Commit Graph

374 Commits

Author SHA1 Message Date
k8s-merge-robot
58783bc256 Merge pull request #21684 from justinsb/fix_21200
Auto commit by PR queue bot
2016-02-23 07:06:08 -08:00
Justin Santa Barbara
050863cf93 AWS kube-up: Build kubeconfig asap
Once we've built the master, we can build kubeconfig.  By doing so, if
we time out waiting for the nodes, the system is still configured
correctly.

In particular, spot instances can be slow to launch.

Related to issue #21200
2016-02-22 10:46:56 -05:00
Justin Santa Barbara
fb12c48cb7 AWS kube-up: Allow spot instances
I think we should probably leave this undocumented for now, until we
have a better way to launch multiple sets of nodes, but it's great for
cost savings while testing!

Fix #21200
2016-02-22 10:42:49 -05:00
k8s-merge-robot
1ce188e557 Merge pull request #21430 from spxtr/e2e-internal
Auto commit by PR queue bot
2016-02-21 12:08:22 -08:00
k8s-merge-robot
d6192dd152 Merge pull request #21342 from spxtr/remove-e2e-test
Auto commit by PR queue bot
2016-02-21 00:39:38 -08:00
k8s-merge-robot
a12a2e984a Merge pull request #21233 from justinsb/aws_wily
Auto commit by PR queue bot
2016-02-20 13:28:08 -08:00
k8s-merge-robot
fe18055adc Merge pull request #21165 from justinsb/aws_master_pd_in_fstab
Auto commit by PR queue bot
2016-02-20 00:29:46 -08:00
k8s-merge-robot
6c5bf89ed3 Merge pull request #21160 from justinsb/aws_bigger_master
Auto commit by PR queue bot
2016-02-20 00:29:45 -08:00
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
k8s-merge-robot
bcabc096f2 Merge pull request #21130 from madhusudancs/daemonset-enable-default
Auto commit by PR queue bot
2016-02-17 22:54:35 -08:00
k8s-merge-robot
78c294a66c Merge pull request #21164 from justinsb/aws_reboot_with_ephemeral
Auto commit by PR queue bot
2016-02-17 17:12:01 -08:00
Joe Finney
2b756ca56d Move E2E kube-up call into test-setup from e2e-internal. 2016-02-17 16:49:07 -08:00
laushinka
7ef585be22 Spelling fixes inspired by github.com/client9/misspell 2016-02-18 06:58:05 +07:00
Madhusudan.C.S
f06a195855 Remove all {KUBE_,}ENABLE_{DEPLOYMENT,DAEMONSET} variables from all our bash scripts. 2016-02-17 15:04:28 -08:00
Madhusudan.C.S
ab1e9c200f Graduate DaemonSet to v1beta1 API, enable it by default. 2016-02-17 15:03:12 -08:00
Justin Santa Barbara
4d2c580781 AWS kube-up: Add support for wily
vivid is technically no longer supported now, so we should support
wily.
2016-02-17 10:45:21 -05:00
Justin Santa Barbara
a16372ce40 AWS: Change master to m3.medium by default
m3.large for > 150 nodes.

t2.micro often runs out of memory.  The t2 class has very
difficult-to-understand behaviour when it runs out of CPU.  The
m3.medium is reasonably affordable, and avoids these problems.

Fix #21151

Issue #18975
2016-02-17 10:24:14 -05:00
Justin Santa Barbara
c68c019150 AWS: Add comments to explain the symlink logic 2016-02-17 09:57:37 -05:00
Justin Santa Barbara
affffdc9ab AWS kube-up: Put the master-pd into /etc/fstab, to mount it early
Otherwise we risk services coming up on the master before the backing
volume is ready.

If we then see the master-pd is already mounted, don't try to remount
it.

Issue #21155
2016-02-17 09:47:01 -05:00
k8s-merge-robot
22ad56dded Merge pull request #21161 from justinsb/aws_bootstrap_on_every_boot
Auto commit by PR queue bot
2016-02-17 06:45:24 -08:00
Joe Finney
6f7182e500 Remove hack/e2e-test.sh in favor of hack/e2e.go. 2016-02-16 14:54:50 -08:00
k8s-merge-robot
784f555fd9 Merge pull request #18821 from rastut/add_aws_detect_root_device
Auto commit by PR queue bot
2016-02-13 10:09:20 -08:00
k8s-merge-robot
f2e845dd6d Merge pull request #21144 from justinsb/aws_fix_project_not_defined
Auto commit by PR queue bot
2016-02-13 08:33:53 -08:00
Justin Santa Barbara
69300c51b3 AWS kube-up: Reboot support with ephemeral storage
If the ephemeral volume is present and mounted, don't try to reinitialize
them.

Don't block the boot if the ephemeral volume is corrupt / missing -
this enables us to cope with a stop/start & presumably also corruption.
In this case, we'll reformat the ephemeral storage.

Fix #21157
2016-02-12 20:31:56 -05:00
Justin Santa Barbara
19cf622e88 AWS kube-up: make the bootstrap script run every boot
This is so we have the same behaviour as on GCE.

This also lets us change the bootstrap script or the config, which is
nice.  Instance data is immutable on AWS once it is booted.

Fix #21150
2016-02-12 20:30:47 -05:00
Janet Kuo
2874f30c05 Enable Deployments by default 2016-02-12 12:29:14 -08:00
Justin Santa Barbara
1852fab029 AWS: Fix kube-up generation of kubeconfig
We were assuming the PROJECT env var was set, which the e2e tests do.
But PROJECT is normally not set on AWS (it is set on GCE); this broke as
part of the harmonization.

Revert to the pre-existing behaviour here, where we use "aws_" as the
prefix.

Fix #21141
2016-02-12 09:53:22 -05:00
Chris Kühl
7fcc40bb27 docs: replace Rocket with rkt 2016-02-09 13:52:25 +01:00
k8s-merge-robot
68a165cda9 Merge pull request #20681 from justinsb/aws_sync_kubeup_gce_step3
Auto commit by PR queue bot
2016-02-09 02:08:48 -08:00
Michael Nikitochkin
55e2906f79 AWS: Allow to configure master ip and services mask
Fix script for case when neeed to setup cluster
in an existen VPC and subnet with ip mask example: 10.0.0.0/8.

Fixed bug to detect ip of master if provided MASTER_RESERVED_IP.

For some reason detecting master ip was moved to volumes and only when MASTER_RESERVED_IP=auto.
If specify IPv4 for MASTER_RESERVED_IP like `52.1.1.1`, than we could
not detect ip even during last steps of setuping cluster.
step the KUBE_MASTER_IP is reseted because there are no tag for the
volume.
2016-02-07 21:59:00 +01:00
Justin Santa Barbara
63743e9c1c AWS kube-up: Use the cert we generated in kubeconfig 2016-02-06 22:52:01 -05:00
Justin Santa Barbara
270fae4d66 kube-up: Support multiple additional IPs for SSL cert
And pass the public IP when setting up AWS
2016-02-06 22:52:01 -05:00
Justin Santa Barbara
fd5c44c72e kube-up: Finish refactor of AWS to reuse GCE script
We override key functions in the GCE script with AWS equivalents.
2016-02-06 22:52:01 -05:00
k8s-merge-robot
9d981b696c Merge pull request #20642 from justinsb/aws_sync_kubeup_gce_step2
Auto commit by PR queue bot
2016-02-06 11:56:47 -08:00
k8s-merge-robot
5b046cccdb Merge pull request #20448 from justinsb/aws_fix_e2e_detect_master
Auto commit by PR queue bot
2016-02-04 16:24:20 -08:00
Justin Santa Barbara
732e76fb4e kube-up: Refactor SSL cert helpers into common
Next step in the the GCE/AWS harmonization
2016-02-04 09:10:41 -05:00
Justin Santa Barbara
0f6b3611d8 AWS: Fix e2e usage of find-tagged-master-ip
In the e2e tests detect-master is called directly.  In turn, it calls
find-tagged-master-ip, which assumed that find-master-pd has already already
been called.  But this wasn't true in the e2e case.

We add a call to find-master-pd; it is idempotent.
2016-02-01 22:49:31 +00:00
k8s-merge-robot
30375b21d6 Merge pull request #19776 from justinsb/aws_sync_kubeup_gce
Auto commit by PR queue bot
2016-02-01 14:44:18 -08:00
k8s-merge-robot
0aca70016c Merge pull request #20066 from justinsb/aws_move_build_runtime_config
Auto commit by PR queue bot
2016-02-01 09:48:27 -08:00
Alex Robinson
f684e80b1f Merge pull request #20067 from justinsb/aws_test_defaults
AWS: Change config-test defaults to match GCE
2016-01-30 17:42:28 -08:00
Justin Santa Barbara
c556e8f0f5 kube-up: Refactor kube-env into common; use in AWS
Use kube-env on the AWS master also
2016-01-28 10:31:51 -08:00
Justin Santa Barbara
4d88a4ed2c kube-up: move sha1sum-file into common
Also hash the tar files in AWS kube-up
2016-01-28 10:19:38 -08:00
Justin Santa Barbara
a70412099a AWS: add retries around apt-get install
Related to issue #17309
2016-01-28 10:19:38 -08:00
Justin Santa Barbara
ae75a7d259 kube-up: move build-runtime-config from AWS & GCE into common 2016-01-28 10:19:38 -08:00
Justin Santa Barbara
0df6f5d8c5 kube-up: move yaml-quote from GCE & AWS to common 2016-01-28 10:19:38 -08:00
k8s-merge-robot
6c63acdd3d Merge pull request #19398 from justinsb/aws_elasticip_for_master
Auto commit by PR queue bot
2016-01-26 04:34:28 -08:00
k8s-merge-robot
8c821700f3 Merge pull request #19389 from justinsb/internal_cidr
Auto commit by PR queue bot
2016-01-26 03:58:48 -08:00
Justin Santa Barbara
a4d6801655 AWS: Change config-test defaults to match GCE
KUBE_ENABLE_DAEMONSETS & KUBE_ENABLE_DEPLOYMENTS default to true in
config-test in GCE, so we should do the same in AWS for e2e sanity.
2016-01-24 10:43:43 -05:00
Justin Santa Barbara
8a7afc459c AWS: Call build-runtime-config where it is used
build-runtime-config was being called in verify-prereqs, which didn't
match how GCE called it, and didn't seem to actually work.

Instead call it just before the master configuration is built.  Also
call it just before the node configuration is built, even though the
nodes don't _currently_ require the runtime_config.
2016-01-24 10:43:15 -05:00
Justin Santa Barbara
d8460d5920 AWS: Pass non-masquerade-cidr through kube-up
Set the environment variable NON_MASQUERADE_CIDR.

Docs in cluster/aws/options.md
2016-01-22 21:01:44 -05:00