Commit Graph

216 Commits

Author SHA1 Message Date
Zach Loafman
616c6be653 Refactor the "gcloud compute instances create" call as well. 2015-04-06 08:35:02 -07:00
Zach Loafman
9e5fd874cc First hack at upgrade script for GCE:
Address #6075: Shoot the master VM while saving the master-pd. This
takes a couple of minor changes to configure-vm.sh, some of which also
would be necessary for reboot. In particular, I changed it so that the
kube-token instance metadata is no longer required after inception;
instead, we mount the master-pd and see if we've already created the
known tokens file before blocking on the instance metadata.

Also partially addresses #6099 in bash by refactoring the kube-push
path.
2015-04-05 11:42:34 -07:00
Filip Grzadkowski
0ffabfa9f7 Add ability to set master/minion size for GCE using envvar. 2015-04-03 11:36:05 +02:00
Eric Tune
59daeabaee Make secrets at cluster startup.
These secrets will be used in subsequent PRs by:
scheduler, controller-manager, monitoring services,
logging services, and skydns.

Each of these services will then be able to stop using kubernetes-ro
or host networking.
2015-04-02 15:58:45 -07:00
CJ Cullen
5e6e67ba59 Add an alternative TokenSource to the GCE CloudProvider. 2015-04-01 17:52:30 -07:00
Jeff Lowdermilk
ce72839153 Merge pull request #6329 from mbforbes/fixUp
Fix pod that's looked for on kube up.
2015-04-01 17:51:37 -07:00
Alex Robinson
4f9b1c486f Don't ever give up on retrying downloads in the GCE download-or-bust function. 2015-04-02 00:40:19 +00:00
Max Forbes
152a461f39 Fix pod that's looked for on kube up. 2015-04-01 17:28:04 -07:00
Jeff Grafton
0d382c03fc Add missing } to fix cluster/gce/util.sh 2015-03-31 14:03:18 -07:00
Robert Bailey
fe00f7abd0 Merge pull request #6168 from brendandburns/docs
Auto-install gcloud and needed components, if necessary
2015-03-31 13:55:46 -07:00
Zach Loafman
6c219885e0 s/apiservers/api_servers/ in Salt
It looks like api_servers finally won this battle. Kill off the
last remaining places passing it, but allow the kubelet Salt to
accept apiservers for a period of time.

(This was bothering my OCD.)
2015-03-31 12:29:46 -07:00
Brendan Burns
98cdf04189 Auto-install gcloud if needed. 2015-03-31 12:10:35 -07:00
Victor Marmol
0c75f26a74 Merge pull request #6148 from zmerlynn/brctl_complaint
Fix "brctl: invalid command" message appearing in startup logs
2015-03-30 08:01:30 -07:00
Zach Loafman
b3636bba34 Fix "brctl: invalid command" message appearing in startup logs
We were actually failing to call brctl in configure-vm.sh. I finally
tracked it down to the attempt to delete the docker0 bridge. This
particular package was getting installed later by Salt anyways, so
all this PR is doing is moving the package install up from Salt to
bash.

Also adds some minor logging.
2015-03-29 18:30:33 -07:00
Zach Loafman
708553f0bb GCE node salt: Use the master FQDN, not the master IP
This will allow us to upgrade the master by full re-provision, without
even bothering to reserve the IP.
2015-03-28 13:44:45 -07:00
Daniel Smith
525bbfd175 Merge pull request #6103 from zmerlynn/remove_gce_node_names
Remove the --machines SaltStack configuration on GCE
2015-03-27 17:55:16 -07:00
Zach Loafman
68ccb97907 Remove the --machines SaltStack configuration on GCE
Per https://github.com/GoogleCloudPlatform/kubernetes/issues/6072#issuecomment-87074456, this is no longer necessary.
We now no longer need a static node list. Woo!
2015-03-27 14:44:19 -07:00
Brendan Burns
6fd376a04f Fix the scripts to optionally skip prompts and also return successfully no matter what. 2015-03-27 13:53:26 -07:00
Brendan Burns
1aa2b97792 Revert "Revert "Update gce starting guide to use get.k8s.io""
This reverts commit b369f2b48f.
2015-03-27 13:46:28 -07:00
Zach Loafman
b369f2b48f Revert "Update gce starting guide to use get.k8s.io" 2015-03-26 19:36:25 -07:00
Brendan Burns
61b624457d Update gce starting guide to use get.k8s.io
Also auto install gcloud if it's not present.
2015-03-26 16:08:56 -07:00
Max Forbes
ddb0a70481 Fix tiny salt logging line. 2015-03-26 14:37:51 -07:00
Brendan Burns
eea09ddfbb Fix the date command to work on OS X 2015-03-26 13:48:59 -07:00
Zach Loafman
ed567736ea Retry apt operations to death
Nodes are probably broken if update or install fails. Don't proceed
if we can't get past these. Also, instead of ignoring the error off
dpkg, use --force depends, which changes the errors to be kinder
warnings for anyone looking through the logs.
2015-03-24 16:11:40 -07:00
derekwaynecarr
ee53dfc741 Turn on namespace lifecycle plug-in 2015-03-24 10:36:06 -04:00
Robert Bailey
c3189657b6 Bump the default boot disk size for GCE from 10GB to 100GB so that there is
more space to store containers and image layers.
2015-03-20 13:06:35 -07:00
Zach Loafman
407d1fec45 Both @satnam6502 and E2E tests confirm: This code is no longer needed
Deletion is wonderful. The only weird thing was where to put the
message about the proxy URLs. Satnam suggested kubectl clusterinfo,
which seemed like a good option to put at the end of cluster turn-up.
2015-03-19 22:23:11 -07:00
Zach Loafman
57cd8165b0 Ensure salt-minion never starts by setting invoke-rc.d policy using
the /usr/sbin/policy-rc.d script and returning 101, per
https://people.debian.org/~hmh/invokerc.d-policyrc.d-specification.txt,
but only for the window while we're installing Salt.

This is a much more fool-proof method than what I was attempting
before. I hunted for how to do this before and clearly failed at my
Google-fu.

Fixes #5621
2015-03-18 16:15:24 -07:00
Vish Kannan
b6e7f470cb Merge pull request #5576 from vmarmol/auth
Add monitoring scope to default GCE config.
2015-03-17 17:22:52 -07:00
Victor Marmol
8e2dc04da8 Add monitoring scope to default GCE config.
This will allow Heapster to run GCM-based monitoring on the nodes.
2015-03-17 17:02:10 -07:00
Dawn Chen
dce75998d0 Upgrade to container-vm-20150315 2015-03-17 16:10:39 -07:00
Zach Loafman
f2de7b4dcc Add --can-ip-forward to master instance
Another piece missing in
https://github.com/GoogleCloudPlatform/kubernetes/pull/5390. The
master should have --can-ip-forward if you're routing to it.
2015-03-13 15:17:07 -07:00
Filipe Brandenburger
0948cb745f Merge pull request #5454 from jlowdermilk/get-password
Make get-password robust against invalid kubeconfig entries
2015-03-13 13:40:55 -07:00
Jeff Lowdermilk
8fef6fb343 Make get-password robust against invalid kubeconfig entries 2015-03-13 13:26:47 -07:00
Jeff Lowdermilk
9b55e1f176 Delete master route on kube-down 2015-03-13 11:00:19 -07:00
Satnam Singh
dfb9f2aa17 Revert "Revert "Actually update binaries during kube-push"" 2015-03-12 18:04:36 -07:00
Satnam Singh
bb2c3037e0 Revert "Actually update binaries during kube-push" 2015-03-12 17:09:50 -07:00
Satnam Singh
c3951d7cf6 Merge pull request #5415 from zmerlynn/fix_binary_push
Actually update binaries during kube-push
2015-03-12 16:51:28 -07:00
Zach Loafman
b02188b7d8 Actually update binaries during kube-push
This was a dumb mistake during a re-factor of configure-vm. I tested
this early, re-factored the tail of this file, spot checked kube-push
and failed to test kube-push properly. My bad.

Fixes #5361. Fixes #5408.
2015-03-12 16:44:32 -07:00
Brendan Burns
5d9fe6212c Add MASTER_IP_RANGE to e2e. 2015-03-12 14:22:03 -07:00
Zach Loafman
9fc5262c14 Infer KUBERNETES_MASTER from hostname rather than requiring it to be explicit
This allows GKE to have the same YAML on both the master and nodes
2015-03-12 11:51:21 -07:00
Satnam Singh
e16bcceceb Merge pull request #5390 from brendandburns/shell_sucks
Configure docker on the master like we do on workers.
2015-03-12 11:42:03 -07:00
Brendan Burns
17ff8fb421 Configure docker on the master like we do on workers. 2015-03-12 10:37:30 -07:00
Alex Robinson
bc655681fe Merge pull request #5352 from zmerlynn/conditional_disk
Don't always run mount-master-pd
2015-03-12 10:19:04 -07:00
Victor Marmol
c03b080328 Merge pull request #5332 from vishh/heapster_e2e
Adding a ginkgo version of monitoring e2e test
2015-03-12 08:37:45 -07:00
Zach Loafman
d4f06cea9f Don't always run mount-master-pd 2015-03-11 19:58:44 -07:00
Vishnu Kannan
e7aca5c337 Enable heapster based cluster monitoring by default on other cluster providers, expect for GKE. 2015-03-11 23:34:04 +00:00
Vishnu Kannan
663bdb91c9 Remove monitoring firewall setup methods. Heapster anf Grafana are now accessible via the API server proxy. 2015-03-11 23:20:45 +00:00
Vishnu Kannan
425dd7e3ee Adding a heapster ginkgo e2e test. 2015-03-11 21:47:47 +00:00
Zach Loafman
e780592315 Merge pull request #5264 from jlowdermilk/kube-up
Standalone kubeconfig for gce kube-up (mulligan)
2015-03-11 11:46:45 -07:00