update golang to v1.17

[Thor-wl]

As kubernetes-sigs/controller-tools#643 mentions, there may be some bugs in package encoding for golang v1.16.x. This PR should be merged before #1910
Signed-off-by: Thor-wl 13164644535@163.com

[Thor-wl]

Not sure which PR has introduces this bug or is there something wrong with this UT case. Ignore it temporarily. Will look into it and then give the fix.

[hwdef]

I met this error, too.
Very confused

[Thor-wl]

I reviewed the UT file allocate_test.go and it was not changed recently. So I guess there may be some wrong update with the allocate.go. @justadogistaken I noticed that only this PR #1906 was merged with this file recently. So can you help give a test about this bug? I've tested for serval times and found that the allocations are not stable. Namely, the output allocations are not unique. There are small probability to be same with expected.

[justadogistaken]

I reviewed the UT file allocate_test.go and it was not changed recently. So I guess there may be some wrong update with the allocate.go. @justadogistaken I noticed that only this PR #1906 was merged with this file recently. So can you help give a test about this bug? I've tested for serval times and found that the allocations are not stable. Namely, the output allocations are not unique. There are small probability to be same with expected.

I reset my commit to the pr ahead of PR #1906, this error exists previously. But this may not be an error. Pods of c1/p1, c1/p2, c2/p1, c2/p2 have same requests. This error will happen if c1/p2 gets into tasks(priorityQueue) first or c2/p2 gets into first instead of p1.

[justadogistaken]

Let's check this out.
TaskStatusIndex will return a taskMap, which is like map[TaskID]*TaskInfo.
And map does not have a concrete order while being accessed.
So it could happen to p2 is accessed before p1 in a pg.
But the result is correct.
The problem is that the "expected: "c2/p1": "n1", "c1/p1": "n1"" in UT is not appropriate.
It should be like "c1/ p1 or p2, c2/ p1 or p2".

			for _, task := range job.TaskStatusIndex[api.Pending] {
				// Skip BestEffort task in 'allocate' action.
				if task.Resreq.IsEmpty() {
					klog.V(4).Infof("Task <%v/%v> is BestEffort task, skip it.",
						task.Namespace, task.Name)
					continue
				}

				tasks.Push(task)
			}

[huone1]

/lgtm

[huone1]

/approve

[justadogistaken]

There was a pr #1898 removing the follwing block.
Tasks will be sort by UID(***-p1, ***-p2) Originally.

	if lv.Pod.CreationTimestamp.Equal(&rv.Pod.CreationTimestamp) {
		return lv.UID < rv.UID
	}
	return lv.Pod.CreationTimestamp.Before(&rv.Pod.CreationTimestamp)

[Thor-wl]

There was a pr #1898 removing the follwing block. Tasks will be sort by UID(***-p1, ***-p2) Originally.

	if lv.Pod.CreationTimestamp.Equal(&rv.Pod.CreationTimestamp) {
		return lv.UID < rv.UID
	}
	return lv.Pod.CreationTimestamp.Before(&rv.Pod.CreationTimestamp)

/cc @xiaoanyunfei Pls help take a look and make corresponding UT cases changed if the allocate logic changes. THX.

[huone1]

/lgtm

[shinytang6]

/lgtm
/approve

[volcano-sh-bot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: huone1, shinytang6

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [shinytang6]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment