[
https://issues.apache.org/jira/browse/YARN-8940?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16661862#comment-16661862
]
Weiwei Yang commented on YARN-8940:
-----------------------------------
Hi [~eyang]
Here is the list of access modes supported in CSI spec
{code}
UNKNOWN = 0;
// Can only be published once as read/write on a single node, at
// any given time.
SINGLE_NODE_WRITER = 1;
// Can only be published once as readonly on a single node, at
// any given time.
SINGLE_NODE_READER_ONLY = 2;
// Can be published as readonly at multiple nodes simultaneously.
MULTI_NODE_READER_ONLY = 3;
// Can be published at multiple nodes simultaneously. Only one of
// the node can be used as read/write. The rest will be readonly.
MULTI_NODE_SINGLE_WRITER = 4;
// Can be published as read/write at multiple nodes
// simultaneously.
MULTI_NODE_MULTI_WRITER = 5;
{code}
for a pre-provisioned volume, we'll verify if the given volume supports this
access mode; for a dynamical provisioned volume, we create the volume and set
the volume using the given mode.
For the mount propagation, it is not defined in the CSI spec, but it says
bq. The CO MUST clearly document the mount propagation requirements for Node
Plugins and the Plugin Supervisor SHALL satisfy the CO’s requirements.
I think I need to dig more when dealing with mount points, this sounds like
different storage system may have different mount propagation requirements. But
this won't surface from the user level, I think.
> Add volume as a top-level attribute in service spec
> ----------------------------------------------------
>
> Key: YARN-8940
> URL: https://issues.apache.org/jira/browse/YARN-8940
> Project: Hadoop YARN
> Issue Type: Sub-task
> Reporter: Weiwei Yang
> Priority: Major
>
> Initial thought:
> {noformat}
> {
> "name": "volume example",
> "version": "1.0.0",
> "description": "a volume simple example",
> "components" :
> [
> {
> "name": "",
> "number_of_containers": 1,
> "artifact": {
> "id": "docker.io/centos:latest",
> "type": "DOCKER"
> },
> "launch_command": "sleep,120",
> "configuration": {
> "env": {
> "YARN_CONTAINER_RUNTIME_DOCKER_RUN_OVERRIDE_DISABLE":"true"
> }
> },
> "resource": {
> "cpus": 1,
> "memory": "256",
> },
> "volumes": [
> {
> "volume" : {
> "type": "s3_csi",
> "id": "5504d4a8-b246-11e8-94c2-026b17aa1190",
> "capability" : {
> "min": "5Gi",
> "max": "100Gi"
> },
> "source_path": "s3://my_bucket/my", # optional for object stores
> "mount_path": "/mnt/data", # required, the mount point in
> docker container
> "access_mode": "SINGLE_READ", # how the volume can be accessed
> }
> }
> ]
> }
> }
> ]
> }
> {noformat}
> Open for discussion.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
