Hi Arik,
unfortunatly I couldn't discuss last friday with a colleague who needs to work
on the cluster this week-end. Discovering the freezed tasks, he found a
workaround by deleting the lines in the job table, and the tasks disappear in
the GUI. Then... He drops the faulty VM :(
But, I have shared the engine.log of april 11 and 12. The export to OVA of the
"CO7_VM1" vm was lauched around 17:01, with other vms (to the same directory on
host "victor"). Other vms exports end successfully but not this one.
Another export of the same vm was launched on april 12 at 10:42, to see if it
says "already working"... But it didn't, and the second task rans with the
first, indefinitly.
This links are active for 2 days :
https://seafile.systea.fr/f/12d90cc5c59b488c9fde/?dl=1
https://seafile.systea.fr/f/ce7cd61231924020bc62/?dl=1
I've not seen something revelant in it, but have a look ?
This morning I tried to reproduce the problem. I've create a vm from template
(50G). Migrate it : OK. Export it to OVA : OK.
I then extended the disk to 100G, then tested again. Migrate it : OK. Export it
to OVA...... OK :( So I wasn't able to reproduce the error.
Thx
Frank
Le Vendredi, Avril 13, 2018 21:53 CEST, Arik Hadas <aha...@redhat.com> a écrit:
On Fri, Apr 13, 2018 at 6:54 PM, fsoyer <fso...@systea.fr> wrote:Hi,
This task is listed (since 2 days) in the GUI / up right "tasks" icon. It is
visibly freezed as no OVA file has been created, but no errors in GUI, just...
it runs. Or : it loops :)
This (test) vm is one on which I have extended the disk (50 -> 100G). Before
being stopped and trying to export it to OVA, it works fine. All other vms
around can be well exported but not this one. I've tried to restart engine,
change SPM, restart one by one each node, but the task is always here. I could
even restart the vm today without error and it works fine ! But... the task
runs...
Today also, I tried to clone the vm : same thing, now I have 2 tasks running
indefinitly :(
Strange bug, where no timeout stopped the tasks in error. I can't see any
revelant things in engine.log or vdsm.log, but probably I've not seen them in
all the messages.
No problem to remove this (test) vm and try on another (test) one (extend disk
to see if this is the reason of the problem). But before I want to kill this
tasks ! Please don't remove that VM yet.It would be appreciated if you could
file a bug and share the engine log that covers the attempt to export this VM
to OVA + the ansible log of that operation.
Thanks.
Frank
Le Vendredi, Avril 13, 2018 16:24 CEST, Arik Hadas <aha...@redhat.com> a écrit:
On Fri, Apr 13, 2018 at 11:14 AM, fsoyer <fso...@systea.fr> wrote:Hi all,
I can't find any exhaustive doc for new vdsm-client. My problem actually is a
task (export a vm to OVA) blocked. I'm afraid you won't find any task in VDSM
for 'export a VM to OVA'.Expoting a VM to OVA is comprised of three steps:1.
Creating temporary cloned disks - you'll find a task of copy-image-group for
each disk.2. Creating the OVA file - that's done by a python script executed by
ansible, there is no task for that in VDSM.3. Removing the temporary cloned
disks - you'll find a task of remove-image for each disk. Can you please
elaborate on the problem you're having - where do you see that task and how can
you see that it's blocked? I found that I can interact with
vdsm-client Task getInfo taskID=<uuid>, and replace getInfo by "stop", BUT :
how can I find this UUID ???
Old "vdsClient -s 0 getAllTasksStatuses" has no equivalent ??
Does someone knows if a complete doc exists dot vdsm-client ?
Thanks
Frank
Le Mercredi, Janvier 25, 2017 12:30 CET, Irit Goihman <igoih...@redhat.com> a
écrit:
Hi All, vdsClient will be removed from master branch today.It is using XMLRPC
protocol which has been deprecated and replaced by JSON-RPC. A new client for
vdsm was introduced in 4.1: vdsm-client.This is a simple client that uses
JSON-RPC protocol which was introduced in ovirt 3.5. The client is not aware of
the available methods and parameters, and you should consult the schema [1] in
order to construct the desired command. Future version should parse the schema
and provide online help. If you're using vdsClient, we will be happy to assist
you in migrating to the new vdsm client. vdsm-client usage: vdsm-client [-h]
[-a ADDRESS] [-p PORT] [--unsecure] [--timeout TIMEOUT] [-f
FILE] namespace method [name=value [name=value] ...] Invoking simple methods:
# vdsm-client Host getVMList['b3f6fa00-b315-4ad4-8108-f73da817b5c5'] For
invoking methods with many or complex parameters, you can read the parameters
from a JSON format file: # vdsm-client Lease info -f lease.json where
lease.json file content is: {
"lease": {
"sd_id":
"75ab40e3-06b1-4a54-a825-2df7a40b93b2",
"lease_id": "b3f6fa00-b315-4ad4-8108-f73da817b5c5"
} }
It is also possible to read parameters from standard input, creating complex
parameters interactively: # cat <<EOF | vdsm-client Lease info -f -
{
"lease": {
"sd_id":
"75ab40e3-06b1-4a54-a825-2df7a40b93b2",
"lease_id": "b3f6fa00-b315-4ad4-8108-f73da817b5c5" }
}
EOF Constructing a command from vdsm schema: Let's take VM.getStats as an
example.This is the entry in the schema: VM.getStats:
added: '3.1'
description: Get statistics about
a running virtual machine. params:
- description: The UUID of the
VM name: vmID
type: *UUID
return:
description: An array
containing a single VmStats record type:
- *VmStats
namespace: VMmethod name:
getStatsparams: vmID The vdsm-client command is: # vdsm-client VM getStats
vmID=b3f6fa00-b315-4ad4-8108-f73da817b5c5 Invoking getVdsCaps command: #
vdsm-client Host getCapabilities Please consult vdsm-client help and man page
for further details and options. [1]
https://github.com/oVirt/vdsm/blob/master/lib/api/vdsm-api.yml --Irit
GoihmanSoftware EngineerRed Hat Israel Ltd.
_______________________________________________
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users
_______________________________________________
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users
