[
https://issues.apache.org/jira/browse/FALCON-710?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14137597#comment-14137597
]
Balu Vellanki commented on FALCON-710:
--------------------------------------
{quote}
is there a particular reason Balu Vellanki we don't want to support time window
on suspend/kill/resume ?
{quote}
Yes, because it can let users make un-intentional kill/suspend calls. I do not
imagine a CLI situation where users will not know what instances they want to
kill/suspend/resume before they make the API call. Is there such a use-case?
IMHO Falcon should not encourage such behavior.
In Falcon dashboard, the UI will list all instances currently running, with a
checkbox next to each instance for POST operations. When the POST call is made,
UI will know the exact instance on which to run these operations.
> As per the latest behaviour of instance request, default values of start/end
> apply to APIs like kill and suspend as well
> ------------------------------------------------------------------------------------------------------------------------
>
> Key: FALCON-710
> URL: https://issues.apache.org/jira/browse/FALCON-710
> Project: Falcon
> Issue Type: Bug
> Components: merlin
> Environment: QA
> Reporter: Karishma Gulati
> Assignee: Balu Vellanki
>
> Given the latest change made in the default behaviour of instance request.
> the default values of start and end should be applicable only when getting
> the status of instances, and not for APIs like kill and suspend as well.
> For eg, if the instances of a process created are from the past, and a
> suspend/kill request is made without specifying the start and end, all
> instances of that process are suspended/killed.
> For process validity : start - 2010-01-02T01:00Z and end - 2010-01-02T01:26Z,
> the URL
> {code}
> Request Url:
> http://192.168.138.137:16000/api/instance/kill/process/agregator-coord16-3ca03cba/?colo=*&user.name=karishma
> {code}
> returns the following response
> {code}
> {
> "status": "SUCCEEDED",
> "message": "ua1/KILL\n",
> "requestId": "ua1/09285aae-f571-408c-a8f4-989cba9cbd19\n",
> "instances": [
> {
> "instance": "2010-01-02T01:25Z",
> "status": "KILLED",
> "logFile":
> "http://ip-192-168-138-137:11000/oozie?job\u003d0000894-140828065226481-oozie-oozi-W";,
> "cluster": "corp-953ec81e",
> "startTime": "2014-09-11T09:29:30Z",
> "endTime": "2014-09-11T09:29:49Z",
> "details": ""
> },
> {
> "instance": "2010-01-02T01:20Z",
> "status": "KILLED",
> "logFile":
> "http://ip-192-168-138-137:11000/oozie?job\u003d0000892-140828065226481-oozie-oozi-W";,
> "cluster": "corp-953ec81e",
> "startTime": "2014-09-11T09:29:28Z",
> "endTime": "2014-09-11T09:29:50Z",
> "details": ""
> },
> {
> "instance": "2010-01-02T01:15Z",
> "status": "KILLED",
> "logFile":
> "http://ip-192-168-138-137:11000/oozie?job\u003d0000890-140828065226481-oozie-oozi-W";,
> "cluster": "corp-953ec81e",
> "startTime": "2014-09-11T09:29:26Z",
> "endTime": "2014-09-11T09:29:52Z",
> "details": ""
> },
> {
> "instance": "2010-01-02T01:10Z",
> "status": "KILLED",
> "logFile":
> "http://ip-192-168-138-137:11000/oozie?job\u003d0000888-140828065226481-oozie-oozi-W";,
> "cluster": "corp-953ec81e",
> "startTime": "2014-09-11T09:29:25Z",
> "endTime": "2014-09-11T09:29:53Z",
> "details": ""
> },
> {
> "instance": "2010-01-02T01:05Z",
> "status": "KILLED",
> "logFile":
> "http://ip-192-168-138-137:11000/oozie?job\u003d0000886-140828065226481-oozie-oozi-W";,
> "cluster": "corp-953ec81e",
> "startTime": "2014-09-11T09:29:24Z",
> "endTime": "2014-09-11T09:29:56Z",
> "details": ""
> },
> {
> "instance": "2010-01-02T01:00Z",
> "status": "KILLED",
> "logFile":
> "http://ip-192-168-138-137:11000/oozie?job\u003d0000884-140828065226481-oozie-oozi-W";,
> "cluster": "corp-953ec81e",
> "startTime": "2014-09-11T09:29:24Z",
> "endTime": "2014-09-11T09:29:58Z",
> "details": ""
> }
> ]
> }
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
