07 +01:00
da5bef307b
Allow updating Increase for ScaleUpRequest in ClusterStateRegistry
2018-12-28 17:17:07 +01:00
60babe7158
Use kubernetes lister for daemonset instead of custom one
...
Also migrate to using apps/v1.DaemonSet instead of old
extensions/v1beta1.
2018-12-28 13:55:41 +01:00
5bc77f051c
UPSTREAM: <carry>: fix calculation of max cluster size
...
When scaling up, the calculation for the maximum size of the cluster
based on `--max-nodes-total` doesn't take into account any nodes that
are in the process of coming up. This allows the cluster to grow
beyond the size specified.
With this change I now see:
scale_up.go:266] 21 other pods are also unschedulable
scale_up.go:423] Best option to resize: openshift-cluster-api/amcdermo-ca-worker-us-east-2b
scale_up.go:427] Estimated 18 nodes needed in openshift-cluster-api/amcdermo-ca-worker-us-east-2b
scale_up.go:432] Capping size to max cluster total size (23)
static_autoscaler.go:275] Failed to scale up: max node total count already reached
2018-12-18 17:05:19 +00:00
016bf7fc2c
Use k8s.io/klog instead github.com/golang/glog
2018-11-26 17:30:31 +01:00
7008fb50be
Merge pull request #1380 from losipiuk/lo/backoff
...
Make Backoff interface
2018-11-07 05:13:43 -08:00
6febc1ddb0
Fix formatted log messages
2018-11-06 14:51:43 +01:00
bf6ff4be8e
Clean up estimators
2018-11-06 14:15:42 +01:00
0e2c3739b7
Use NodeGroup as key in Backoff
2018-10-30 18:17:26 +01:00
55fc1e2f00
Store NodeGroup in ScaleUpRequest and ScaleDownRequest
2018-10-30 18:03:04 +01:00
6f5e6aab6f
Move node group balancing to processor
...
The goal is to allow customization of this logic
for different use-case and cloudproviders.
2018-10-25 14:04:05 +02:00
a266420f6a
Recalculate clusterStateRegistry after adding multiple node groups
2018-10-02 17:15:20 +02:00
437efe4af6
If possible use nodeInfo based on created node group
2018-10-02 15:46:45 +02:00
8179e4e716
Refactor the scale-(up|down) status processors so that they have more info available
...
Replace the simple boolean ScaledUp property of ScaleUpStatus with a more
comprehensive ScaleUpResult. Add more possible values to ScaleDownResult.
Refactor the processors execution so that they are always executed every
iteration, even if RunOnce exits earlier.
2018-09-20 17:12:02 +02:00
bf8cfef10b
NodeGroupManager.CreateNodeGroup can return extra created node groups.
2018-09-19 13:55:51 +02:00
705a6d87e2
fixup! Call CheckPodsSchedulableOnNode in scale_up.go via caching layer
2018-09-17 13:01:19 +02:00
0ad4efe920
Call CheckPodsSchedulableOnNode in scale_up.go via caching layer
2018-09-13 17:01:15 +02:00
b88e6019f7
code review fixes 3
2018-08-28 18:11:04 +02:00
5620f76c62
Pass NoScaleUpInfo to ScaleUpStatus processor
2018-08-28 14:26:03 +02:00
cd9808185e
Report reason why pod didn't trigger scale-up
2018-08-28 14:11:36 +02:00
398a1ac153
Fix error on node info not found for group
2018-07-23 11:16:12 +02:00
1dd0147d9e
Add more events for CA
2018-07-09 15:42:05 +08:00
800ee56b34
Refactor and extend GPU metrics error types
2018-07-05 13:13:11 +02:00
aae4d1270a
Make GetGpuTypeForMetrics more robust
2018-06-26 21:35:16 +02:00
5eb7021f82
Add GPU-related scaled_up & scaled_down metrics ( #974 )
...
* Add GPU-related scaled_up & scaled_down metrics
* Fix name to match SD naming convention
* Fix import after master rebase
* Change the logic to include GPU-being-installed nodes
2018-06-22 21:00:52 +02:00
99c8c51bb3
Create NodeGroupManager which is responsible for creating/deleting node groups.
2018-06-14 16:11:32 +02:00
b7323bc0d1
Respect GPU limits in scale_up
2018-06-14 15:46:58 +02:00
dfcbedb41f
Take into consideration nodes from not autoscaled groups when enforcing resource limits
2018-06-14 15:31:40 +02:00
9f75099d2c
Restructure checking resource limits in scale_up.go
...
Preparatory work for before introducing GPU limits
2018-06-13 19:00:37 +02:00
be3dd85503
Update scheduler cache package
2018-06-11 13:54:12 +08:00
9c61477d25
Do not return error when getting cpu/memory capacity of node
2018-06-08 15:04:57 +02:00
b8ae6df5d3
Add post scale up status processor.
2018-06-06 13:34:49 +02:00
856855987b
Move some GKE-specific logic outside core
...
No change in actual logic being executed. Added a new
NodeGroupListProcessor interface to encapsulate the existing logic.
Moved PodListProcessor and refactor how it's passed around
to make it consistent and easy to add similar interfaces.
2018-05-29 12:57:19 +02:00
6761d7f354
Execute predicates only for similar pods.
2018-05-29 09:36:11 +02:00
4c710950de
Move ClusterStateRegistry to StaticAutoscaler
...
AutoscalingContext is basically a configuration and few static helpers
and API handles.
ClusterStateRegistry is state and thus moved to other state-keeping
objects.
2018-05-24 13:03:01 +02:00
bfb70e40ee
Allow passing taints to Node Group creation.
2018-05-18 14:33:33 +02:00
88b769b324
Refactor cluster autoscaler builder and add pod list processor.
2018-04-26 12:37:51 +02:00
feb4ad9e14
Add utility for limiting logging
2018-03-22 12:57:22 +01:00
04bec08e84
Compilation fix
2018-03-20 20:11:36 +01:00
b7f8622eb2
Create node groups with GPU in scale-up.go
...
This is still not implemented in cloudprovider.
Extended NewNodeGroup inteface to have a way of passing
parameters for more complex resources.
2017-12-11 13:12:22 +01:00
Łukasz Osipiuk
Chris Bradfield
Jiaxin Shan
Andrew McDermott
Pengfei Ni
Vivek Bagade
Kubernetes Prow Robot
Maciej Pytel
Aleksandra Malinowska
Łukasz Osipiuk
Jakub Tużnik
Karol Gołąb
Krzysztof Jastrzebski
Beata Skiba
Joachim Bartosik
Marcin Wielgus