For example , I "inspect busybox", the Size and VirtualSize are both equal to 1112820, but in docker, the Size is 0 and the VirtualSize is 1112820.So this maybe is a little carelessness.

First I run a container called "parent",and stop it.Then I create a container "child-first" with "--link parent:parent", and I try to start it, it return the container name as if it successed.
But in fact it should return the error message "Error response from daemon: Cannot start container child-first: Cannot link to a non running container: /parent AS /child-first/parent"," Error: failed to start containers: [child-first]"。

hyper has the container name regex([a-z0-9]([-a-z0-9]*[a-z0-9])?). So, we cant use name like xl_gao. But the random name appears name like cranky_archimedes. I think its a tiny bug, but its confusing.

How to Reproduce:

hyper run  -d  busybox false

hyper run -d --name container1 busybox top
hyper run -it --link container1:alias1 busybox sh 

ping alias1 works
ping container1 failed

The status code of successfully rm snapshot should be 204, but in fact the status code is 201.

How to Reproduce:

xlgao@xlgao-pc:~/gocode/src/github.com/hyperhq/hypercli/hyper$ hyper -H tcp://147.75.195.39:6443 run -d busybox sh
54d49f51e70efb4e65f35b8cddc9c9149d54b04ae529f05a6323f739872ed914
xlgao@xlgao-pc:~/gocode/src/github.com/hyperhq/hypercli/hyper$ hyper -H tcp://147.75.195.39:6443 exec -i 54d49f51e70efb4e65f35b8cddc9c9149d54b04ae529f05a6323f739872ed914 /bin/sh

but '_' is using in default names.

I have a lot of volumes and snapshots which are dependent with each other. After some operations, I deleted all the snapshots, while there are some volumes I can't remove, Error Message "Volume still has 1 dependent snapshots."

hyper info :
Containers: 7
 Running: 1
 Paused: 0
 Stopped: 6
Images: 4
Server Version: Hyper_ beta
Storage Driver: hyper
Execution Driver: hyper.sh
Plugins: 
 Volume: hyper storage
 Network: hyper network
 Authorization: Access-Key
Kernel Version: 4.0.0
CPUs: 0
Total Memory: 0 B
ID: f7929220601e4823974196051bb80fe4
Debug mode (client): false
Debug mode (server): false

But I couldn't reproduce this bug anymore, maybe It's caused by server updating recently.
If I can reproduce this in the future, I'll update this issue later.

But truely, there is some inconsistence in snapshot when I try to reproduce the bug.
How to reproduce:

hyper create volume --name test-volume-1
hyper create snapshot --name test-snap-1 -v test-volume-1
hyper run --name test-rm -v test-volume-1:/opt/data -d busybox sh
hyper rm -fv test-rm

If you run command hyper snapshot ls, there is no snapshot test-snap-1 anymore. But It do exist.
you can delete it again by run hyper snapshot rm test-snap-1 or create a volume base on test-snap-1. I think it's command rm -v which cause the inconsistence.

"DELETE", "/containers/$container_id?force=1"

https://docs.hyper.sh/Reference/API/1.23/Container/remove.html

Api succeed:
want code 200, got 204

At first I create a volume "test", I attach it with a container.Then I stop and rm the container.But I cannot rm "test", because " Volume test is still attached, detach volume first.".In fact the container has already disappeared.
And if I create the snapshot based on the above volume "test", it also will meet the error " The server encountered an internal error or misconfiguration...".

The option of "-f" is useless.As long as the argument satisfy the format "key=value", then it will print all of the volume.Of course, the format "dangling=invalid" which should be invalid is OK.
It is better to add the feature when I pass "-f dangling=true",it will only print the volume in use.And other invalid argument like "-f FOO=123" will return error.

I run a container with "-v /data", when I inspect that container, the Mounts item is still empty!

(1) First, run a container with command "hypercli run -di busybox cat".
Then request "POST", "/containers/"+cid+"/attach?stream=1&stdin=1&stdout=1", and read from stdout which should success, but timeout.
(2) Run a container with command "hypercli run -di busybox /bin/sh -c cat >&2"
Then request "POST", "/containers/"+cid+"/attach?stream=1&stdin=1&stderr=1" , and read from stderr which should success,but timeout
Request "POST", "/containers/"+cid+"/attach?stream=1&stdin=1&stdout=1",and read from stderr which should timeout, but err is nil
(3)Run a container with command "hypercli run -dit busybox /bin/sh -c cat >&2"
Then request "POST", "/containers/"+cid+"/attach?stream=1&stdin=1&stdout=1", and read from stdout which should success but timeout

we should pass this test case, but for now ,we can not pass it.

dockerCmd(c, "run", "-d", "-t", "--name", name, "busybox", "/bin/sh")

status, body, err := sockRequest("GET", fmt.Sprintf("/containers/%s/logs", name), nil)
c.Assert(status, checker.Equals, http.StatusBadRequest)
c.Assert(err, checker.IsNil)

expected := "Bad parameters: you must choose at least one stream"

I get

... obtained int = 500
... expected int = 400

How to reproduce:

hyper run -d -t hyperhq/test-port-list sh
hyper port $container_id
hyper port $container_id
......

Return value list is not in order

Start a running container using api , get status code 204, want 304.

there two issues about exec api

• 1. call exec start after stopping container, code like this

id = createExec(c, "test")
dockerCmd(c, "stop", "test")
startExec(c, id, http.StatusNotFound)

... obtained int = 200
... expected int = 409

• 2. call exec start multiple times, code like this

runSleepingContainer(c, "-d", "--name", "test")
execID := createExec(c, "test")
startExec(c, execID, http.StatusOK)
startExec(c, execID, http.StatusConflict)

... obtained int = 200
... expected int = 409

1. When kill a running container, get error message "Error response from daemon: Cannot kill container X: Container X is not running".
   How to Reproduce:
   hyper run -d busybox top
hyper kill $contianer_id
2. When kill a stoped container, we should get an error message, something like we got from [1]. While we only got a container id.

How to Reproduce:

hyper -H run -e LALA=value1 -e LALA=value2 -d busybox top

I get failed when I go through these test cases.

func (s *DockerSuite) TestLogsFollowGoroutinesWithStdout(c *check.C)

func (s *DockerSuite) TestLogsFollowGoroutinesNoOutput(c *check.C)

hyper run -d --name A busybox top
hyper run -d --name B --link A:aliasA busybox top
hyper inspect -f {{ .HostConfig.Links }} $container_id

Got: "A:aliasA"
Expected: "/A:/B/aliasA"

When I rm a volume in use,it return the status code 400, but it should return 409 which means statusConflict

Can't delete container bc137a8a38b5 in pkt environment.

When I remove a nonexistent container, I get a panic. The error message is shown below.

panic: runtime error: invalid memory address or nil pointer dereference
[signal 0xb code=0x1 addr=0x20 pc=0x53ffb1]

goroutine 1 [running]:
panic(0xa332a0, 0xc82000e190)
    /opt/go/src/runtime/panic.go:464 +0x3e6
encoding/json.(*Decoder).refill(0xc8203dc000, 0x0, 0x0)
    /opt/go/src/encoding/json/stream.go:152 +0x281
encoding/json.(*Decoder).readValue(0xc8203dc000, 0x1, 0x0, 0x0)
    /opt/go/src/encoding/json/stream.go:128 +0x413
encoding/json.(*Decoder).Decode(0xc8203dc000, 0x902120, 0xc82025b100, 0x0, 0x0)
    /opt/go/src/encoding/json/stream.go:57 +0x159
github.com/docker/engine-api/client.(*Client).ContainerRemove(0xc8200a6240, 0x7ffde265d190, 0x4, 0xc820000000, 0x0, 0x0, 0x0, 0x0, 0x0)
    /home/xlgao/gocode/src/github.com/hyperhq/hypercli/vendor/src/github.com/docker/engine-api/client/container_remove.go:26 +0x50e
github.com/hyperhq/hypercli/api/client.(*DockerCli).removeContainer(0xc8200a61b0, 0x7ffde265d190, 0x4, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0)
    /home/xlgao/gocode/src/github.com/hyperhq/hypercli/api/client/rm.go:55 +0xcb
github.com/hyperhq/hypercli/api/client.(*DockerCli).CmdRm(0xc8200a61b0, 0xc8200780b0, 0x1, 0x1, 0x0, 0x0)
    /home/xlgao/gocode/src/github.com/hyperhq/hypercli/api/client/rm.go:31 +0x5c4
reflect.callMethod(0xc8202c48d0, 0xc820199c28)
    /opt/go/src/reflect/value.go:631 +0x1fc
reflect.methodValueCall(0xc8200780b0, 0x1, 0x1, 0x1, 0xc8202c48d0, 0x0, 0x0, 0xc8202c48d0, 0x0, 0x468724, ...)
    /opt/go/src/reflect/asm_amd64.s:29 +0x36
github.com/hyperhq/hypercli/cli.(*Cli).Run(0xc8202c4540, 0xc8200780a0, 0x2, 0x2, 0x0, 0x0)
    /home/xlgao/gocode/src/github.com/hyperhq/hypercli/cli/cli.go:89 +0x384
main.main()
    /home/xlgao/gocode/src/github.com/hyperhq/hypercli/hyper/hyper.go:63 +0x435

hyper run -d busybox top
hyper create --link $container_id busybox top

Got error "Error response from daemon: Link alias ($container_id provided) should fit RFC11123 format a-z0-9?"

When I hyper stats a non-existent container, the return message is not stable.Sometimes it may return the message that the container is not exist.But sometimes it may return the normal stats as if it is an existent container.
Especially when I hyper stats a non-existent container with "--no-stream=false", at first it may return the stats, but after a while the stats disappear as if it realize the mistake!

How to Reproduce:

hyper run -d --name child busybox top
hyper run -d --name parent --link=child:linkedone busybox top
hyper ps --format {{.Names}} --no-trunc

should get

parent
child,parent/linkedone

actually get

parent
child

At first , I run a container "hypercli run -itd busybox cat", and then attach it , input "hello\n", in test case TestAttachDetachTruncatedID(docker_cli_attach_unix_test.go)，it read an empty string and err = EOF.But if I do all of it by hand ,everything seems OK , its so confusing.And delay is useless

[root@81b7659c2a09 integration-cli]# hyper -H ${DOCKER_HOST} rmi ""
Error response from daemon: 404: Page Not Found
Obvisouly,this corner is not considered.In docker, the daemon will return "image name cannot be blank"

In both hyper and docker, volume can be created in two ways:

1. volume create [--name] --- created volume
2. run -v /foo --- temp volume
   Both kind of volume will be removed by hyper if rm -v or --rm takes effect. While docker will only remove temp volume created by 2nd way.
   Concerned test case: TestRunNamedVolumesFromNotRemoved

test case: TestRunRestartMaxRetries

hyper run -d --restart=on-failure:3 busybox false
hyper inpsect -f {{ .RestartCount }} $container_id

want 3, get 0

func (s *DockerSuite) TestLogsSince(c *check.C)
This test case cannot pass by ci, but I can pass this by myself.

docker run -e FOO busybox env
Get result:

PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
HOSTNAME=2da5648c6781
HOME=/root

hyper run -e FOO busybox env
Get result:

hyper: Error response from daemon: Misconfigured environment FOO, Env name should match [A-Za-z_][A-Za-z0-9_]*.

enviroment FOO is not configured on local host.

Expect:
behavior should be consist with docker Or use a more inspiring error message.

https://github.com/hyperhq/hypercli/blob/integration-test/integration-cli/todo/cli/docker_cli_run_test.go#L772-L879

I create a container by nonexistent image such as "test234".As usual it should return the error message "No such image: test456:latest",but in hyper, it will response "No such image: test456",without tag "latest".

hyper run -d busybox sh -c "sleep 60s"
hyper kill $container_id
hyper ps -n=1

After kill this container, it's still running.

testcase: TestRunLeakyFileDescriptors

hyper run busybox ls -C /proc/self/fd

get result: "0 1 2 21 22 23 3", not "0 1 2 3"

Now, hyper logs could only get a running container's logs, which is different from docker. Is that normal or just a bug not been fixed yet?
Similar to this,

• If we get logs from a stopped container, there will be an error.
• If we restart a container, logs from the previous running will no longer be exist.

How to reproduce:

hyper run -h foo.bar.baz busybox hostname

Got error timeout to wait container state

"scratch" is a reserved name.

$ hyper pull scratch
Using default tag: latest
Error response from daemon: The server encountered an internal error or misconfiguration...

$ docker pull scratch
Using default tag: latest
Error response from daemon: 'scratch' is a reserved name

"POST", "/containers/create", config

If config is empty, should get error message "Config cannot be empty in order to create a container\n"
not "The server encountered an internal error or misconfiguration...\n"

How to Reproduce:

hypercli run -d --name t1 busybox top
hypercli run -d --name t2 busybox top
hypercli run -d --name t3 busybox true
hypercli run -d --name t4 busybox top

hypercli ps -f "before=t3" will get error like this Error response from daemon: The server encountered an internal error or misconfiguration....

hypercli ps -f "before=t3'ID" will not get any container.

"GET" "/containers/json?all=1"

If result is blank, want "[]\n", get "[]"

At first I "hypercli stats" , and then I run a new container。But the stats of the new container won't show up in the previous "hypercli stats" console.

hyper run bosybox foo
$?

want exit code 127, get 255.
Related test code: https://github.com/hyperhq/hypercli/blob/integration-test/integration-cli/todo/cli/docker_cli_run_test.go#L3882-L3919

All the fip apis are cool, except fip release. I can't pass the test case about fip release.

The test code is test code about fip release.

we should get 204, but we get 200. And after I called fip release api, I run hyper fip ls, I find that ip isn't released.

(1)In the API doc, all the APIs of snapshot miss 's'.
For example, snapshot creating is "/snapshot/create", but in fact it should be "/snapshots/create".
(2)In the doc of snapshot create:" POST /snapshot/createi?name=snapshot-tardis&volume=tardis",there is no "i" after "create".

How to Reproduce:

hyper exec container-name sh -c "while ! test -e /execid1; do sleep 1; done"
hyper inspect -f {{.ExecIDs}} container-name