Add some clarity around the state.json file #211
Conversation
runtime.md
Outdated
There was a problem hiding this comment.
Maybe we can say Linux/Unix? Thoughts?
There was a problem hiding this comment.
On Fri, Oct 02, 2015 at 01:40:43PM -0700, Vincent Batts wrote:
+On Linux based systems the metadata MUST be stored in
/run/opencontainer/containers.Maybe we can say Linux/Unix? Thoughts?
“Systems with POSIX paths”? I'm fine leaving this as “Linux” and
opting in other architectures as we add support for them.
Why don't we require a particular path on Windows?
There was a problem hiding this comment.
used Linux/Unix.
If someone from MSFT tells me what dir to use I can add it.
There was a problem hiding this comment.
that's where we look forward to the PoC implementations to help define that.
On Fri, Oct 2, 2015 at 4:59 PM, Doug Davis [email protected] wrote:
In runtime.md
#211 (comment):@@ -2,22 +2,26 @@
State
-The runtime state for a container is persisted on disk so that external tools can consume and act on this information.
-The runtime state is stored in a JSON encoded file.
-It is recommended that this file is stored in a temporary filesystem so that it can be removed on a system reboot.
-On Linux based systems the state information should be stored in/run/opencontainer/containers.
-The directory structure for a container is/run/opencontainer/containers/<containerID>/state.json.-By providing a default location that container state is stored external applications can find all containers running on a system.
-*
version(string) Version of the OCI specification used when creating the container.
-*id(string) ID is the container's ID.
-*pid(int) Pid is the ID of the main process within the container.-*
bundlePath(string) BundlePath is the path to the container's bundle directory.+Metadata for containers are stored on disk so that external tools can consume and act on this information.
+It is recommended that this data be stored in a temporary filesystem so that it can be removed on a system reboot.
+On Linux based systems the metadata MUST be stored in/run/opencontainer/containers.used Linux/Unix.
If someone from MSFT tells me what dir to use I can add it.—
Reply to this email directly or view it on GitHub
https://github.com/opencontainers/specs/pull/211/files#r41067051.
runtime.md
Outdated
There was a problem hiding this comment.
The README for this project suggests (to me) either “OCS” or “OCIS” or “OCI Specification”. I suggest we pick one abbreviation and have the README say:
The Open Container Initiative Specifications (OCIS) …
or whatever. And then use that short name throughout the spec. Until we do that, I'd probably just leave this as “OCI specification”.
There was a problem hiding this comment.
hmmm. OCF had been mentioned, but not much defined.
Perhaps just "OpenContainers spec" or similar for now.
|
I like this PR a lot, but it's going to conflict with #145, and I
think those (minor) changes are useful. I'm fine if you just squash
them into this branch, or if I rebase mine on this, or if you rebase
yours on mine, …. If I can help manage that conflict in any way, let
me know ;).
|
duglin
force-pushed
the
tweakRuntime
branch
2 times, most recently
from
4665ce0 to
6a3e10b
Compare
|
@wking I think I got your edits from the other one |
runtime.md
Outdated
There was a problem hiding this comment.
I'd change to “is the ID of the main process” to avoid doubling up on “process”.
|
I agree that we need to tighten up the format of the ID. I was going to leave that for a different PR so that this is more about just cleaning up the text, but if we can agree quickly (w/o much bike shedding) then I'll add it. To get the ball rolling, what about IDs must conform to: |
runtime.md
Outdated
There was a problem hiding this comment.
Maybe qualify this with “Linux/Unix” somehow? Something like “For example, a Linux/Unix directory /run/opencontainer/containers/173975398351 corresponds to a container with an ID of 173975398351.”
duglin
force-pushed
the
tweakRuntime
branch
2 times, most recently
from
ee6dc1c to
7d2f0cc
Compare
|
On Fri, Oct 02, 2015 at 02:27:12PM -0700, Doug Davis wrote:
Fair. So far, it's just been silence, so I don't think bikeshedding
I think you want ‘-’ too, so you can use UUIDs (links in my thread). |
While there may be the need for rigidity, I'm not sure that needs to be the case right now. UUID needs to be allowed. But possibly colons too. |
|
Let's save the ID format for another PR. |
|
sounds good On Mon, Oct 5, 2015 at 10:17 AM, Doug Davis [email protected]
|
runtime.md
Outdated
|
LGTM |
Signed-off-by: Doug Davis <[email protected]>
|
LGTM |
Add some clarity around the state.json file
We already require it for Linux/Unix-based systems [1], so we don't have to repeat it here. And other systems will use different paths, which we haven't specified yet. When I asked why we didn't specify a path for Windows [2], Vincent said we were waiting on help from PoC implementations [3]. So this commit punts the location to the "State" section, and lets the "Lifecycle" section just focus on when the write-to-filesystem happens. There's also discussion about removing the filesystem state registry commpletely [4], in which case we'd want to remove the whole line from the lifecycle. [1]: opencontainers@7713efc#diff-b84a8d65d8ed53f4794cd2db7e8ea731L7 [2]: opencontainers#211 (comment) [3]: opencontainers#211 (comment) [4]: https://groups.google.com/a/opencontainers.org/forum/#!topic/dev/q6TYqVZOcX8 Subject: removal of /run/opencontainer/containers Date: Wed, 25 Nov 2015 14:29:35 +0000 Message-ID: <CAD2oYtNipt3i_C6=J4Bc-jwauo5YAvKXUqTROnPNP3vZ9+C5Vw@mail.gmail.com> Signed-off-by: W. Trevor King <[email protected]>
We already require it for Linux/Unix-based systems [1], so we don't have to repeat it here. And other systems will use different paths, which we haven't specified yet. When I asked why we didn't specify a path for Windows [2], Vincent said we were waiting on help from PoC implementations [3]. So this commit punts the location to the "State" section, and lets the "Lifecycle" section just focus on when the write-to-filesystem happens. There's also discussion about removing the filesystem state registry completely [4], in which case we'd want to remove the whole line from the lifecycle. [1]: opencontainers@7713efc#diff-b84a8d65d8ed53f4794cd2db7e8ea731L7 [2]: opencontainers#211 (comment) [3]: opencontainers#211 (comment) [4]: https://groups.google.com/a/opencontainers.org/forum/#!topic/dev/q6TYqVZOcX8 Subject: removal of /run/opencontainer/containers Date: Wed, 25 Nov 2015 14:29:35 +0000 Message-ID: <CAD2oYtNipt3i_C6=J4Bc-jwauo5YAvKXUqTROnPNP3vZ9+C5Vw@mail.gmail.com> Signed-off-by: W. Trevor King <[email protected]>
We already require it for Linux/Unix-based systems [1], so we don't have to repeat it here. And other systems will use different paths, which we haven't specified yet. When I asked why we didn't specify a path for Windows [2], Vincent said we were waiting on help from PoC implementations [3]. So this commit punts the location to the "State" section, and lets the "Lifecycle" section just focus on when the write-to-filesystem happens. There's also discussion about removing the filesystem state registry completely [4], in which case we'd want to remove the whole line from the lifecycle. [1]: opencontainers@7713efc#diff-b84a8d65d8ed53f4794cd2db7e8ea731L7 [2]: opencontainers#211 (comment) [3]: opencontainers#211 (comment) [4]: https://groups.google.com/a/opencontainers.org/forum/#!topic/dev/q6TYqVZOcX8 Subject: removal of /run/opencontainer/containers Date: Wed, 25 Nov 2015 14:29:35 +0000 Message-ID: <CAD2oYtNipt3i_C6=J4Bc-jwauo5YAvKXUqTROnPNP3vZ9+C5Vw@mail.gmail.com> Signed-off-by: W. Trevor King <[email protected]>
5 participants
Signed-off-by: Doug Davis [email protected]