Eric Paris
c560f41a4b
Update docs which were incorrectly using _ in flag names
...
Each of these was verified that the actual flag declaration correctly
used -
2015-08-17 11:18:43 -07:00
Eric Paris
86ca88be61
convert cluster/ from --flag_name= --flag-name=
2015-08-07 11:00:13 -04:00
Robert Bailey
8df33bc1a7
Register the kubelet on the master node with an apiserver. This option is
...
separated from the apiserver running locally on the master node so that it
can be optionally enabled or disabled as needed.
Also, fix the healthchecking configuration for the master components, which
was previously only working by coincidence:
If a kubelet doesn't register with a master, it never bothers to figure out
what its local address is. In which case it ends up constructing a URL like
http://:8080/healthz for the http probe. This happens to work on the master
because all of the pods are using host networking and explicitly binding to
127.0.0.1. Once the kubelet is registered with the master and it determines
the local node address, it tries to healthcheck on an address where the pod
isn't listening and the kubelet periodically restarts each master component
when the liveness probe fails.
2015-08-06 13:39:32 -07:00
derekwaynecarr
0de0e3e9f1
Fix salt configuration on systemd setups
2015-08-05 17:01:43 -04:00
gmarek
5bb7eec5ab
Use salt pillars to pass test arguments during cluster creation.
2015-07-31 09:32:01 +02:00
Alex Robinson
94ae0a9a0d
Make the kubelet on a GCE master check instance metadata for manifests.
...
Primary motivation: enable GKE and other cluster-as-a-service folks to
easily run additional logic on the master without having to modify salt
or SSH to the master after it's been created.
2015-07-30 21:04:34 +00:00
Justin Santa Barbara
89900fa19b
AWS: Ubuntu Vivid, set cgroup_root to docker
...
Otherwise libcontainer generates systemd unit file names that start with "-",
which are not valid.
2015-07-08 17:00:43 -04:00
Justin Santa Barbara
a0ae1d90ba
Salt: don't assume that master => cbr-cidr
...
In particular, this is required for vagrant
2015-06-25 09:17:00 -04:00
Maxwell Forbes
3afda5d566
Merge pull request #10312 from dchen1107/cleanup
...
Take 2: Fix the race between configuring cbr0 and restarting static pods
2015-06-24 17:59:50 -07:00
Dawn Chen
6ddfa512de
Revert "Revert "Fix the race between configuring cbr0 and restarting static pods""
...
This reverts commit fd0a95dd12
.
2015-06-24 11:10:10 -07:00
Piotr Szczesniak
fd0a95dd12
Revert "Fix the race between configuring cbr0 and restarting static pods"
2015-06-24 09:56:49 +02:00
derekwaynecarr
db202d4904
Remove nginx from vagrant
2015-06-23 13:07:50 -04:00
Brendan Burns
192ffdfb25
Fix the container bridge so that it can create cbr0
...
Fix the kubelet so that it tries to sync status, even if Docker is down
2015-06-22 23:18:01 -07:00
Justin Santa Barbara
16d6ba8a9d
AWS: Stop the master kubelet from registering as a node (like GCE does)
2015-06-12 18:33:45 -04:00
derekwaynecarr
2168cee414
Upgrade to Fedora 21, Docker 1.6, clean-up SDN
2015-06-04 10:59:23 -04:00
Robert Bailey
1dad552edc
Prevent access to the /run and /exec endpoints on the master's kubelet
...
so that users can't run arbitrary code on the master.
2015-06-02 14:42:11 -07:00
Justin Santa Barbara
c4a2631593
Mount logic breaks if /var/lib/kubelet is a symlink
...
Pass the correct kubelet root-dir on AWS
2015-05-29 20:13:09 -04:00
Victor Marmol
c97dda068d
Create a /system system container in Debian.
2015-05-26 20:52:09 -07:00
Victor Marmol
656be14e84
Run Docker containers as top-level containers in Debian.
...
Part of #8215
2015-05-22 13:27:25 -07:00
CJ Cullen
bf646abf8d
Revert "Revert "Modify nodes to register directly with the master.""
...
This reverts commit c53786ab31
.
2015-05-20 14:21:03 -07:00
Quinton Hoole
c53786ab31
Revert "Modify nodes to register directly with the master."
2015-05-20 13:47:51 -07:00
Robert Bailey
8e356f8439
Configure the cloud provider for the kubelet.
2015-05-19 09:55:07 -07:00
Robert Bailey
01467e0bb8
Modify nodes to register directly with the master.
...
- Delete nodes when they are no longer ready and don't exist in the
cloud provider.
- Label each node with it's hostname.
- Add flag to skip node registration.
- Add a test for registering an existing node.
2015-05-19 09:55:07 -07:00
CJ Cullen
31ea7d1295
Put cbr0-modifying stuff behind a flag. Address some other comments.
2015-05-12 23:00:28 -07:00
Justin Santa Barbara
57f7b658bb
AWS: Change apiserver to listen on 443 directly, not through nginx
...
Mirrors changes in GCE. I think the same changes will be needed for vagrant.
2015-05-02 16:59:04 -04:00
Robert Bailey
dc45f7f9e6
Remove nginx and replace basic auth with bearer token auth for GCE.
...
- Configure the apiserver to listen securely on 443 instead of 6443.
- Configure the kubelet to connect to 443 instead of 6443.
- Update documentation to refer to bearer tokens instead of basic auth.
2015-04-22 11:11:20 -07:00
Justin Santa Barbara
29592356f7
Rename salt minion_ip to hostname_override
...
Prep work for making AWS node identifier -> instance id
2015-04-21 06:24:49 -07:00
Justin Santa Barbara
2426366ec8
Update cAdvisor with moved docker root on AWS
...
We set up a symlink now, and we also pass docker_root into the kubelet.
The symlink is probably sufficient, but doing both feels safer.
2015-04-07 10:40:22 -07:00
Robert Bailey
40ebed845f
Remove salt configuration that matches the default settings in
...
the Kubelet.
2015-04-01 12:13:40 -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
Abhishek Shah
d84ff79cfd
Added a catchall to set api server ip address in kubelet cmd params
2015-03-27 08:44:52 -07:00
Abhishek Shah
f35a352782
Modified salt config to get master ip address correctly
2015-03-26 21:17:06 -07:00
Eric Tune
59a06b5073
Don't set --etcd_servers argument for kubelet.
...
This is ignored by all recent kubelet versions when api_servers is
set. And it is set in this salt file.
2015-02-11 15:53:26 -08:00
Justin Santa Barbara
7790435915
Fix for bool vs string comparisons in salt templates
...
Fixes #4118
2015-02-04 14:09:10 -05:00
Joe Beda
228659a177
Fix up for posix flags
2015-01-15 09:12:19 -08:00
Tim Hockin
59164ca844
Make DNS an option for cluster turnup
2014-12-29 09:18:12 -08:00
Brendan Burns
878d85df0c
Add a log level flag.
2014-12-15 12:00:29 -08:00
Brendan Burns
149881e13e
Actually set repository QPS in the kubelet.
2014-12-11 12:06:40 -08:00
Alen Komljen
d83d269c27
Remove empty lines when file is generated from jinja template
2014-12-02 15:00:58 +01:00
Eric Tune
46dcacfa93
Kubelet talks securely to apiserver.
...
Configure apiserver to serve Securely on port 6443.
Generate token for kubelets during master VM startup.
Put token into file apiserver can get and another file the kubelets can get.
Added e2e test.
2014-11-18 00:23:41 -08:00
Tim Hockin
440d14c239
Use '=' for flags or else parsing gets confused sometimes
2014-10-06 12:25:56 -07:00
Brendan Burns
bb829ce8f6
fix salt.
2014-09-17 17:13:41 -07:00
Brendan Burns
ffa4500e09
Add a salt flag for enabling privileged containers.
2014-09-17 14:25:28 -07:00
derekwaynecarr
c8c5435656
Enforce --address flag is set on all Linux flavors
2014-08-07 17:54:49 -05:00
Tim Hockin
7e9c142392
Remove duplicate -config param
...
This is duplicated in cluster/saltbase/salt/kubelet/default
cluster/saltbase/salt/kubelet/initd
2014-08-07 10:53:08 -07:00
Tim Hockin
b0ffa8c8f6
Set kubelet -address flag to 0.0.0.0 via salt
...
Currently it binds to the eth0 IP only, this includes localhost. The flag
default is localhost only, but we need it looser than that.
2014-08-06 18:11:35 -07:00
derekwaynecarr
69ae2fe4bb
Initial vagrant setup and e2e testing support
2014-07-24 16:32:36 -04:00
Victor Marmol
842907613b
Run cAdvisor on all minions.
2014-07-01 11:36:33 -07:00
Joe Beda
894a7e3282
Move everything out of src and reorganize scripts.
...
Fixed up some scripts to be more robust. Changed the e2e test setup to use g1-small instances. Fixed up documentation to reflect the new script locations. Disabled the "curl | bash" cluster launch as it hasn't been well tested and doesn't include the cloudcfg tool yet.
2014-06-06 21:41:19 -07:00