github/kubernetes
4
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
44,995 Commits 1 Branch 31 Tags 1,019 MiB
e9e8fe6c32
Commit Graph

77 Commits

Author SHA1 Message Date
Anthony Yeh
e9e8fe6c32 RC/RS: Fix ignoring inactive Pods. 2017-03-06 15:51:53 -08:00
Anthony Yeh
2c2fc9c707 RC/RS: Don't log Pod events unless some RC/RS actually cares. 2017-03-01 15:43:51 -08:00
Anthony Yeh
01d025a7cc ControllerRefManager: Don't always filter inactive Pods. 
Some controllers, like DaemonSet, want to see all Pods.
 2017-03-01 15:43:51 -08:00
Anthony Yeh
db6665251a RC/RS: Don't requeue on error inside sync function. 
Returning an error from the sync function already triggers a requeue
in processNextWorkItem().
 2017-03-01 15:43:51 -08:00
Anthony Yeh
ca13b9e532 RC/RS: Use ControllerRef to route watch events. 
This is part of the completion of ControllerRef, as described here:

https://github.com/kubernetes/community/blob/master/contributors/design-proposals/controller-ref.md#watches

This also removes the need for the Pod->Controller mapping cache in RC
and RS. This mapping is now persisted in the Pod's ControllerRef
instead.
 2017-03-01 15:43:50 -08:00
Chao Xu
c3baf402f5 gc changes 2017-02-28 23:05:41 -08:00
Kubernetes Submit Queue
2681e38d3a Merge pull request #42097 from kargakis/address-mismatched-available-replicas 
Automatic merge from submit-queue

Enqueue controllers after minreadyseconds when all pods are ready

@janetkuo this should address https://github.com/kubernetes/kubernetes/issues/41697#issuecomment-281851377. Impossible to unit test this but it should stabilize some of our deployment e2e tests that occasionally fail because of availableReplicas not being updated.

It should also fix https://github.com/kubernetes/kubernetes/issues/41641

Eventually I would like AddAfter to be able to cancel previous invocations of the same key so I opened https://github.com/kubernetes/client-go/issues/131

@kubernetes/sig-apps-bugs
 2017-02-27 22:09:46 -08:00
Michail Kargakis
9eab226947 Enqueue controllers after minreadyseconds when all pods are ready 2017-02-27 19:28:15 +01:00
Mayank Kumar
6b35ff72ce Enable Garbage collection by default for RS and RC 2017-02-23 22:37:52 -08:00
Anthony Yeh
70bd5fdfe5 Refactor ControllerRefManager 
To prepare for implementing ControllerRef across all controllers,
this pushes the common adopt/orphan logic into ControllerRefManager
so each controller doesn't have to duplicate it.

This also shares the adopt/orphan logic between Pods and ReplicaSets,
so it lives in only one place.
 2017-02-16 15:09:17 -08:00
deads2k
fd34b11e13 react to informer updates 2017-02-13 09:18:32 -05:00
Andy Goldstein
70c6087600 Replace hand-written informers with generated ones 
Replace existing uses of hand-written informers with generated ones.
 Follow-up commits will switch the use of one-off informers to shared
 informers.
 2017-02-06 13:49:27 -05:00
deads2k
8a12000402 move client/record 2017-01-31 19:14:13 -05:00
deads2k
2c1c0f3f72 move workqueue to client-go 2017-01-30 09:08:21 -05:00
deads2k
b0b156b381 make tools/cache authoritative 2017-01-25 08:29:45 -05:00
deads2k
1ce0637b27 move listers out of cache to reduce import tree 2017-01-20 15:01:38 -05:00
deads2k
77b4d55982 mechanical 2017-01-16 09:35:12 -05:00
deads2k
6a4d5cd7cc start the apimachinery repo 2017-01-11 09:09:48 -05:00
NickrenREN
639572ac68 fix redundant alias and remove unused function 2017-01-09 17:13:09 +08:00
Michail Kargakis
e5b586b5b0 Share rc cache from the rc manager 2017-01-03 16:59:09 +01:00
Michail Kargakis
7ef3e6f7c9 controller: wait for all pods to be deleted before Recreating 2016-12-15 19:55:18 +01:00
Chao Xu
03d8820edc rename /release_1_5 to /clientset 2016-12-14 12:39:48 -08:00
Kubernetes Submit Queue
15f9572b8c Merge pull request #38613 from kargakis/do-not-adopt-when-deleted 
Automatic merge from submit-queue

controller: adopt pods only when controller is not deleted

When a replica set is deleted it will continue adopting pods thus driving the worker that handles it in erroring out because the adoption is [always cancelled](59c313730c/pkg/controller/controller_ref_manager.go (L110)) in the controller reference manager.
```
E1212 14:40:31.245773    7964 replica_set.go:616] cancel the adopt attempt for pod e2e-tests-deployment-2rr3m_test-rollover-deployment-1981456318-73c3m_791e16cb-c070-11e6-a234-68f72840e7df because the controlller is being deleted
E1212 14:40:31.258462    7964 replica_set.go:616] cancel the adopt attempt for pod e2e-tests-deployment-2rr3m_test-rollover-deployment-1981456318-73c3m_791e16cb-c070-11e6-a234-68f72840e7df because the controlller is being deleted
E1212 14:40:31.259131    7964 replica_set.go:616] cancel the adopt attempt for pod e2e-tests-deployment-2rr3m_test-rollover-deployment-1981456318-73c3m_791e16cb-c070-11e6-a234-68f72840e7df because the controlller is being deleted
E1212 14:40:31.259149    7964 replica_set.go:616] cancel the adopt attempt for pod e2e-tests-deployment-2rr3m_test-rollover-deployment-1981456318-wrmt8_791e3d46-c070-11e6-a234-68f72840e7df because the controlller is being deleted
I1212 14:40:31.268012    7964 deployment_controller.go:314] Error syncing deployment e2e-tests-deployment-2rr3m/test-rollover-deployment: Operation cannot be fulfilled on deployments.extensions "test-rollover-deployment": the object has been modified; please apply your changes to the latest version and try again
E1212 14:40:31.277252    7964 replica_set.go:616] cancel the adopt attempt for pod e2e-tests-deployment-2rr3m_test-rollover-deployment-1981456318-73c3m_791e16cb-c070-11e6-a234-68f72840e7df because the controlller is being deleted
E1212 14:40:31.277276    7964 replica_set.go:616] cancel the adopt attempt for pod e2e-tests-deployment-2rr3m_test-rollover-deployment-1981456318-wrmt8_791e3d46-c070-11e6-a234-68f72840e7df because the controlller is being deleted
E1212 14:40:31.277287    7964 replica_set.go:616] cancel the adopt attempt for pod e2e-tests-deployment-2rr3m_test-rollover-deployment-1981456318-bmqpn_81482114-c070-11e6-a234-68f72840e7df because the controlller is being deleted
E1212 14:40:31.289148    7964 replica_set.go:616] cancel the adopt attempt for pod e2e-tests-deployment-2rr3m_test-rollover-deployment-1981456318-b6s4x_82fa8343-c070-11e6-a234-68f72840e7df because the controlller is being deleted
E1212 14:40:31.289169    7964 replica_set.go:616] cancel the adopt attempt for pod e2e-tests-deployment-2rr3m_test-rollover-deployment-1981456318-73c3m_791e16cb-c070-11e6-a234-68f72840e7df because the controlller is being deleted
E1212 14:40:31.289176    7964 replica_set.go:616] cancel the adopt attempt for pod e2e-tests-deployment-2rr3m_test-rollover-deployment-1981456318-wrmt8_791e3d46-c070-11e6-a234-68f72840e7df because the controlller is being deleted
E1212 14:40:31.289181    7964 replica_set.go:616] cancel the adopt attempt for pod e2e-tests-deployment-2rr3m_test-rollover-deployment-1981456318-bmqpn_81482114-c070-11e6-a234-68f72840e7df because the controlller is being deleted
```

@kubernetes/deployment @caesarxuchao
 2016-12-13 04:57:49 -08:00
Michail Kargakis
ec2c79a35e controller: adopt pods only when controller is not deleted 2016-12-12 15:12:44 +01:00
Michail Kargakis
9c7b39066e Log enqueueing replica sets for availability checks 2016-12-12 14:09:16 +01:00
Clayton Coleman
c52d510a24
refactor: generated 2016-12-10 18:05:53 -05:00
Michail Kargakis
267dae6435 controller: requeue replica sets for availability checks 2016-12-05 02:41:15 +01:00
Clayton Coleman
3454a8d52c
refactor: update bazel, codec, and gofmt 2016-12-03 19:10:53 -05:00
Clayton Coleman
5df8cc39c9
refactor: generated 2016-12-03 19:10:46 -05:00
Clayton Coleman
35a6bfbcee
generated: refactor 2016-11-23 22:30:47 -06:00
Chao Xu
7eeb71f698 cmd/kube-controller-manager 2016-11-23 15:53:09 -08:00
Michail Kargakis
2491216222 Replica set/rc controller changes for Conditions 2016-11-02 10:30:09 +01:00
Chao Xu
850729bfaf include multiple versions in clientset 
update client-gen to use the term "internalversion" rather than "unversioned";
leave internal one unqualified;
cleanup client-gen
 2016-10-29 13:30:47 -07:00
Jan Chaloupka
6079053407 Update clientset generator to use RESTClient interface instead of the RESTClient data type 2016-10-21 10:13:51 +02:00
deads2k
b471398f1f convert replica set controller to shared informer 2016-10-18 08:13:37 -04:00
m1093782566
a76a743ecd fix replica set hot loop 
Change-Id: I5176eb9350324de8e7b2035686c4e2c2abc5ef3d
 2016-10-08 15:59:22 +08:00
deads2k
c30b2efc46 update replicaset lister 2016-10-05 15:20:27 -04:00
Michail Kargakis
f7c232b8c6 extensions: add minReadySeconds/availableReplicas to replica sets 2016-09-28 11:06:40 +02:00
Mike Danese
a765d59932 move informer and controller to pkg/client/cache 
Signed-off-by: Mike Danese <mikedanese@google.com>
 2016-09-15 12:50:08 -07:00
deads2k
4317173d3f add names for workqueues to gather controller latency/depth metrics 2016-08-30 09:51:50 -04:00
Wojciech Tyczynski
e9d5be628a Don't validate selector that is already validated 2016-08-22 09:39:32 +02:00
Michail Kargakis
ab3bc03a4f api: add readyReplicas in RC/RS 2016-08-21 14:49:45 +02:00
Wojciech Tyczynski
331083727f Change podNamespacer API 2016-08-17 16:55:01 +02:00
Wojciech Tyczynski
acfd8c787f Avoid computing super-expensive DeepEqual on every pod update 2016-08-11 08:35:36 +02:00
Dominika Hodovska
816f6d32ca Collapse duplicate informer creation paths 2016-08-04 09:02:13 +02:00
Michal Rostecki
59ca5986dd Print/log pointers of structs with %#v instead of %+v 
There are many places in k8s where %+v is used to format a pointer
to struct, which isn't working as expected.

Fixes #26591
 2016-08-01 22:27:56 +02:00
Wojciech Tyczynski
85de930a66 ReplicaSet controller can set/remove ControllerRef 2016-07-21 11:36:52 +02:00
Wojciech Tyczynski
2794cf538c Use sharedPodInformer in ReplicaSet controller 2016-07-20 12:25:26 +02:00
David McMahon
ef0c9f0c5b Remove "All rights reserved" from all the headers. 2016-06-29 17:47:36 -07:00
Chao Xu
63fb075f0a fix updatePod of replication controller manager and replica set controller to 
handle pod label updates that match no rc or rs
 2016-06-15 10:34:26 -07:00