Skip to content
This repository was archived by the owner on May 6, 2020. It is now read-only.

One proxy per vm#167

Merged
sboeuf merged 6 commits into
clearcontainers:masterfrom
jodh-intel:one-proxy-per-vm
Nov 13, 2017
Merged

One proxy per vm#167
sboeuf merged 6 commits into
clearcontainers:masterfrom
jodh-intel:one-proxy-per-vm

Conversation

@jodh-intel
Copy link
Copy Markdown

Allow the proxy to be run for a single pod.

This branch allows multiple proxies to be run on a single system, one per pod. Note that it also permits this behaviour even when the proxy is run as a system-level systemd service.

James O. D. Hunt added 6 commits November 13, 2017 15:25
Correct an error message containing a typographical mistake.

Signed-off-by: James O. D. Hunt <james.o.hunt@intel.com>
getSocketPath() now accepts a URI (specified by a new "-uri" CLI option)
that specifies the socket to use. If the specified socket path is empty,
the old behaviour is applied.

Signed-off-by: James O. D. Hunt <james.o.hunt@intel.com>
Contributions-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
Add the syslog logrus package.

Signed-off-by: James O. D. Hunt <james.o.hunt@intel.com>
Write log output to syslog rather than stderr when running as
a pod instance.

Signed-off-by: James O. D. Hunt <james.o.hunt@intel.com>
Don't run KSM code when running a pod-specific instance.

Signed-off-by: James O. D. Hunt <james.o.hunt@intel.com>
Signal the proxy to shutdown when unregisterVM() is called if running
as a pod instance. Note that the shutdown cannot be immediate: it must
first send back the response frame and wait for the hypervisor to stop.

Fixes clearcontainers#164.

Signed-off-by: James O. D. Hunt <james.o.hunt@intel.com>
@jodh-intel
Copy link
Copy Markdown
Author

Tested locally to ensure it is possible to run a system-level cc-proxy instance plus pod-level instances.

@sboeuf
Copy link
Copy Markdown

sboeuf commented Nov 13, 2017

@jodh-intel oh ok so you mean with this patch, we can still use the proxy as a single instance ?

@jodh-intel
Copy link
Copy Markdown
Author

Hi @sboeuf - yes, we can retain the systemd-spawned instance for the time being, then remove that functionality later. This multi-step process allows us to keep disruption to a minimum. See clearcontainers/runtime#795.

@sboeuf
Copy link
Copy Markdown

sboeuf commented Nov 13, 2017

@jodh-intel that's perfect to go step by step, I agree we don't want to break everyhting :)

@clearcontainersbot
Copy link
Copy Markdown

kubernetes qa-passed 👍

Copy link
Copy Markdown

@sboeuf sboeuf left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Perfect PR :)

@sboeuf
Copy link
Copy Markdown

sboeuf commented Nov 13, 2017

LGTM

@sboeuf
Copy link
Copy Markdown

sboeuf commented Nov 13, 2017

@amshinde please take a look and merge this so that we can move forward with containers/virtcontainers#483

Copy link
Copy Markdown
Contributor

@amshinde amshinde left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants