OSDir

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: 4.11.1 RC / VMs with UserDispersingPlanner won't deploy


Hi, Boris, Devs.

I use local-storage-only hosts, KVM hypervisor, no shared storage at all.
All local hosts are marked as PERMIT (storage tag) and PERMIT, HIGHCPU
(host tag). I deployed SO with PERMIT, HIGHCPU (host tags) and PERMIT
(storage tags). All logs I have at current runlevel are listed in the
thread. I will not have a chance to set log level to Debug until the 18
June due to vacations.

BTW, my cluster heads were upgraded from 4.10, but compute nodes are
created from scratch. I might stumble upon some artifacts because it's not
a clear deployment.



2018-05-31 14:42 GMT+07:00 Boris Stoyanov <boris.stoyanov@xxxxxxxxxxxxx>:

> Hi Ivan,
>
> I’m using a SO with the UD Planner and I have no tags on my hosts, I was
> able to deploy VMs on both KVM and VMware hosts.
> Could this be an issue with your env? Can you send logs we could look at?
>
> Regards,
> Boris Stoyanov.
>
>
> boris.stoyanov@xxxxxxxxxxxxx
> www.shapeblue.com
> 53 Chandos Place, Covent Garden, London  WC2N 4HSUK
> @shapeblue
>
>
>
> > On 30 May 2018, at 14:00, Ivan Kudryavtsev <kudryavtsev_ia@xxxxxxxxx>
> wrote:
> >
> > Hello, Dag, no, I have a single enabled cluster with enabled hosts, where
> > vms are deployed just fine from service offerings without host tags and
> > without UserDispersingPlanner, but with storage tags (since I have local
> > storage). But if I use a service offering with host tags which match host
> > tags exactly, literally and (or) UserDispersingPlanner is set, then
> things
> > stop work.
> >
> > During problem discovery I removed tags and left only
> > UserDispersingPlanner, vms won't deploy as well. I removed planner and
> left
> > tags, vms won't deploy as well. Only when no host tags and no
> > UserDispersingPlanner they deploy.
> >
> > 2018-05-30 17:54 GMT+07:00 Dag Sonstebo <Dag.Sonstebo@xxxxxxxxxxxxx>:
> >
> >> Hi Ivan,
> >>
> >> OK – understand – if those are the tags used then IMO it should work,
> and
> >> yes agree, userdisperseplanning should’t kick in with 0 VMs.
> >>
> >> Could the issue be with tagging rather than userdispersing:
> >>>> (logid:5ee25964) The clusterId list for the given offering tag: []
> >>
> >> This looks like it can’t find any clusters with given tag (even though
> you
> >> have them set)
> >>
> >>>> 2018-05-30 16:31:35,486 DEBUG [c.c.d.FirstFitPlanner]
> >>>> (API-Job-Executor-10:ctx-b36a77ea job-156334 ctx-9c358aab)
> >>>> (logid:5ee25964) No clusters found after removing disabled
> >> clusters and clusters in avoid list, returning.
> >>
> >> I appreciate I’m asking the obvious here so forgive me – but you’ve
> >> definitely not got disabled clusters (“after removing disabled
> clusters”)?
> >>
> >> Regards,
> >> Dag Sonstebo
> >> Cloud Architect
> >> ShapeBlue
> >>
> >> On 30/05/2018, 11:43, "Ivan Kudryavtsev" <kudryavtsev_ia@xxxxxxxxx>
> wrote:
> >>
> >>    Hi, Dag. Thank you very much, I really know that
> >>
> >>    What I have in my cloud:
> >>
> >>    Two local storage KVM hosts:
> >>    Host tags: PERMIT,HIGHCPU
> >>    Storage tags: PERMIT
> >>
> >>    I have service offering with tags:
> >>    Host tags: PERMIT,HIGHCPU
> >>    Storage tags: PERMIT
> >>
> >>    I don't have hosts without that labels at all, but, VM doesn't deploy
> >> even
> >>    if no anti-affinity is set, so no constraints should be breached.
> >> Also, I
> >>    don't catch how UserDispersingPlanner may interfere, if I have 0 VMs
> >>    deployed.
> >>
> >>    It worked just fine in previous releases (at least 4.10, 4.9.3, 4.3).
> >> But
> >>    now, in 4.11.1 RC it doesn't.
> >>
> >>
> >>    2018-05-30 17:36 GMT+07:00 Dag Sonstebo <Dag.Sonstebo@xxxxxxxxxxxxx
> >:
> >>
> >>> Hi Ivan,
> >>>
> >>> Host tags are an absolute requirement – so if host tag matching can
> >> not be
> >>> met due to affinity or planner then the VMs won’t start.
> >>>
> >>> Regards,
> >>> Dag Sonstebo
> >>> Cloud Architect
> >>> ShapeBlue
> >>>
> >>> On 30/05/2018, 11:33, "Ivan Kudryavtsev" <kudryavtsev_ia@xxxxxxxxx>
> >> wrote:
> >>>
> >>>    Sorry for spamming your e-mails, I finally have got the exact
> >> behavior:
> >>>
> >>>    If
> >>>      VM host tags set AND (OR) VM UerDispersingPlanner set
> >>>    then not able to deploy vm
> >>>
> >>>    Anti-affinity works fine (if first two are not set).
> >>>
> >>>
> >>>
> >>>    2018-05-30 17:18 GMT+07:00 Ivan Kudryavtsev <
> >> kudryavtsev_ia@xxxxxxxxx
> >>>> :
> >>>
> >>>> To be more precise, I see that error only under "admin" user.
> >> Users
> >>> from
> >>>> domains can create VMS with anti-affinity groups fine.
> >>>>
> >>>> 2018-05-30 16:38 GMT+07:00 Ivan Kudryavtsev <
> >>> kudryavtsev_ia@xxxxxxxxx>:
> >>>>
> >>>>> Hello, Devs. I have a fresh 4.11.1 RC with two KVM hosts
> >> deployed.
> >>>>>
> >>>>> First, I unable to deploy instances with UserDispersingPlanner
> >>> planner.
> >>>>> The errors look like:
> >>>>>
> >>>>> 2018-05-30 16:31:35,372 DEBUG [c.c.a.ApiServlet]
> >>>>> (qtp1856056345-27:ctx-3fb8db1f ctx-d8415fd2) (logid:dbe8f2ce)
> >>> ===END===
> >>>>> 91.221.61.126 -- GET  command=deployVirtualMachine&r
> >>>>> esponse=json&zoneid=d477bb3f-3592-4503-8f2a-da3d878dd476&tem
> >>>>> plateid=369815dc-0622-4645-8ca4-c6432baebe89&hypervisor=KVM&
> >>>>> serviceofferingid=eab6a6fa-ae7d-4e39-9ddf-32285cdfd20b&
> >>>>> affinitygroupids=c2ca0ddf-9e87-4d7c-9e91-3f6eb56f30f4&
> >>>>> securitygroupids=c1ec899c-c69b-11e7-bdcf-0242ac110004&
> >>>>> displayname=aaa4&name=aaa4&_=1527672695121
> >>>>> 2018-05-30 16:31:35,456 DEBUG [c.c.d.
> >> DeploymentPlanningManagerImpl]
> >>>>> (API-Job-Executor-10:ctx-b36a77ea job-156334 ctx-9c358aab)
> >>>>> (logid:5ee25964) DeploymentPlanner allocation algorithm: null
> >>>>> 2018-05-30 16:31:35,456 DEBUG [c.c.d.
> >> DeploymentPlanningManagerImpl]
> >>>>> (API-Job-Executor-10:ctx-b36a77ea job-156334 ctx-9c358aab)
> >>>>> (logid:5ee25964) Trying to allocate a host and storage pools
> >> from
> >>> dc:1,
> >>>>> pod:null,cluster:null, requested cpu: 9000, requested ram:
> >>> 12884901888
> >>>>> 2018-05-30 16:31:35,456 DEBUG [c.c.d.
> >> DeploymentPlanningManagerImpl]
> >>>>> (API-Job-Executor-10:ctx-b36a77ea job-156334 ctx-9c358aab)
> >>>>> (logid:5ee25964) Is ROOT volume READY (pool already
> >> allocated)?: No
> >>>>> 2018-05-30 16:31:35,460 DEBUG [o.a.c.a.
> >> HostAntiAffinityProcessor]
> >>>>> (API-Job-Executor-10:ctx-b36a77ea job-156334 ctx-9c358aab)
> >>>>> (logid:5ee25964) Processing affinity group test for VM Id:
> >> 2149
> >>>>> 2018-05-30 16:31:35,473 DEBUG [c.c.d.
> >> DeploymentPlanningManagerImpl]
> >>>>> (API-Job-Executor-10:ctx-b36a77ea job-156334 ctx-9c358aab)
> >>>>> (logid:5ee25964) Deploy avoids pods: [], clusters: [], hosts:
> >> []
> >>>>> 2018-05-30 16:31:35,474 DEBUG [c.c.d.FirstFitPlanner]
> >>>>> (API-Job-Executor-10:ctx-b36a77ea job-156334 ctx-9c358aab)
> >>>>> (logid:5ee25964) Searching all possible resources under this
> >> Zone: 1
> >>>>> 2018-05-30 16:31:35,476 DEBUG [c.c.d.FirstFitPlanner]
> >>>>> (API-Job-Executor-10:ctx-b36a77ea job-156334 ctx-9c358aab)
> >>>>> (logid:5ee25964) Listing clusters in order of aggregate
> >> capacity,
> >>> that have
> >>>>> (atleast one host with) enough CPU and RAM capacity under this
> >>> Zone: 1
> >>>>> 2018-05-30 16:31:35,479 DEBUG [c.c.d.FirstFitPlanner]
> >>>>> (API-Job-Executor-10:ctx-b36a77ea job-156334 ctx-9c358aab)
> >>>>> (logid:5ee25964) Removing from the clusterId list these
> >> clusters
> >>> from avoid
> >>>>> set: []
> >>>>> 2018-05-30 16:31:35,486 DEBUG [c.c.d.FirstFitPlanner]
> >>>>> (API-Job-Executor-10:ctx-b36a77ea job-156334 ctx-9c358aab)
> >>>>> (logid:5ee25964) The clusterId list for the given offering
> >> tag: []
> >>>>> 2018-05-30 16:31:35,486 DEBUG [c.c.d.FirstFitPlanner]
> >>>>> (API-Job-Executor-10:ctx-b36a77ea job-156334 ctx-9c358aab)
> >>>>> (logid:5ee25964) No clusters found after removing disabled
> >> clusters
> >>> and
> >>>>> clusters in avoid list, returning.
> >>>>> 2018-05-30 16:31:35,490 DEBUG [c.c.v.UserVmManagerImpl]
> >>>>> (API-Job-Executor-10:ctx-b36a77ea job-156334 ctx-9c358aab)
> >>>>> (logid:5ee25964) Destroying vm VM[User|i-2-2149-VM] as it
> >> failed to
> >>> create
> >>>>> on Host with Id:null
> >>>>>
> >>>>> It's not in the "deployment.planners.exclude" btw.
> >>>>>
> >>>>> I removed from offerings that attribute:
> >>>>> update service_offering_view set deployment_planner = null;
> >>>>>
> >>>>> and they can be deployed right now.
> >>>>>
> >>>>>
> >>>>> Second, VMs with anti-affinity also won't deploy:
> >>>>>
> >>>>> The message is exactly the same by meaning.
> >>>>>
> >>>>>
> >>>>>
> >>>>> --
> >>>>> With best regards, Ivan Kudryavtsev
> >>>>> Bitworks Software, Ltd.
> >>>>> Cell: +7-923-414-1515
> >>>>> WWW: http://bitworks.software/ <http://bw-sw.com/>
> >>>>>
> >>>>>
> >>>>
> >>>>
> >>>> --
> >>>> With best regards, Ivan Kudryavtsev
> >>>> Bitworks Software, Ltd.
> >>>> Cell: +7-923-414-1515
> >>>> WWW: http://bitworks.software/ <http://bw-sw.com/>
> >>>>
> >>>>
> >>>
> >>>
> >>>    --
> >>>    With best regards, Ivan Kudryavtsev
> >>>    Bitworks Software, Ltd.
> >>>    Cell: +7-923-414-1515
> >>>    WWW: http://bitworks.software/ <http://bw-sw.com/>
> >>>
> >>>
> >>>
> >>> Dag.Sonstebo@xxxxxxxxxxxxx
> >>> www.shapeblue.com
> >>> 53 Chandos Place, Covent Garden, London  WC2N 4HSUK
> >>> @shapeblue
> >>>
> >>>
> >>>
> >>>
> >>
> >>
> >>    --
> >>    With best regards, Ivan Kudryavtsev
> >>    Bitworks Software, Ltd.
> >>    Cell: +7-923-414-1515
> >>    WWW: http://bitworks.software/ <http://bw-sw.com/>
> >>
> >>
> >>
> >> Dag.Sonstebo@xxxxxxxxxxxxx
> >> www.shapeblue.com
> >> 53 Chandos Place, Covent Garden, London  WC2N 4HSUK
> >> @shapeblue
> >>
> >>
> >>
> >>
> >
> >
> > --
> > With best regards, Ivan Kudryavtsev
> > Bitworks Software, Ltd.
> > Cell: +7-923-414-1515
> > WWW: http://bitworks.software/ <http://bw-sw.com/>
>
>


-- 
With best regards, Ivan Kudryavtsev
Bitworks Software, Ltd.
Cell: +7-923-414-1515
WWW: http://bitworks.software/ <http://bw-sw.com/>