Migrating a VM to another ressource pool fails : Hub "Resources" doesn't meet compute requirement:

[thgreyprnc]

Hello

I have an issue when trying to migrate a VM to another ressource pool

     -within vCD 8.20 I have 2 ressource pools, both associated to the same Provider VDC. those 2 ressource pools are related, within vCenter, to 2 separate clusters : "SOURCE" and "DEST"

     -all hosts from both clusters are connected to the same DVS switch.

     -One host of "DEST" has an FC adapter trough which he is able to see all datastores from the SOURCE cluster. (( The reason of doing so is because DEST will be using vSAN in the future, so that host will be our "bridge" between the legacy FC storage and vSAN : once a VM is going to be vMotioned to it, we will update it's storage policy to vSAN thus triggering a storage vMotion))

     -EVC is enabled in "sandy bridge" mode on both clusters.

     -a vMotion of a VM initiated from vsphere web client between "SOURCE" and "DEST" works perfectly fine.

     -All VM's are located in the ressource pool "SOURCE". The goal is to migrate them to "DEST".

     -from vCD when I try to migrate a VM to my ressource pool "DEST", I am granted with the following error :

               [ 30836e4b-f761-438d-b263-9aecd5878a0d ] The operation failed because no suitable resource was found.

               Hub "Resources" doesn't meet compute requirement:

               MemoryMBMin:819, mandatory:true, checkMinHostCapacity:false.

                -

                The operation failed because no suitable resource was found.

               Hub "Resources" doesn't meet compute requirement:

               MemoryMBMin:819, mandatory:true, checkMinHostCapacity:false.

                -

                PlacementException NO_FEASIBLE_PLACEMENT_SOLUTION

This error seems to indicate that there are not enough memory ressources available in the "DEST" pool.

The VDC in which the VM I am trying to migrate is sitting is an "Allocation pool" and the parameter "Make Allocation pool Org VDCs elastic" is enabled in vCD.

My DEST cluster is new, not running any VM with a total of 1.35Tb of ram available.

Now, if you look at the attached screen shots, I noticed that my destination cluster is not displaying any "total" ressources but maybe it's normal ?

Of course, In the org VDC related to the VM I am trying to migrate, I defined all reservations on the lowest setting I could.

I feel I am missing something obvious but... what ?!?

Thanks !

[thgreyprnc]

I saw this : Unable to power on VM from vCloud Director

Could it be linked to the fact that my DEST cluster only has 1 host out of 6 which is able to see Datastores from the SOURCE cluster ?

[thgreyprnc]

Pfff 😕 still 0 answers...

It's hard to get how vCD is actually following each single aspect of murphy's law... and nobody on internet seems to have any clue...

Seriously, each single little thing that can go wrong does go wrong with this piece of ****. For me this product is a giant fail and I already advised my company to move FAR AWAY from that as soon as possible !

I simply want the damn thing to execute a simple vmotion from a VM to another ressource pool on a destination host which has exact same storage/network access as all my source hosts. Or a least trying to start the vmotion, not simply spitting me that meaningless error message !!! Or why does that vmotion perfectly work when executed from vsphere web client ? On a destination empty cluster how could there be processor or memory ressources missing ? On the other hand why the heck would vCD care about the status of other hosts of this cluster ? Just DO what I tell you to do !!!! This is real BullshitⓇ

[mhampto]

From what you have provided this seems like the most likely cause of the error. It is looking at the entire VCD environment (if I understand the configuration correctly) and all hosts would temporarily need access while migrating.

[thgreyprnc]

Ok thanks for your input.

If this is the case, sorry but it is pure non-sense !

My destination cluster is already running in a vsan streched config... Which means that half of the hosts are located on plant-B where I don't have any FC connectivity to plant-A.

This means we forget about the idea of adding an FC adapter on all hosts.

Furthermore, it means that I basically have 1 single choice :

destroy my vsan config, leave only 3 hosts in my DEST cluster PLANT-A side, add an fc adapter to my 2 remaining hosts and blablabla. when all vm's will be migrated to DEST, I re-add my PLANT-B hosts, re-enable vsan and adjust my storage policies within vCD vm's.

But again sorry...all this because of this retarded test that refuses to vmotion only because all my hosts are not able to see my datastores ? I tested, putting other hosts in maintenance or even disabling them within vCD > same result.

I mean this is not acceptable, we are talking about service provider software... Building an overlay on top of vCenter OK but it has to give, at least, the very basic functionality everyone is used to have for over a decade !

I consider that not allowing a vMotion to happen on the only basis that only 1 destination host has the same datastore layout is non-sense. Or please someone explain me why and who VMware is to tell me what I should do with MY environment.

If I take the decision to move for, a very restricted period, my environment via 1 single host, please just let it be MY decision.

I deeply hate this "I am the one telling you how you do things" approach. Reminds me Microsoft back in the years

Seriously VMware ? :smileyangry:

[thgreyprnc]

Ok so now my 3 hosts running on plant-A are equiped with FC adapters.

I disabled vSAN and removed plant-B hosts.

This means that my 3 destinatino hosts are now rigorously in the exact same network/storage layout as my source hosts.

Guess what ? The migration initiated from vCD fails with the exact same ******* error message :smileyangry:

This has to be a bad joke ! What a piece of SHIT !!!

[IamTHEvilONE]

based on the original post, there is no CPU or Memory allotment to the org or provider vDC in the destination.

Are you attempting to use the:

1. Migrate feature under manage and monitor > resource pools > right click the VM and migrate

or

2. Go into vCenter and vMotion

The first one has a series of pre-checks to ensure that the destination Org vDC has enough compute resources available.  Option 2 just does the vCenter side and not really optimal when using vCloud Director.

The screen capture for HCC shows 0 total CPU and Memory Reservations available, which means you can't put anything into there with a requirement on Memory or CPU.  However, I think this screen capture is from vCenter and not from vCloud Director.  If the HCC resource pool represents your Provider vDC in the destination, then you failed to set this up correctly.  you need to set the resources to have a minimum = maximum = some value ... otherwise there is no CPU or Memory to pull from.

Can you set min = 50GHz, max = 50 GHz for CPU and min = 50 GB, max = 50 GB for RAM in the vCenter side?  (or some non-zero value)  Just compare your source and destination between the two.

Can you provide a screen capture of the vCloud Director Allocations page of the source and destination Org vDC you are attempting to migrate to and from?