Kubernetes Submit Queue
06c1f2ba2c
Merge pull request #34707 from gmarek/master
...
Automatic merge from submit-queue
Bump log level in case of Node eviction
2016-10-13 05:37:10 -07:00
gmarek
41278b4c6b
Bump log level in case of Node eviction
2016-10-13 13:26:16 +02:00
gmarek
8b7e9d303c
Handle DeletedFinalStateUnknown in NodeController
2016-10-13 11:31:04 +02:00
Justin Santa Barbara
dc2a9d6f5c
Fix typo: initilizing -> initializing
2016-10-07 13:29:22 -04:00
deads2k
0961784a9b
switch node controller to shared informers
2016-09-29 09:16:41 -04:00
gmarek
cb0a13c1e5
Move orphaned Pod deletion logic to PodGC
2016-09-28 13:58:31 +02:00
Justin Santa Barbara
54195d590f
Use strongly-typed types.NodeName for a node name
...
We had another bug where we confused the hostname with the NodeName.
To avoid this happening again, and to make the code more
self-documenting, we use types.NodeName (a typedef alias for string)
whenever we are referring to the Node.Name.
A tedious but mechanical commit therefore, to change all uses of the
node name to use types.NodeName
Also clean up some of the (many) places where the NodeName is referred
to as a hostname (not true on AWS), or an instanceID (not true on GCE),
etc.
2016-09-27 10:47:31 -04:00
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
gmarek
c40a36cab0
Change the eviction metric type and fix rate-limited-timed-queue
2016-09-08 12:20:51 +02:00
Kubernetes Submit Queue
9dfe8df7cd
Merge pull request #32107 from xingzhou/km_bug
...
Automatic merge from submit-queue
Used goroutine to launch node controller's internalPodInformer.
Fixes #32103
2016-09-06 11:50:28 -07:00
Kubernetes Submit Queue
afef4b6938
Merge pull request #32070 from gmarek/nodecontroller
...
Automatic merge from submit-queue
Sleep between NodeStatus update retries
Just a thing I found when looking into other problems.
This is pretty much no-risk change fixing wrong behavior. Do you think it should go in 1.4? @pwittrock
2016-09-06 03:51:20 -07:00
Xing Zhou
46c302b4c2
Used goroutine to launch node controller's internalPodInformer.
...
Node controller's internalPodInformer will block main thread
if it is not started as a go routine. This patch fixed this
by runing internalPodInformer as a go routine.
2016-09-06 15:39:13 +08:00
Wojciech Tyczynski
b69d516763
NodeController listing nodes from apiserver cache
2016-09-05 16:52:57 +02:00
gmarek
bac603afd6
Sleep between NodeStatus update retries
2016-09-05 12:29:28 +02:00
gmarek
ea2d19f5d7
Remove unused argument to NodeController.Run
2016-08-30 14:24:56 +02:00
gmarek
5d8cb17efa
Add cluster health metrics to NodeController
2016-08-18 15:11:10 +02:00
Kubernetes Submit Queue
f9190ed61a
Merge pull request #30138 from gmarek/flags
...
Automatic merge from submit-queue
Expose flags for new NodeEviction logic in NodeController
Fix #28832
Last PR from the NodeController NodeEviction logic series.
cc @davidopp @lavalamp @mml
2016-08-18 00:41:28 -07:00
bprashanth
15c9826061
Nodecontroller doesn't flip readiness on pods if kubeletVersion < 1.2.0
2016-08-17 15:33:35 -07:00
gmarek
4cf698ef04
Expose flags for new NodeEviction logic in NodeController
2016-08-17 10:43:24 +02:00
AdoHe
b2ab4c6d9b
fix node controller event uid issue
2016-08-14 09:41:20 +08:00
Dominika Hodovska
816f6d32ca
Collapse duplicate informer creation paths
2016-08-04 09:02:13 +02:00
gmarek
66224ce0bd
Change eviction logic in NodeController and make it Zone-aware
2016-08-02 14:21:52 +02:00
k8s-merge-robot
59836d6dbd
Merge pull request #24841 from sjenning/shared-informer
...
Automatic merge from submit-queue
update node controller to use shared pod informer
continuing work from #24470 and #23575
2016-08-02 03:45:01 -07:00
Matt T. Proud
76aab29ede
pkg/controller/node/nodecontroller: simplify mutex
...
Similar to #29598 , we can rely on the zero-value construction behavior
to embed `sync.Mutex` into parent structs.
2016-07-26 07:06:16 +02:00
Seth Jennings
db6026c82a
node controller use shared pod informer
2016-07-20 15:26:19 -05:00
Seth Jennings
6d77f53af4
refactor maybeDeleteTerminatingPod
2016-07-20 15:26:19 -05:00
gmarek
56006fac43
Retry assigning CIDRs
2016-07-18 17:06:04 +02:00
Prashanth Balasubramanian
2f9516db30
List all nodes and occupy cidr map before starting allocations
2016-07-16 13:54:01 -07:00
gmarek
f6b1c316e9
Allow switching rate limiter inside RateLimitedQueue
2016-07-14 15:38:14 +02:00
gmarek
5677a9845e
Split NodeController rate limiters between zones
2016-07-13 14:09:19 +02:00
gmarek
fd600ab65c
Add hooks for cluster health detection
2016-07-12 15:10:58 +02:00
gmarek
7524da877e
Reduce tightness of coupling in NodeController
2016-07-12 11:00:41 +02:00
gmarek
7f5f9d3a6f
Move CIDR allocation logic away from nodecontroller.go
2016-07-12 09:40:43 +02:00
David McMahon
ef0c9f0c5b
Remove "All rights reserved" from all the headers.
2016-06-29 17:47:36 -07:00
Oleg Shaldybin
a58b4cf59d
Don't panic in NodeController if pod update fails
...
Previously it was trying to use a nil pod variable if error was returned
from the pod update call.
2016-06-28 11:54:13 -07:00
goltermann
218645b346
Fix several spelling errors in comments.
2016-06-17 10:41:18 -07:00
gmarek
7cac170214
AllocateOrOccupyCIDR returs quickly
2016-05-31 09:11:42 +02:00
k8s-merge-robot
577cdf937d
Merge pull request #26415 from wojtek-t/network_not_ready
...
Automatic merge from submit-queue
Add a NodeCondition "NetworkUnavaiable" to prevent scheduling onto a node until the routes have been created
This is new version of #26267 (based on top of that one).
The new workflow is:
- we have an "NetworkNotReady" condition
- Kubelet when it creates a node, it sets it to "true"
- RouteController will set it to "false" when the route is created
- Scheduler is scheduling only on nodes that doesn't have "NetworkNotReady ==true" condition
@gmarek @bgrant0607 @zmerlynn @cjcullen @derekwaynecarr @danwinship @dcbw @lavalamp @vishh
2016-05-29 03:06:59 -07:00
Alex Robinson
d577550dd0
Merge pull request #26054 from gmarek/flags
...
Make service-range flag in controller-manager optional
2016-05-27 14:26:15 -07:00
gmarek
7bdf480340
Node is NotReady until the Route is created
2016-05-27 19:29:51 +02:00
Zach Loafman
cb69960742
nodecontroller: Fix log message on successful update
2016-05-25 14:44:15 -07:00
gmarek
08385b2c5f
Make service-range flag in controller-manager optional
2016-05-23 09:37:53 +02:00
gmarek
1d89d2f2d2
Add few log lines to NodeController
2016-05-23 08:49:11 +02:00
mqliang
17d5a302bb
make podcidr mask size configurable
2016-05-20 20:44:40 +08:00
mqliang
cf7a3475f3
Don't allow node controller to allocate into service CIDR range
2016-05-20 20:44:40 +08:00
mqliang
69b8453fa0
cidr allocator
2016-05-20 20:44:40 +08:00
gmarek
6d27009db1
NodeController doesn't evict Pods if no Nodes are Ready
2016-05-17 23:03:21 +02:00
mqliang
c10f43a2e5
implement AddIndexers for SharedIndexInformer
2016-05-06 21:23:18 +08:00
mqliang
9011207f18
add namespace index to rc and pod
2016-05-06 17:12:36 +08:00
gmarek
3171aac57c
Generated clients can return their RESTClients, RESTClient can return its RateLimiter
2016-04-27 22:15:10 +02:00
k8s-merge-robot
72e51dacfe
Merge pull request #24034 from AdoHe/log_spam
...
Automatic merge from submit-queue
remove log spam from nodecontroller
@thockin @quinton-hoole ptal.
2016-04-21 12:11:05 -07:00
AdoHe
e52f71f78d
remove log spam from nodecontroller
2016-04-10 22:51:29 -04:00
k8s-merge-robot
7d7ca5ab72
Merge pull request #23608 from caesarxuchao/mv-typed-clients
...
Automatic merge from submit-queue
Move typed clients into clientset folder
Move typed clients from `pkg/client/typed/` to `pkg/client/clientset_generated/${clientset_name}/typed`.
The first commit changes the client-gen, the last commit updates the doc, other commits are just moving things around.
@lavalamp @krousey
2016-04-02 19:31:40 -07:00
Chao Xu
49559a3332
Generate the typed clients under the clientset folder
2016-03-31 15:28:45 -07:00
André Cruz
34a273f1eb
Fixed typo.
2016-03-29 16:10:11 +01:00
k8s-merge-robot
e44ad7a083
Merge pull request #22735 from resouer/throttle-dev
...
Auto commit by PR queue bot
2016-03-26 06:44:48 -07:00
goltermann
32d569d6c7
Fixing all the "composite literal uses unkeyed fields" Vet errors.
2016-03-25 15:25:09 -07:00
harry
8472cfa214
Refactor throttle into util pkg
...
Fix missing throttle.go
2016-03-25 08:32:23 +08:00
Mike Danese
e0431d8409
Revert "Revert "continuously delete pods on nodes that don't exist""
...
This reverts commit da0a72f2c2
2016-03-08 11:00:35 -08:00
Marek Grabowski
da0a72f2c2
Revert "continuously delete pods on nodes that don't exist"
2016-03-08 09:38:53 +01:00
Mike Danese
c404e7c6d1
continuously delete pods on nodes that don't exist
2016-03-07 16:01:33 -08:00
k8s-merge-robot
dc46ae031d
Merge pull request #22336 from cjcullen/evict
...
Auto commit by PR queue bot
2016-03-07 08:42:40 -08:00
CJ Cullen
e7fc608df7
Immediately evict pods and delete node when cloud says node is gone.
2016-03-07 06:07:51 -08:00
Chao Xu
828fe5df2f
remove V(2)
2016-03-03 14:54:33 -08:00
k8s-merge-robot
ec77d0841d
Merge pull request #22376 from caesarxuchao/add-log-nodecontroller
...
Auto commit by PR queue bot
2016-03-02 13:36:34 -08:00
Chao Xu
f8dd7fe1de
log the succussful forceful deletion
2016-03-02 11:18:55 -08:00
CJ Cullen
506df24c1f
Revert "Evict pods w/o rate-limit when cloud says node is gone."
2016-03-01 11:04:18 -08:00
k8s-merge-robot
fec00b535f
Merge pull request #21187 from cjcullen/evict
...
Auto commit by PR queue bot
2016-03-01 10:45:22 -08:00
Kris
e664ef922f
Move restclient to its own package
2016-02-29 12:05:13 -08:00
CJ Cullen
3a8c7a7074
Evict pods w/o rate-limit when cloud says node is gone.
2016-02-29 10:33:50 -08:00
Mike Danese
a50bc3da10
revert deletePods changes in #22100
2016-02-27 08:40:50 -08:00
Mike Danese
c1a7e280a3
fix pod eviction for gracefully terminationg pods
2016-02-26 17:34:15 -08:00
Fabio Yeon
7d0684e9c4
Merge pull request #21628 from smarterclayton/suppress_debug_logging
...
Reduce volume of logs generated at v(3)
2016-02-26 15:47:31 -08:00
gmarek
e99ad585ce
Decrease verbosity in NodeController
2016-02-23 17:03:38 +01:00
Clayton Coleman
ae2f6a833a
Reduce volume of logs generated at v(3)
...
Node controller is generating a huge amount of logging at v(3) that is
more appropriate for v(5). Split the log into two levels and ensure it
also ends up on one line (so grep works).
The pod manager generates a v(4) pod output on sync that always contains
a newline - since the size of the pod is so excessive in output, kick it
to v(5) for deep debugging (we're pretty happy with this loop).
2016-02-20 15:29:05 -05:00
Andy Goldstein
49fb450a36
Fix type for DaemonSet informer
2016-02-18 14:54:39 -05:00
k8s-merge-robot
5acdb92126
Merge pull request #21177 from laushinka/spelling-fixes
...
Auto commit by PR queue bot
2016-02-18 10:29:49 -08:00
gmarek
0d7f3bca0a
Increase logging verbosity to help debugging #21458
2016-02-18 11:25:38 +01:00
laushinka
7ef585be22
Spelling fixes inspired by github.com/client9/misspell
2016-02-18 06:58:05 +07:00
Chao Xu
97aecd002a
remove underscore in imported pkg names
2016-02-16 10:54:51 -08:00
CJ Cullen
52b16129dc
Re-GET nodes during CIDR allocation (to avoid cascading bad resource version).
2016-02-12 13:08:54 -08:00
Jan Chaloupka
4389b3f0d6
Rewritte util.* -> wait.* wherever reasonable
2016-02-07 12:02:20 +01:00
Chao Xu
184440f8ef
rename release_1_2 to internalclientset
2016-02-05 14:02:28 -08:00
Chao Xu
1b047f8e67
rename legacy to core
2016-02-04 14:26:56 -08:00
Chao Xu
f9f5736b01
grep sed
2016-02-03 13:06:07 -08:00
Chao Xu
fe7887f1ec
replace the client with clientset in controllers
2016-02-02 20:28:45 -08:00
harry
1032067ff9
Replace runtime reference by pkg
2016-02-01 21:06:44 +08:00
Matt Liggett
0ba1b49b42
When a node becomes unreachable, do not evict DaemonSet-managed pods.
...
Part of the graduation requirement for DaemonSet spelled out in #15310 .
2016-01-28 11:08:52 -08:00
David Oppenheimer
85f88b8645
Run gofmt.
2016-01-25 23:25:38 -08:00
David Oppenheimer
2a6da5871a
Revert #19580 and #17754 . Fixes #20048 .
2016-01-25 23:19:22 -08:00
Alex Mohr
94b2490eba
Merge pull request #19580 from WeixuZhuang/node
...
resolve the bug when cluster CIDR is not /8
2016-01-21 10:34:19 -08:00
Minhan Xia
01829432db
update pod status once node becomes NotReady
2016-01-14 17:41:36 -08:00
Weixu Zhuang
2af68d25b8
resolve the bug when cluster CIDR is not /8
...
We will have the rigth formula to generate correct maxCIDRs now.
Previous code assume cluster CIDR is /8 which may not be true.
Now it generates maxCIDR based on the info of cluster IP.
2016-01-13 11:55:19 -08:00
Weixu Zhuang
3928bd6e76
Fix TODO in pkg/controller/nodecontroller.go line 472
...
The code now calculates and find out the CIDRs for every node in any sync period.
I will fix this TODO by maintaining a set for available CIDRs left. Firstly, I will
insert 256 CIDRs into the available set. Once someone get one CIDR, remove this CIDR
from the available set. If one node get deleted, we will reinsert the CIDR associates
with this node back to available CIDR. Once there are nothing left in available CIDR set,
generate another 256 CIDRs and insert them into the available set. As a result, we do not
need to generate CIDRs in every monitor process and we only need to assign CIDR to node
which does not have it.
This commit also fix the error that CIDR may overflow when we use the function
generateCIDRs. There will be no more ip overflowing, all assigan CIDR will be valid
2015-12-28 11:15:38 -08:00
Wojciech Tyczynski
960808bf08
Switch to versioned ListOptions in client.
2015-12-14 14:26:09 +01:00
Wojciech Tyczynski
b0fcb5adef
Pass ListOptions to List in ListWatch.
2015-12-07 11:53:53 +01:00
Mike Danese
f784be33ac
actually validate semver in node controller rather than prefix checking
2015-12-03 14:49:44 -08:00
Wojciech Tyczynski
6dcb689d4e
Simplify List() signature in clients.
2015-12-03 09:54:07 +01:00
k8s-merge-robot
8a8639d7af
Merge pull request #17863 from wojtek-t/only_list_options_in_watch
...
Auto commit by PR queue bot
2015-12-02 06:28:28 -08:00
Wojciech Tyczynski
8343c8ce6c
Pass ListOptions to List() methods.
2015-12-01 15:00:36 +01:00
