Join GitHub today
GitHub is home to over 40 million developers working together to host and review code, manage projects, and build software together.
Sign up[WIP] Virt apiserver #355
[WIP] Virt apiserver #355
Conversation
stu-gott
force-pushed the
stu-gott:virt-apiserver
branch
2 times, most recently
from
70a697f
to
7d9b522
This comment has been minimized.
This comment has been minimized.
|
Pfeh - This looks good so far. |
This comment has been minimized.
This comment has been minimized.
mpolednik
commented
Sep 6, 2017
|
I believe this feature lacks a proposal. Could you create a doc explaining the scope, which services are affected and how new services would be developed when using the aggregated apiserver? The doc being part of this PR is fine. |
This comment has been minimized.
This comment has been minimized.
|
@mpolednik You are correct! I accidentally created this PR a while back, and it's very much a work in progress currently. I'll definitely add a proposal/explanation for these changes before removing the "[WIP]" |
| volumeMounts: | ||
| - mountPath: /var/run/virt-apiserver-cert | ||
| name: virt-apiserver-cert | ||
| - name: etcd |
This comment has been minimized.
This comment has been minimized.
jhernand
Oct 2, 2017
What is the reason for having a separate instance of etcd can't the API server use the same instance than the rest of the cluster? If the administrator of the cluster configures it for high availability and replication of the master, then this separate instance of etcd won't benefit from that, if I understand correctly.
Signed-off-by: Stu Gott <sgott@redhat.com>
Signed-off-by: Stu Gott <sgott@redhat.com>
Signed-off-by: Stu Gott <sgott@redhat.com>
Signed-off-by: Stu Gott <sgott@redhat.com>
Signed-off-by: Stu Gott <sgott@redhat.com>
Signed-off-by: Stu Gott <sgott@redhat.com>
Signed-off-by: Stu Gott <sgott@redhat.com>
Signed-off-by: Stu Gott <sgott@redhat.com>
Signed-off-by: Stu Gott <sgott@redhat.com>
This comment has been minimized.
This comment has been minimized.
|
Great work, due to the missing in-cluster storage for user API servers we decided to stick with CCRDs for the time being, thus this user API server can be closed. Once Kubernetes gains in-cluster user API server storage, then we might want to revisit this topic again. Thanks Stu. |
stu-gott commentedAug 3, 2017
•
edited