"Bypass VLAN id/range overlap" argument missing?

[Yordan Kostov]

Dear everyone,

I would like to use "Bypass VLAN id/range overlap" to create 2 
shared networks that span across one VLAN.
Looking in "cloudmonkey create network" help menu I cannot see 
such argument.
Is there something I am missing or it is not implemented?

Parameters
==
startip = (string) the beginning IP address in the network IP range
externalid = (string) ID of the network in an external system.
networkdomain = (string) network domain
account = (string) account that will own the network
isolatedpvlantype = (string) the isolated private VLAN type for this network
ip6cidr = (string) the CIDR of IPv6 network, must be at least /64
displaytext = (string) the display text of the network
acltype = (string) Access control type; supported values are account and 
domain. In 3.0 all shared networks should have aclType=Domain, and all isolated 
networks - Account. Account means that only the account owner can use the 
network, domain - all accounts in the domain can use the network
zoneid = (uuid) the zone ID for the network
name = (string) the name of the network
netmask = (string) the netmask of the network. Required for shared networks and 
isolated networks when it belongs to VPC
endip = (string) the ending IP address in the network IP range. If not 
specified, will be defaulted to startIP
networkofferingid = (uuid) the network offering ID
displaynetwork = (boolean) an optional field, whether to the display the 
network to the end user or not.
aclid = (uuid) Network ACL ID associated for the network
endipv6 = (string) the ending IPv6 address in the IPv6 network range
ip6gateway = (string) the gateway of the IPv6 network. Required for Shared 
networks
gateway = (string) the gateway of the network. Required for shared networks and 
isolated networks when it belongs to VPC
domainid = (uuid) domain ID of the account owning a network
startipv6 = (string) the beginning IPv6 address in the IPv6 network range
subdomainaccess = (boolean) Defines whether to allow subdomains to use networks 
dedicated to their parent domain(s). Should be used with aclType=Domain, 
defaulted to allow.subdomain.network.access global config if not specified
projectid = (uuid) an optional project for the SSH key
physicalnetworkid = (uuid) the physical network ID the network belongs to
vpcid = (uuid) the VPC network belongs to
isolatedpvlan = (string) the isolated private VLAN for this network

Regards,
Jordan

Re: Routers on specific cluster

[Vivek Kumar]

Sure Andrija, I get you point. Will do the rest of the testing.


> On 02-Jun-2021, at 1:45 AM, Andrija Panic  wrote:
> 
> Just to add on what Abhishek said - IF you configure "
> system.vm.default.hypervisor with a preferred value, I don't think that
> tags will be considered at all (or might only be considered within the
> chosen hypervisor's clusters - but I think they are not... based on a
> recent case I've seen)
> 
> So you might want to NOT specify this setting, and then try to use tags on
> Service Offering for the VR
> 
> On Mon, 31 May 2021 at 04:27, Harikrishna Patnala <
> harikrishna.patn...@shapeblue.com> wrote:
> 
>> Hi Vivek,
>> 
>> I could found the global setting "system.vm.default.hypervisor" which is
>> considered while deploying the router, you may check that. Host tags in the
>> system service offering should also work, can you go through the logs while
>> the virtual router is getting deployed to observe which pod/cluster/host is
>> picked while allocating resources.
>> 
>> Regards,
>> Harikrishna
>> 
>> Harikrishna Patnala
>> Software Engineer
>> harikrishna.patn...@shapeblue.com
>> www.shapeblue.com
>> 
>> 
>> 
>> 
>> 
>> 
>> --
>> *From:* Vivek Kumar 
>> *Sent:* Friday, May 28, 2021 6:19 PM
>> *To:* users@cloudstack.apache.org 
>> *Subject:* Re: Routers on specific cluster
>> 
>> Hey Andrija,
>> 
>> Is there any setting in global setting where  we define that VR should go
>> to any particular cluster, as you mentioned in previous thread. I know
>> there is no as such performance or functionality issue just curious to
>> know.
>> 
>> 
>> Regards,
>> Vivek Kumar
>> 
>>> On 28-May-2021, at 4:45 AM, Andrija Panic 
>> wrote:
>>> 
>>> host tags should work for all Service Offerings - unless you are trying
>> to
>>> create VR on VMware cluster, while your global setting is saying it
>> should
>>> always be KVM (and yes, you should keep it on KVM - much quicker to
>>> provision/configure VR than on VMware) - did you try setting host tags on
>>> the Service Offering for Virtual Router (or whatever offering you are
>> using
>>> for the VR) ?
>>> 
>>> On Thu, 27 May 2021 at 05:20, Vivek Kumar
>> 
>>> wrote:
>>> 
 Hello Guys,
 
 So I am using ACS 4.13 with Advance zone. I have 5 KVM cluster and 5
 VMware Cluster. So whenever I provision any VPC for any of the cluster,
>> VPC
 VR always go and sit on KVM hosts, we have around 200 VPC VR and all are
 there in KVM zone.  Well for functionality prospective I don’t have any
 issues but is there any way to provision router in specific cluster as
 well.
 
 
 Storage tag and host tag doesn’t work in case of VR’s offering.
 
 
 Regards
 Vivek Kumar
>>> 
>>> 
>>> 
>>> --
>>> 
>>> Andrija Panić
>> 
>> 
> 
> -- 
> 
> Andrija Panić

Re: Issue with live migrations

[Jeremy Hansen]

Thank you for your help.  Set up a tagged VLAN and live migrations are working 
like a champ now.

My next issue…

I launched a second guest and noticed it did not allocate a new public IP.  My 
expectation was that it would allocate another public IP and the second 
instance would have its own virtual router/firewall/port forwarding, etc, but 
that doesn’t seem to be the case.  I can configured the firewall on the 
existing virtual router to forward to the second instance, but I’d prefer it 
just allocate another public IP from the range and allow me to configure each 
instance as a separate entity.  Is this possible?

Thanks
-jeremy


> On Jun 1, 2021, at 4:32 AM, Alessandro Caviglione  
> wrote:
> 
> No, but if you're using Advanced Networking you've to manage the trunk
> configuration on the switch ports.
> Every Virtual Router has a public IP on a specific interface and a specific
> VLAN and a private IP on another interface in another vlan that acts as a
> gateway for the instance.
> So, if VR and instance are in the same host, packets do not exit from host
> and instance are reachable, but if they run on different hosts the packets
> must go through a switch that should be configured to allow traffic on that
> private VLAN.
> So, in fact, you can allow a single switch port to forward traffic on
> multiple VLAN, this is called Trunk.
> 
> On Tue, Jun 1, 2021 at 12:31 PM Jeremy Hansen  wrote:
> 
>> Ahh, so I just noticed that I could also migrate the router and when I do
>> that, everything works as expected again.  So is this how migrates work?
>> Do routers always have to move with the instance?
>> 
>> -jeremy
>> 
>> 
>> On Jun 1, 2021, at 3:26 AM, Jeremy Hansen  wrote:
>> 
>> One thing I noticed, and I don’t know if this is expected, but the virtual
>> router my instance is using remains on the original VM host.  If I log in
>> to the guest instance while it’s on the second VM host, I can no longer
>> ping the private net gateway:
>> 
>> 
>> 
>> and the virtual router can’t ping the private IP
>> 
>> 
>> 
>> If I migrate the instance back, it ping reachable again:
>> 
>> 
>> 
>> 
>> Thanks
>> -jeremy
>> 
>> 
>> 
>> On Jun 1, 2021, at 3:07 AM, Jeremy Hansen  wrote:
>> 
>> Hmm. VM hosts are on the same vlan. Port configuration on both hosts is
>> exactly the same. I’m not saying it isn’t vlan related but I would expect
>> neither host to work if it was a vlan issue?
>> 
>> Thanks
>> 
>> On Jun 1, 2021, at 2:47 AM, Alessandro Caviglione 
>> wrote:
>> 
>> I think that your switch doesn't have ports tagged with all required
>> VLANs...
>> 
>> On Tue, Jun 1, 2021 at 11:20 AM Jeremy Hansen  wrote:
>> 
>> 
>> I set up another VM host so now I have two VM hosts.  I’m experimenting
>> with live migrations.  It looks like the VM migrates, but I can no longer
>> SSH to the guest once it’s migrated.  I’m using an isolated network.  I
>> setup firewall rules to allow all, along with port forwarding rules for
>> ssh.  I’m using the CentOS 5.5 template.  When I migrate to the second VM
>> host, ping/icmp still works, but I’m unable to ssh to the VM guest.  If I
>> migrate back to the original VM host, everything is fine again.  Any clues
>> what I’m doing wrong on the second VM host or why rules would change on the
>> migration?
>> 
>> Thanks
>> -jeremy
>> 
>> 
>> 
>> 
>> 
>> 



signature.asc
Description: Message signed with OpenPGP

Re: VMware Instance Error

[Abhishek Kumar]

Hi Mike,

For your deployment/disk controller error, you should check the template you 
are using for the VM. From the logs you have shared there is something wrong 
with the StartCommand.

---
org.apache.cloudstack.storage.to.TemplateObjectTO":{"id":"0","format":"ISO","accountId":"0","hvm":"false","bootable":"false"
---
Not sure why id is 0 there and the format is ISO. Are you using an ISO to 
deploy the VM?

Regards,
Abhishek

From: Andrija Panic 
Sent: 02 June 2021 01:38
To: users 
Subject: Re: VMware Instance Error

Mike,

are you in a position to share that specific template (if licencing allows,
if it's trial, etc) - ping me on email directly - and I can give it a spin
in the lab on 4.15
4.15.1 should bring this new VMware template behaviour (read from OVF,
instead of allowing you to specify OS, NIC/disk controllers, etc) - AS
OPTIONAL - so you can wait for 4.15.1 if you use non-standard OVA
appliances (or, for that matter, the ones that simply don't work with ACS,
for whatever reasons)

Best,


 

On Tue, 1 Jun 2021 at 21:46, Corey, Mike  wrote:

> Another development in my troubleshooting this...
>
> I used a vendor (OEM) supplied ova appliance as a template and downloaded
> in the same manner.  NO SETTINGS were created for the template by ACS.
>
> If the expected behavior is for ACS to add settings based off of the OVF
> descriptor - my case IS NOT.
>
> I could really use some guidance here on troubleshooting as to why the
> template settings are NOT being created by ACS.
>
> Again - anyone running ACS 4.15 with VMware and having success with using
> templates?
>
> Any other logs I can check to determine why ACS is not creating these
> custom setting attributes on the templates?
>
> Many thanks!
>
> Mike
>
> -Original Message-
> From: Corey, Mike 
> Sent: Tuesday, June 1, 2021 8:47 AM
> To: users@cloudstack.apache.org
> Subject: [CAUTION] RE: VMware Instance Error
>
> Thanks for the reply Abhishek!
>
> 1- ACS deploys the VM from template; however, it fails and shows Error in
> ACS-Instances UI.  So the VM remains in vCenter - that VM has the CPU/RAM
> as the template NOT what the ACS service offering (Small Instance) I
> selected.  I suppose ACS is not finishing the configuration because the
> deployment fails.
>
> 3 - This is the issue I need help to troubleshoot.  ACS IS NOT creating
> any settings on the Windows 10 and CentOS template I've uploaded into ACS.
> Both template OVF files have the hardware descriptions used by the template
> VM.  However, NO SETTINGS are being added to the ACS template.
>
> How can this be troubleshot?  Obviously, ACS is NOT populating the
> settings on the template as expected - what could be the issue?  What
> commands occur when the template is downloaded and populate the settings of
> the template (so I can search for errors)?  Is something supposed to be
> written to ACS but is not due to permissions or file doesn't exist?
>
> Item 1 & 4 - Yes, the VM in vCenter is named the same as the ones that
> fail deployment in ACS (i-2-##-vm) - the VM has resources of the template
> (not service offering) and powers ON and loads OS.
>
> 5 - snippet from logs below - diskcontroller is being identified by OVF -
> nothing is set by ACS on template.
>
> INFO  [c.c.h.v.r.VmwareResource] (DirectAgent-496:ctx-f077199e
> usphlmvesxt01.phl.global.corp.sap, job-164/job-166, cmd: StartCommand)
> (logid:6c707559) Executing resource StartCommand:
> {"vm":{"id":"26","name":"i-2-26-VM","bootloader":"HVM","state":"Starting","type":"User","cpus":"1","minSpeed":"166","maxSpeed":"500","minRam":"(512.00
> MB) 536870912","maxRam":"(512.00 MB)
> 536870912","hostName":"June1-01","arch":"x86_64","os":"Windows 10
> (64-bit)","platformEmulator":"windows9_64Guest","bootArgs":"","enableHA":"false","limitCpuUse":"false","enableDynamicallyScaleVm":"false","vncPassword":"X1k9AVje8WIPzXb_8P_88Q","params":{"deployvm":"true","dataDiskController":"scsi","memoryOvercommitRatio":"1.0","Message.ReservedCapacityFreed.Flag":"false","nestedVirtualizationFlag":"false","cpuOvercommitRatio":"3.0","vmware.reserve.mem":"false","vmware.reserve.cpu":"false","nicAdapter":"E1000","rootDiskController":"lsilogicsas"},"uuid":"dafd18a5-04a9-4bda-b701-0d24e4b1e61d","enterHardwareSetup":"false","disks":[{"data":{"org.apache.
> cloudstack.storage.to
> .VolumeObjectTO":{"uuid":"973a1795-4569-4769-afcb-93e3911be12e","volumeType":"ROOT","dataStore":{"org.
> apache.cloudstack.storage.to.PrimaryDataStoreTO":{"uuid":"9df43ec0-1e52-3cc3-8e5a-bf2313bbbd8d","id":"1","poolType":"NetworkFilesystem","host":"usphlcv302.phl.global.corp.sap","path":"/usphlmvesx_lab_cspri01","port":"2049","url":"NetworkFilesystem:///usphlmvesx_lab_cspri01/?ROLE=Primary=9df43ec0-1e52-3cc3-8e5a-bf2313bbbd8d","isManaged":"false"}},"name":"ROOT-26-0","size":"(80.00
> GB)
> 

Re: Routers on specific cluster

[Andrija Panic]

Just to add on what Abhishek said - IF you configure "
system.vm.default.hypervisor with a preferred value, I don't think that
tags will be considered at all (or might only be considered within the
chosen hypervisor's clusters - but I think they are not... based on a
recent case I've seen)

So you might want to NOT specify this setting, and then try to use tags on
Service Offering for the VR

On Mon, 31 May 2021 at 04:27, Harikrishna Patnala <
harikrishna.patn...@shapeblue.com> wrote:

> Hi Vivek,
>
> I could found the global setting "system.vm.default.hypervisor" which is
> considered while deploying the router, you may check that. Host tags in the
> system service offering should also work, can you go through the logs while
> the virtual router is getting deployed to observe which pod/cluster/host is
> picked while allocating resources.
>
> Regards,
> Harikrishna
>
> Harikrishna Patnala
> Software Engineer
> harikrishna.patn...@shapeblue.com
> www.shapeblue.com
>
>
>
>
>
>
> --
> *From:* Vivek Kumar 
> *Sent:* Friday, May 28, 2021 6:19 PM
> *To:* users@cloudstack.apache.org 
> *Subject:* Re: Routers on specific cluster
>
> Hey Andrija,
>
> Is there any setting in global setting where  we define that VR should go
> to any particular cluster, as you mentioned in previous thread. I know
> there is no as such performance or functionality issue just curious to
> know.
>
>
> Regards,
> Vivek Kumar
>
> > On 28-May-2021, at 4:45 AM, Andrija Panic 
> wrote:
> >
> > host tags should work for all Service Offerings - unless you are trying
> to
> > create VR on VMware cluster, while your global setting is saying it
> should
> > always be KVM (and yes, you should keep it on KVM - much quicker to
> > provision/configure VR than on VMware) - did you try setting host tags on
> > the Service Offering for Virtual Router (or whatever offering you are
> using
> > for the VR) ?
> >
> > On Thu, 27 May 2021 at 05:20, Vivek Kumar
> 
> > wrote:
> >
> >> Hello Guys,
> >>
> >> So I am using ACS 4.13 with Advance zone. I have 5 KVM cluster and 5
> >> VMware Cluster. So whenever I provision any VPC for any of the cluster,
> VPC
> >> VR always go and sit on KVM hosts, we have around 200 VPC VR and all are
> >> there in KVM zone.  Well for functionality prospective I don’t have any
> >> issues but is there any way to provision router in specific cluster as
> >> well.
> >>
> >>
> >> Storage tag and host tag doesn’t work in case of VR’s offering.
> >>
> >>
> >> Regards
> >> Vivek Kumar
> >
> >
> >
> > --
> >
> > Andrija Panić
>
>

-- 

Andrija Panić

Re: Prediction or studies for KVM live migration.

[Andrija Panic]

..as I assumed - and it exports volume to Secondary Storage as the
intermediate location - all good.
So "live storage migration" doesn't work, as expected.

cheers,

On Fri, 28 May 2021 at 14:11, Kalil de Albuquerque Carvalho <
kalil.carva...@hybriddc.com.br> wrote:

> Hello Andrija.
>
> Sorry for the mess. Instance migration, just about VM's, memory, vCPU
> works fine, on the fly. But if is needed to migrate a volume from one
> primary storage to other is needed the VM must be halted.
>
> I had tested of migration from a Gluster primary Storage to a NFS primary
> storage, and contrariwise, and worked well but just with powerd off VM's.
>
> Best regars.
> Em 27/05/2021 20:18, Andrija Panic escreveu:
>
> I understood you said that LIVE storage migration (migrate VM's volumes
> (with the VM)) works while VM is RUNNING.
>
> Are you now saying this is NOT working (which is what I would expect), and
> that only stopped VM migration is possible from GLuster to NFS?
>
> best,
>
> On Wed, 26 May 2021 at 22:04, Kalil de Albuquerque Carvalho <
> kalil.carva...@hybriddc.com.br> wrote:
>
>> Hello Andrija.
>>
>> Gluster as primary storage works fine. Storage migration, with powered
>> off VM's, it's working  too. My problem is just doing this with VM's
>> running.
>>
>> I'm using Ubuntu 20.02, CentOS 7 and Windows 10 and 2016 for testing and
>> not working.
>>
>> But it is the life, thaks all
>>
>> Best regards,
>>
>>
>> Em 26/05/2021 16:41, Andrija Panic escreveu:
>>
>> I thought I replied to this one, but I dont' see my email...
>>
>> So, from CEPH/NFS to SolidFire should work (in this direction only)  - or
>> let me say "used to work" (haven't tested it recently) - this was developed
>> for my ex-company where I use to work, by Mike Tutkowski from NetApp)
>> Also, my understanding is that it's also possible to migrate VMs using
>> local storage from host to host (whole VM with its disks) - @Gabriel
>> Beims Bräscher  can confirm this, afaik?
>>
>> If you are using Ubuntu - all fine - qemu-kvm supports live storage
>> migrations from Ubuntu 14.04 at least, an onwards.
>> If you are using CentOS 7, you have to use qemu-kvm-ev from the oVirt
>> repo ONLY - all other versions of qemu-kvm do NOT support storage live
>> migration (Redhat revoked it for $$$ reasons, while it was working fine in
>> CentOS6)
>>
>> If your tested it and it worked from Gluster to NFS - that's (great) news
>> (for me).
>>
>> Hope that helps,
>>
>> Cheers,
>>
>> On Wed, 26 May 2021 at 20:06, Wido den Hollander  wrote:
>>
>>>
>>>
>>> On 26/05/2021 13:55, Kalil de Albuquerque Carvalho wrote:
>>> > Hello Wido.
>>> >
>>> > Sorry about that. I was not so clear, or made some misunderstanding.
>>> >
>>> > Doing some corrections, I've tested migration from Gluster to NFS, and
>>> > the reverse, and every think worked well. So, please, disregard this
>>> > part of my question. I should would did this test before made the
>>> question.
>>> >
>>> > My question, now, is when will be support, if will be, with the
>>> running
>>> > VM's. Today, I'm testing the version 4.15, just working with power off
>>> > VM's.
>>> >
>>> Aha, you mean live storage migration between different types of primary
>>> storage.
>>>
>>> That is indeed not supported with KVM and also not on the roadmap at the
>>> moment.
>>>
>>> Wido
>>>
>>> > Best regards.
>>> >
>>> > Em 26/05/2021 04:08, Wido den Hollander escreveu:
>>> >>
>>> >>
>>> >> On 25/05/2021 13:32, Kalil de Albuquerque Carvalho wrote:
>>> >>> Hello all.
>>> >>>
>>> >>> Reading the manual I discovery that live migration is not support
>>> for
>>> >>> KVM hypervisor. I was wander if there are studies or predictions for
>>> >>> this features on KVM hosts.
>>> >>>
>>> >>
>>> >> Where did you read this? Live Migration with the KVM hypervisor works
>>> >> just fine.
>>> >>
>>> >> Wido
>>> >>
>>> >>> Yet on the manual citation, it said that migration just can occur
>>> >>> from CEPH/NFS to "SolidFire Managed Storage". On my tests we are
>>> >>> using Gluster as Primary Storage and not appear any storage to
>>> >>> migrate to. We created tow differents Primary Storages for this kind
>>> >>> of tests. Is that correct, migration in this case just will occur
>>> >>> from/to CEPH/NFS? If yes, will be same future release that will be
>>> >>> possible migration between Guster storages?
>>> >>>
>>> >>> Best regars.
>>> >>>
>>> >>
>>>
>>
>>
>> --
>>
>> Andrija Panić
>>
>>
>
> --
>
> Andrija Panić
>
>

-- 

Andrija Panić

Re: VMware Instance Error

[Andrija Panic]

Mike,

are you in a position to share that specific template (if licencing allows,
if it's trial, etc) - ping me on email directly - and I can give it a spin
in the lab on 4.15
4.15.1 should bring this new VMware template behaviour (read from OVF,
instead of allowing you to specify OS, NIC/disk controllers, etc) - AS
OPTIONAL - so you can wait for 4.15.1 if you use non-standard OVA
appliances (or, for that matter, the ones that simply don't work with ACS,
for whatever reasons)

Best,

On Tue, 1 Jun 2021 at 21:46, Corey, Mike  wrote:

> Another development in my troubleshooting this...
>
> I used a vendor (OEM) supplied ova appliance as a template and downloaded
> in the same manner.  NO SETTINGS were created for the template by ACS.
>
> If the expected behavior is for ACS to add settings based off of the OVF
> descriptor - my case IS NOT.
>
> I could really use some guidance here on troubleshooting as to why the
> template settings are NOT being created by ACS.
>
> Again - anyone running ACS 4.15 with VMware and having success with using
> templates?
>
> Any other logs I can check to determine why ACS is not creating these
> custom setting attributes on the templates?
>
> Many thanks!
>
> Mike
>
> -Original Message-
> From: Corey, Mike 
> Sent: Tuesday, June 1, 2021 8:47 AM
> To: users@cloudstack.apache.org
> Subject: [CAUTION] RE: VMware Instance Error
>
> Thanks for the reply Abhishek!
>
> 1- ACS deploys the VM from template; however, it fails and shows Error in
> ACS-Instances UI.  So the VM remains in vCenter - that VM has the CPU/RAM
> as the template NOT what the ACS service offering (Small Instance) I
> selected.  I suppose ACS is not finishing the configuration because the
> deployment fails.
>
> 3 - This is the issue I need help to troubleshoot.  ACS IS NOT creating
> any settings on the Windows 10 and CentOS template I've uploaded into ACS.
> Both template OVF files have the hardware descriptions used by the template
> VM.  However, NO SETTINGS are being added to the ACS template.
>
> How can this be troubleshot?  Obviously, ACS is NOT populating the
> settings on the template as expected - what could be the issue?  What
> commands occur when the template is downloaded and populate the settings of
> the template (so I can search for errors)?  Is something supposed to be
> written to ACS but is not due to permissions or file doesn't exist?
>
> Item 1 & 4 - Yes, the VM in vCenter is named the same as the ones that
> fail deployment in ACS (i-2-##-vm) - the VM has resources of the template
> (not service offering) and powers ON and loads OS.
>
> 5 - snippet from logs below - diskcontroller is being identified by OVF -
> nothing is set by ACS on template.
>
> INFO  [c.c.h.v.r.VmwareResource] (DirectAgent-496:ctx-f077199e
> usphlmvesxt01.phl.global.corp.sap, job-164/job-166, cmd: StartCommand)
> (logid:6c707559) Executing resource StartCommand:
> {"vm":{"id":"26","name":"i-2-26-VM","bootloader":"HVM","state":"Starting","type":"User","cpus":"1","minSpeed":"166","maxSpeed":"500","minRam":"(512.00
> MB) 536870912","maxRam":"(512.00 MB)
> 536870912","hostName":"June1-01","arch":"x86_64","os":"Windows 10
> (64-bit)","platformEmulator":"windows9_64Guest","bootArgs":"","enableHA":"false","limitCpuUse":"false","enableDynamicallyScaleVm":"false","vncPassword":"X1k9AVje8WIPzXb_8P_88Q","params":{"deployvm":"true","dataDiskController":"scsi","memoryOvercommitRatio":"1.0","Message.ReservedCapacityFreed.Flag":"false","nestedVirtualizationFlag":"false","cpuOvercommitRatio":"3.0","vmware.reserve.mem":"false","vmware.reserve.cpu":"false","nicAdapter":"E1000","rootDiskController":"lsilogicsas"},"uuid":"dafd18a5-04a9-4bda-b701-0d24e4b1e61d","enterHardwareSetup":"false","disks":[{"data":{"org.apache.
> cloudstack.storage.to
> .VolumeObjectTO":{"uuid":"973a1795-4569-4769-afcb-93e3911be12e","volumeType":"ROOT","dataStore":{"org.
> apache.cloudstack.storage.to.PrimaryDataStoreTO":{"uuid":"9df43ec0-1e52-3cc3-8e5a-bf2313bbbd8d","id":"1","poolType":"NetworkFilesystem","host":"usphlcv302.phl.global.corp.sap","path":"/usphlmvesx_lab_cspri01","port":"2049","url":"NetworkFilesystem:///usphlmvesx_lab_cspri01/?ROLE=Primary=9df43ec0-1e52-3cc3-8e5a-bf2313bbbd8d","isManaged":"false"}},"name":"ROOT-26-0","size":"(80.00
> GB)
> 85899345920","volumeId":"17","vmName":"i-2-26-VM","accountId":"2","format":"OVA","provisioningType":"THIN","id":"17","deviceId":"0","bytesReadRate":"(0
> bytes) 0","bytesWriteRate":"(0 bytes) 0","iopsReadRate":"(0 bytes)
> 0","iopsWriteRate":"(0 bytes)
> 0","hypervisorType":"VMware","directDownload":"false","deployAsIs":"true"}},"diskSeq":"0","type":"ROOT","_details":{"storageHost":"usphlcv302.phl.global.corp.sap","managed":"false","storagePort":"2049","volumeSize":"(80.00
> GB) 
> 

RE: VMware Instance Error

[Corey, Mike]

Another development in my troubleshooting this...

I used a vendor (OEM) supplied ova appliance as a template and downloaded in 
the same manner.  NO SETTINGS were created for the template by ACS.  

If the expected behavior is for ACS to add settings based off of the OVF 
descriptor - my case IS NOT.  

I could really use some guidance here on troubleshooting as to why the template 
settings are NOT being created by ACS.

Again - anyone running ACS 4.15 with VMware and having success with using 
templates?

Any other logs I can check to determine why ACS is not creating these custom 
setting attributes on the templates?

Many thanks!

Mike 

-Original Message-
From: Corey, Mike  
Sent: Tuesday, June 1, 2021 8:47 AM
To: users@cloudstack.apache.org
Subject: [CAUTION] RE: VMware Instance Error

Thanks for the reply Abhishek!

1- ACS deploys the VM from template; however, it fails and shows Error in 
ACS-Instances UI.  So the VM remains in vCenter - that VM has the CPU/RAM as 
the template NOT what the ACS service offering (Small Instance) I selected.  I 
suppose ACS is not finishing the configuration because the deployment fails.

3 - This is the issue I need help to troubleshoot.  ACS IS NOT creating any 
settings on the Windows 10 and CentOS template I've uploaded into ACS.  Both 
template OVF files have the hardware descriptions used by the template VM.  
However, NO SETTINGS are being added to the ACS template.  

How can this be troubleshot?  Obviously, ACS is NOT populating the settings on 
the template as expected - what could be the issue?  What commands occur when 
the template is downloaded and populate the settings of the template (so I can 
search for errors)?  Is something supposed to be written to ACS but is not due 
to permissions or file doesn't exist?

Item 1 & 4 - Yes, the VM in vCenter is named the same as the ones that fail 
deployment in ACS (i-2-##-vm) - the VM has resources of the template (not 
service offering) and powers ON and loads OS.   

5 - snippet from logs below - diskcontroller is being identified by OVF - 
nothing is set by ACS on template.

INFO  [c.c.h.v.r.VmwareResource] (DirectAgent-496:ctx-f077199e 
usphlmvesxt01.phl.global.corp.sap, job-164/job-166, cmd: StartCommand) 
(logid:6c707559) Executing resource StartCommand: 
{"vm":{"id":"26","name":"i-2-26-VM","bootloader":"HVM","state":"Starting","type":"User","cpus":"1","minSpeed":"166","maxSpeed":"500","minRam":"(512.00
 MB) 536870912","maxRam":"(512.00 MB) 
536870912","hostName":"June1-01","arch":"x86_64","os":"Windows 10 
(64-bit)","platformEmulator":"windows9_64Guest","bootArgs":"","enableHA":"false","limitCpuUse":"false","enableDynamicallyScaleVm":"false","vncPassword":"X1k9AVje8WIPzXb_8P_88Q","params":{"deployvm":"true","dataDiskController":"scsi","memoryOvercommitRatio":"1.0","Message.ReservedCapacityFreed.Flag":"false","nestedVirtualizationFlag":"false","cpuOvercommitRatio":"3.0","vmware.reserve.mem":"false","vmware.reserve.cpu":"false","nicAdapter":"E1000","rootDiskController":"lsilogicsas"},"uuid":"dafd18a5-04a9-4bda-b701-0d24e4b1e61d","enterHardwareSetup":"false","disks":[{"data":{"org.apache.cloudstack.storage.to.VolumeObjectTO":{"uuid":"973a1795-4569-4769-afcb-93e3911be12e","volumeType":"ROOT","dataStore":{"org.apache.cloudstack.storage.to.PrimaryDataStoreTO":{"uuid":"9df43ec0-1e52-3cc3-8e5a-bf2313bbbd8d","id":"1","poolType":"NetworkFilesystem","host":"usphlcv302.phl.global.corp.sap","path":"/usphlmvesx_lab_cspri01","port":"2049","url":"NetworkFilesystem:///usphlmvesx_lab_cspri01/?ROLE=Primary=9df43ec0-1e52-3cc3-8e5a-bf2313bbbd8d","isManaged":"false"}},"name":"ROOT-26-0","size":"(80.00
 GB) 
85899345920","volumeId":"17","vmName":"i-2-26-VM","accountId":"2","format":"OVA","provisioningType":"THIN","id":"17","deviceId":"0","bytesReadRate":"(0
 bytes) 0","bytesWriteRate":"(0 bytes) 0","iopsReadRate":"(0 bytes) 
0","iopsWriteRate":"(0 bytes) 
0","hypervisorType":"VMware","directDownload":"false","deployAsIs":"true"}},"diskSeq":"0","type":"ROOT","_details":{"storageHost":"usphlcv302.phl.global.corp.sap","managed":"false","storagePort":"2049","volumeSize":"(80.00
 GB) 
85899345920"}},{"data":{"org.apache.cloudstack.storage.to.TemplateObjectTO":{"id":"0","format":"ISO","accountId":"0","hvm":"false","bootable":"false","directDownload":"false","deployAsIs":"false"}},"diskSeq":"3","type":"ISO"}],"nics":[{"deviceId":"0","networkRateMbps":"200","defaultNic":"true","pxeDisable":"false","nicUuid":"7587229e-bbb3-4e65-b093-fa9ff690bb43","details":{"MacAddressChanges":"true","ForgedTransmits":"true","PromiscuousMode":"false"},"dpdkEnabled":"false","uuid":"62bfbed1-d07c-4886-b4f4-bf594543bb28","ip":"10.4.80.238","netmask":"255.255.240.0","gateway":"10.4.80.1","mac":"1e:00:9d:00:00:4d","dns1":,"dns2":"","broadcastType":"Vlan","type":"Guest","broadcastUri":"vlan://
 ","isolationUri":"vlan:// 

Re: 4.15.1 Snapshot UEFI Legacy: Unable to start VM due to Unsupported Configuration: IDE controllers are unsupported for this QEMU binary or machine type

[Udo Müller]

Hi Slavka,

This ISO is just the iPXE template because we deploy our VM automatically.

Since we deploy Windows and Linux i have to upload it twice? Once categorized 
with Linux, the other time with Windows?

But since i want to have the same behaviour as with our old cloudstack, which 
is /dev/sdX naming for the root device, Windows is the way to go, correct?

I will try it out. Thx for your help!

But cloudstack has to be changed in order to get a vm created with UEFI boot 
and an other OS than the named ones.

At least an error must be shown when using UEFI boot with a template 
categorized Other and deployed on KVM.

Gruß Udo Müller

> Am 01.06.2021 um 19:01 schrieb Slavka Peleva :
> 
> Hi Udo,
> 
> I saw where is the problem, the OS type of your ISO is "Other (32 bit)" on
> which CloudStack sets "ide" for bus.
> 'sata' is set on Windows OS types;
> 'virtio' on these -  Ubuntu, Fedora ,CentOS,Red Hat Enterprise Linux,Debian
> GNU/Linux,FreeBSD,Oracle,Other PV;
> 'scsi' on "Other PV Virtio-SCSI" OS types;
> the rest gets 'ide' for bus
> 
> Regards,
> Slavka
> 
>> On Tue, Jun 1, 2021 at 6:24 PM Udo Müller  wrote:
>> 
>> Hi Slavka,
>> 
>> Deployment is done via WebUI. In this case i choose the target host
>> (since only this one has debug enabled), then choose service, disk and
>> network offering, activate advanced boot options and check UEFI Legacy.
>> 
>> After setting the name the vm is started.
>> 
>> The XML of the VM is attached.
>> 
>> There you see
>> 
>> 
>> 
>> But the VM will work with sata instead of ide. I tested it by defining a
>> VM via virt-manager and choosing sata for the bus type.
>> 
>> See also:
>> 
>> https://bugs.launchpad.net/nova/+bug/1831538
>> 
>> Regards Udo
>> 
>> 
>>> Am 01.06.21 um 16:16 schrieb Slavka Peleva:
>>> Hi Udo,
>>> 
>>> Could you please share more information about the deployment of a VM?
>>> Are you creating a new instance from ISO or a template? Also, can you set
>>> the DEBUG level at the agent, and is it possible to share the XML that
>>> CloudStack sends to libvirt (it's in the
>>> /var/log/cloudstack/agent/agent.log)?
>>> 
>>> One more question did you update and the management with the 4.15.1.0
>>> version?
>>> 
>>> Kind regards,
>>> Slavka
>>> 
>>>> On Tue, Jun 1, 2021 at 4:33 PM Udo Müller  wrote:
>>> 
>>>> Hi guys,
>>>> 
>>>> 
>>>> I played around with cloudstack 4.15 and UEFI boot until someone here
>>>> pointed me to a PR which fixes the problem with the missing q35 chipset.
>>>> 
>>>> Thats the reason why I just installed the 4.15.1 SNAPSHOT.
>>>> 
>>>> Now I get the following problem:
>>>> 
>>>> Unable to start VM on Host... due to Unsupported Configuration: IDE
>>>> controllers
>>>> are unsupported for this QEMU binary or machine type
>>>> 
>>>> I am using cloudstack on CentOS 7.9 with
>>>> 
>>>> # rpm -qa | egrep "(qemu|libvirtd|OVMF|cloudstack)"
>>>> *cloudstack*-agent-4.15.1.0-SNAPSHOT.20210601.el7.x86_64
>>>> *qemu*-kvm-ev-2.12.0-44.1.el7_8.1.x86_64
>>>> ipxe-roms-*qemu*-20180825-3.git133f4c.el7.noarch
>>>> libvirt-daemon-driver-*qemu*-4.5.0-36.el7.x86_64
>>>> *OVMF*-20180508-6.gitee3198e672e2.el7.noarch
>>>> *qemu*-img-ev-2.12.0-44.1.el7_8.1.x86_64
>>>> *cloudstack*-common-4.15.1.0-SNAPSHOT.20210601.el7.x86_64
>>>> *qemu*-kvm-common-ev-2.12.0-44.1.el7_8.1.x86_64
>>>> 
>>>> 1) Is the PR regarding the q35 chipset already included in the SNAPSHOT?
>>>> I guess yes, because with 4.15.0 and i440 chipset the VM could be
>>>> created and was running.
>>>> 
>>>> 2) What do I have to change inside my settings to switch from IDE to
>>>> SCSI or SATA?
>>>> 
>>>> Regards Udo
>>>> 
>> 

Re: 4.15.1 SNAPSHOT: cant create users to account

[Abhishek Kumar]

Hi Udo,

This should fix the problem, https://github.com/apache/cloudstack/pull/5063


Regards,
Abhishek

From: Slavka Peleva 
Sent: 01 June 2021 21:41
To: users@cloudstack.apache.org 
Subject: Re: 4.15.1 SNAPSHOT: cant create users to account

Hi Udo,

I was able to reproduce the problem with the new UI. With the old UI, you
can create a user. You can open an issue on GitHub for the problem

Regards,
Slavka


 

On Tue, Jun 1, 2021 at 6:46 PM Udo Müller  wrote:

> Oops
>
>
> now the screenshots are attached.
>
> Regards udo
>
> Am 01.06.21 um 17:43 schrieb Udo Müller:
> > Oh and btw: The Add user-Popup is rendered wrong in Safari, but is
> > working fine in Chrome and Firefox.
> >
> > See attached screenshots.
> >
> > Is it possible for me to change anything to get this back to work?
> >
> > Regards Udo
> >
> >
> > Am 01.06.21 um 17:33 schrieb Udo Müller:
> >> Hi all,
> >>
> >> The following was working with 4.15.0.
> >>
> >>
> >> Status quo: 4.15.1 SNAPSHOT, database new initialized. Configuration
> >> done with the same settings as 4.15.0
> >>
> >>
> >> i am trying to add a user to the account admin. So i click on
> >>
> >> Accounts -> admin -> View Users -> Add User
> >>
> >> Popup Window opens fine.
> >>
> >>
> >> - When i enter all required fields (username, password 2x, email,
> >> first and last name) and hit ok, i get the error, that the parameter
> >> "account" is missing.
> >>
> >> - When i do like above and then click on the drop-down field Domain
> >> it is empty. Same for drop-down Account.
> >>
> >>
> >> Means i cant create users to accounts which is absolutely needed for
> >> my application.
> >>
> >>
> >> management.log
> >>
> >> 2021-06-01 15:31:27,334 DEBUG [c.c.a.ApiServlet]
> >> (qtp1630678941-3995:ctx-cff2deda) (logid:0a4b2d51) ===START===
> >> 192.168.7.244 -- POST  command=createUser=json
> >> 2021-06-01 15:31:27,339 DEBUG [c.c.a.ApiServer]
> >> (qtp1630678941-3995:ctx-cff2deda ctx-f77a5abe) (logid:0a4b2d51) CIDRs
> >> from which account 'Acct[a7e89cef-c2ca-11eb-8de2-525400f5bc6b-admin]'
> >> is allowed to perform API calls: 0.0.0.0/0,::/0
> >> 2021-06-01 15:31:27,341 INFO  [c.c.a.ApiServer]
> >> (qtp1630678941-3995:ctx-cff2deda ctx-f77a5abe) (logid:0a4b2d51)
> >> Unable to execute API command createuser due to missing parameter
> >> account
> >>
> >> apilog.log
> >>
> >> 2021-06-01 15:31:27,342 INFO  [a.c.c.a.ApiServer]
> >> (qtp1630678941-3995:ctx-cff2deda ctx-f77a5abe) (logid:0a4b2d51)
> >> (userId=2 accountId=2 sessionId=node0cacgcc2ki0zr1kjau5c1p59w511)
> >> 192.168.7.244 -- POST command=createUser=json 431 Unable to
> >> execute API command createuser due to missing parameter account
> >>
> >>
>

RE: hide network from lower domains?

[Yordan Kostov]

FYI I figured it out!

Cloudmonkey create network command has attribute subdomainaccess (Boolean) if 
set to true child domains will not see the network.
I just tested it and shared network on Root level is not seen by its subdomains!

Best regards,
Jordan

-Original Message-
From: Yordan Kostov  
Sent: Tuesday, June 1, 2021 8:10 PM
To: users@cloudstack.apache.org
Subject: RE: hide network from lower domains?


[X] This message came from outside your organization


Hey Dan,

The problem is that the network itself is managed by external firewall 
and not by the virtual Router.
The VR only provides DNS, DHCP and USERDATA services.

What I am working towards is - lets say we have 2 teams and 2 networks 
managed by external firewall.
Each team should be able to see and deploy VMs in their own network but 
not have access or visibility to the other team network.

Regards,
Jordan

-Original Message-
From: Daan Hoogland 
Sent: Tuesday, June 1, 2021 6:59 PM
To: users 
Subject: Re: hide network from lower domains?


[X] This message came from outside your organization


Jordan, is there a reason it must be a shared network? It seems to me you want 
an isolated network or a VPC.

On Tue, Jun 1, 2021 at 5:19 PM Yordan Kostov  wrote:

> Hey everyone,
>
> I am playing around with shared networks and I noticed 
> that when network is created for the ROOT user the lower level domains 
> see it too.
> Is there a way to hide that network from the lower 
> level domain users?
>
> Here is my setup:
>
>   *   Domains
>  *   ROOT
>  *   ROOT/DEVELOPERS
>   *   Accounts - there is a separate account ACC-DEV for ROOT/DEVELOPERS
> domain
>
> One thing I tried was to create a network with offering that is not 
> public but the result is the same.
>
> Best regards,
> Jordan
>


--
Daan

11!

RE: hide network from lower domains?

[Yordan Kostov]

Hey Dan,

The problem is that the network itself is managed by external firewall 
and not by the virtual Router.
The VR only provides DNS, DHCP and USERDATA services.

What I am working towards is - lets say we have 2 teams and 2 networks 
managed by external firewall.
Each team should be able to see and deploy VMs in their own network but 
not have access or visibility to the other team network.

Regards,
Jordan

-Original Message-
From: Daan Hoogland  
Sent: Tuesday, June 1, 2021 6:59 PM
To: users 
Subject: Re: hide network from lower domains?


[X] This message came from outside your organization


Jordan, is there a reason it must be a shared network? It seems to me you want 
an isolated network or a VPC.

On Tue, Jun 1, 2021 at 5:19 PM Yordan Kostov  wrote:

> Hey everyone,
>
> I am playing around with shared networks and I noticed 
> that when network is created for the ROOT user the lower level domains 
> see it too.
> Is there a way to hide that network from the lower 
> level domain users?
>
> Here is my setup:
>
>   *   Domains
>  *   ROOT
>  *   ROOT/DEVELOPERS
>   *   Accounts - there is a separate account ACC-DEV for ROOT/DEVELOPERS
> domain
>
> One thing I tried was to create a network with offering that is not 
> public but the result is the same.
>
> Best regards,
> Jordan
>


--
Daan

Re: 4.15.1 Snapshot UEFI Legacy: Unable to start VM due to Unsupported Configuration: IDE controllers are unsupported for this QEMU binary or machine type

[Slavka Peleva]

Hi Udo,

I saw where is the problem, the OS type of your ISO is "Other (32 bit)" on
which CloudStack sets "ide" for bus.
'sata' is set on Windows OS types;
'virtio' on these -  Ubuntu, Fedora ,CentOS,Red Hat Enterprise Linux,Debian
GNU/Linux,FreeBSD,Oracle,Other PV;
'scsi' on "Other PV Virtio-SCSI" OS types;
the rest gets 'ide' for bus

Regards,
Slavka

On Tue, Jun 1, 2021 at 6:24 PM Udo Müller  wrote:

> Hi Slavka,
>
> Deployment is done via WebUI. In this case i choose the target host
> (since only this one has debug enabled), then choose service, disk and
> network offering, activate advanced boot options and check UEFI Legacy.
>
> After setting the name the vm is started.
>
> The XML of the VM is attached.
>
> There you see
>
> 
>
> But the VM will work with sata instead of ide. I tested it by defining a
> VM via virt-manager and choosing sata for the bus type.
>
> See also:
>
> https://bugs.launchpad.net/nova/+bug/1831538
>
> Regards Udo
>
>
> Am 01.06.21 um 16:16 schrieb Slavka Peleva:
> > Hi Udo,
> >
> > Could you please share more information about the deployment of a VM?
> > Are you creating a new instance from ISO or a template? Also, can you set
> > the DEBUG level at the agent, and is it possible to share the XML that
> > CloudStack sends to libvirt (it's in the
> > /var/log/cloudstack/agent/agent.log)?
> >
> > One more question did you update and the management with the 4.15.1.0
> > version?
> >
> > Kind regards,
> > Slavka
> >
> > On Tue, Jun 1, 2021 at 4:33 PM Udo Müller  wrote:
> >
> >> Hi guys,
> >>
> >>
> >> I played around with cloudstack 4.15 and UEFI boot until someone here
> >> pointed me to a PR which fixes the problem with the missing q35 chipset.
> >>
> >> Thats the reason why I just installed the 4.15.1 SNAPSHOT.
> >>
> >> Now I get the following problem:
> >>
> >> Unable to start VM on Host... due to Unsupported Configuration: IDE
> >> controllers
> >> are unsupported for this QEMU binary or machine type
> >>
> >> I am using cloudstack on CentOS 7.9 with
> >>
> >> # rpm -qa | egrep "(qemu|libvirtd|OVMF|cloudstack)"
> >> *cloudstack*-agent-4.15.1.0-SNAPSHOT.20210601.el7.x86_64
> >> *qemu*-kvm-ev-2.12.0-44.1.el7_8.1.x86_64
> >> ipxe-roms-*qemu*-20180825-3.git133f4c.el7.noarch
> >> libvirt-daemon-driver-*qemu*-4.5.0-36.el7.x86_64
> >> *OVMF*-20180508-6.gitee3198e672e2.el7.noarch
> >> *qemu*-img-ev-2.12.0-44.1.el7_8.1.x86_64
> >> *cloudstack*-common-4.15.1.0-SNAPSHOT.20210601.el7.x86_64
> >> *qemu*-kvm-common-ev-2.12.0-44.1.el7_8.1.x86_64
> >>
> >> 1) Is the PR regarding the q35 chipset already included in the SNAPSHOT?
> >> I guess yes, because with 4.15.0 and i440 chipset the VM could be
> >> created and was running.
> >>
> >> 2) What do I have to change inside my settings to switch from IDE to
> >> SCSI or SATA?
> >>
> >> Regards Udo
> >>
>

Re: 4.15.1 SNAPSHOT: cant create users to account

[Slavka Peleva]

Hi Udo,

I was able to reproduce the problem with the new UI. With the old UI, you
can create a user. You can open an issue on GitHub for the problem

Regards,
Slavka

On Tue, Jun 1, 2021 at 6:46 PM Udo Müller  wrote:

> Oops
>
>
> now the screenshots are attached.
>
> Regards udo
>
> Am 01.06.21 um 17:43 schrieb Udo Müller:
> > Oh and btw: The Add user-Popup is rendered wrong in Safari, but is
> > working fine in Chrome and Firefox.
> >
> > See attached screenshots.
> >
> > Is it possible for me to change anything to get this back to work?
> >
> > Regards Udo
> >
> >
> > Am 01.06.21 um 17:33 schrieb Udo Müller:
> >> Hi all,
> >>
> >> The following was working with 4.15.0.
> >>
> >>
> >> Status quo: 4.15.1 SNAPSHOT, database new initialized. Configuration
> >> done with the same settings as 4.15.0
> >>
> >>
> >> i am trying to add a user to the account admin. So i click on
> >>
> >> Accounts -> admin -> View Users -> Add User
> >>
> >> Popup Window opens fine.
> >>
> >>
> >> - When i enter all required fields (username, password 2x, email,
> >> first and last name) and hit ok, i get the error, that the parameter
> >> "account" is missing.
> >>
> >> - When i do like above and then click on the drop-down field Domain
> >> it is empty. Same for drop-down Account.
> >>
> >>
> >> Means i cant create users to accounts which is absolutely needed for
> >> my application.
> >>
> >>
> >> management.log
> >>
> >> 2021-06-01 15:31:27,334 DEBUG [c.c.a.ApiServlet]
> >> (qtp1630678941-3995:ctx-cff2deda) (logid:0a4b2d51) ===START===
> >> 192.168.7.244 -- POST  command=createUser=json
> >> 2021-06-01 15:31:27,339 DEBUG [c.c.a.ApiServer]
> >> (qtp1630678941-3995:ctx-cff2deda ctx-f77a5abe) (logid:0a4b2d51) CIDRs
> >> from which account 'Acct[a7e89cef-c2ca-11eb-8de2-525400f5bc6b-admin]'
> >> is allowed to perform API calls: 0.0.0.0/0,::/0
> >> 2021-06-01 15:31:27,341 INFO  [c.c.a.ApiServer]
> >> (qtp1630678941-3995:ctx-cff2deda ctx-f77a5abe) (logid:0a4b2d51)
> >> Unable to execute API command createuser due to missing parameter
> >> account
> >>
> >> apilog.log
> >>
> >> 2021-06-01 15:31:27,342 INFO  [a.c.c.a.ApiServer]
> >> (qtp1630678941-3995:ctx-cff2deda ctx-f77a5abe) (logid:0a4b2d51)
> >> (userId=2 accountId=2 sessionId=node0cacgcc2ki0zr1kjau5c1p59w511)
> >> 192.168.7.244 -- POST command=createUser=json 431 Unable to
> >> execute API command createuser due to missing parameter account
> >>
> >>
>

Re: 4.15.0 and 4.15.1 SNAPSHOT: Infrastructure Hosts For input string error

[Daan Hoogland]

We've seen things like this and they have to do with for instance Dutch or
German locales used in the UI. I thought this was solved in 4.15.1 and up.
try using english and see if that works.

On Tue, Jun 1, 2021 at 5:40 PM Udo Müller  wrote:

> Hi guys,
>
>
> next one:
>
> I had this problem with 4.15.0 and also with 4.15.1 SNAPSHOT
>
> I have successfully done the configuration wizard and was able to add
> the first host to my cluster. After finishing it i wanted to add more
> hosts. So i clicked on Infrastructure -> Hosts and get the error message
> popup:
>
> For input string 0,58
>
> The number varies. See attached screenshot.
>
> I am able to add additional hosts with the cloudstack-setup-agent CLI
> but via WebUI.
>
>
> 2021-06-01 11:36:12,942 DEBUG [c.c.a.ApiServlet]
> (qtp1630678941-19:ctx-4b7e9914) (logid:7aca0057) ===START===
> 192.168.7.244 -- GET
>
> listall=true=routing=1=20=listHostsMetrics=json
> 2021-06-01 11:36:12,948 DEBUG [c.c.a.ApiServer]
> (qtp1630678941-19:ctx-4b7e9914 ctx-a4662b81) (logid:7aca0057) CIDRs from
> which account 'Acct[a7e89cef-c2ca-11eb-8de2-525400f5bc6b-admin]' is
> allowed to perform API calls: 0.0.0.0/0,::/0
> 2021-06-01 11:36:12,959 WARN [c.c.a.d.ParamGenericValidationWorker]
> (qtp1630678941-19:ctx-4b7e9914 ctx-a4662b81) (logid:7aca0057) Received
> unknown parameters for command listHostsMetrics. Unknown parameters :
> listall
> 2021-06-01 11:36:12,960 DEBUG [c.c.a.q.QueryManagerImpl]
> (qtp1630678941-19:ctx-4b7e9914 ctx-a4662b81) (logid:7aca0057)
>  >>>Searching for hosts>>>
> 2021-06-01 11:36:12,971 DEBUG [c.c.a.q.QueryManagerImpl]
> (qtp1630678941-19:ctx-4b7e9914 ctx-a4662b81) (logid:7aca0057)
>  >>>Generating Response>>>
> 2021-06-01 11:36:13,009 INFO  [c.c.a.ApiServer]
> (qtp1630678941-19:ctx-4b7e9914 ctx-a4662b81) (logid:7aca0057) For input
> string: "1,55"
>
>

-- 
Daan

Re: hide network from lower domains?

[Daan Hoogland]

Jordan, is there a reason it must be a shared network? It seems to me you
want an isolated network or a VPC.

On Tue, Jun 1, 2021 at 5:19 PM Yordan Kostov  wrote:

> Hey everyone,
>
> I am playing around with shared networks and I noticed
> that when network is created for the ROOT user the lower level domains see
> it too.
> Is there a way to hide that network from the lower level
> domain users?
>
> Here is my setup:
>
>   *   Domains
>  *   ROOT
>  *   ROOT/DEVELOPERS
>   *   Accounts - there is a separate account ACC-DEV for ROOT/DEVELOPERS
> domain
>
> One thing I tried was to create a network with offering that is not public
> but the result is the same.
>
> Best regards,
> Jordan
>


-- 
Daan

Re: 4.15.1 SNAPSHOT: cant create users to account

[Udo Müller]

Oops


now the screenshots are attached.

Regards udo

Am 01.06.21 um 17:43 schrieb Udo Müller:
Oh and btw: The Add user-Popup is rendered wrong in Safari, but is 
working fine in Chrome and Firefox.


See attached screenshots.

Is it possible for me to change anything to get this back to work?

Regards Udo


Am 01.06.21 um 17:33 schrieb Udo Müller:

Hi all,

The following was working with 4.15.0.


Status quo: 4.15.1 SNAPSHOT, database new initialized. Configuration 
done with the same settings as 4.15.0



i am trying to add a user to the account admin. So i click on

Accounts -> admin -> View Users -> Add User

Popup Window opens fine.


- When i enter all required fields (username, password 2x, email, 
first and last name) and hit ok, i get the error, that the parameter 
"account" is missing.


- When i do like above and then click on the drop-down field Domain 
it is empty. Same for drop-down Account.



Means i cant create users to accounts which is absolutely needed for 
my application.



management.log

2021-06-01 15:31:27,334 DEBUG [c.c.a.ApiServlet] 
(qtp1630678941-3995:ctx-cff2deda) (logid:0a4b2d51) ===START=== 
192.168.7.244 -- POST  command=createUser=json
2021-06-01 15:31:27,339 DEBUG [c.c.a.ApiServer] 
(qtp1630678941-3995:ctx-cff2deda ctx-f77a5abe) (logid:0a4b2d51) CIDRs 
from which account 'Acct[a7e89cef-c2ca-11eb-8de2-525400f5bc6b-admin]' 
is allowed to perform API calls: 0.0.0.0/0,::/0
2021-06-01 15:31:27,341 INFO  [c.c.a.ApiServer] 
(qtp1630678941-3995:ctx-cff2deda ctx-f77a5abe) (logid:0a4b2d51) 
Unable to execute API command createuser due to missing parameter 
account


apilog.log

2021-06-01 15:31:27,342 INFO  [a.c.c.a.ApiServer] 
(qtp1630678941-3995:ctx-cff2deda ctx-f77a5abe) (logid:0a4b2d51) 
(userId=2 accountId=2 sessionId=node0cacgcc2ki0zr1kjau5c1p59w511) 
192.168.7.244 -- POST command=createUser=json 431 Unable to 
execute API command createuser due to missing parameter account

Re: 4.15.1 SNAPSHOT: cant create users to account

[Udo Müller]

Oh and btw: The Add user-Popup is rendered wrong in Safari, but is 
working fine in Chrome and Firefox.


See attached screenshots.

Is it possible for me to change anything to get this back to work?

Regards Udo


Am 01.06.21 um 17:33 schrieb Udo Müller:

Hi all,

The following was working with 4.15.0.


Status quo: 4.15.1 SNAPSHOT, database new initialized. Configuration 
done with the same settings as 4.15.0



i am trying to add a user to the account admin. So i click on

Accounts -> admin -> View Users -> Add User

Popup Window opens fine.


- When i enter all required fields (username, password 2x, email, 
first and last name) and hit ok, i get the error, that the parameter 
"account" is missing.


- When i do like above and then click on the drop-down field Domain it 
is empty. Same for drop-down Account.



Means i cant create users to accounts which is absolutely needed for 
my application.



management.log

2021-06-01 15:31:27,334 DEBUG [c.c.a.ApiServlet] 
(qtp1630678941-3995:ctx-cff2deda) (logid:0a4b2d51) ===START=== 
192.168.7.244 -- POST  command=createUser=json
2021-06-01 15:31:27,339 DEBUG [c.c.a.ApiServer] 
(qtp1630678941-3995:ctx-cff2deda ctx-f77a5abe) (logid:0a4b2d51) CIDRs 
from which account 'Acct[a7e89cef-c2ca-11eb-8de2-525400f5bc6b-admin]' 
is allowed to perform API calls: 0.0.0.0/0,::/0
2021-06-01 15:31:27,341 INFO  [c.c.a.ApiServer] 
(qtp1630678941-3995:ctx-cff2deda ctx-f77a5abe) (logid:0a4b2d51) Unable 
to execute API command createuser due to missing parameter account


apilog.log

2021-06-01 15:31:27,342 INFO  [a.c.c.a.ApiServer] 
(qtp1630678941-3995:ctx-cff2deda ctx-f77a5abe) (logid:0a4b2d51) 
(userId=2 accountId=2 sessionId=node0cacgcc2ki0zr1kjau5c1p59w511) 
192.168.7.244 -- POST command=createUser=json 431 Unable to 
execute API command createuser due to missing parameter account

4.15.0 and 4.15.1 SNAPSHOT: Infrastructure Hosts For input string error

[Udo Müller]

Hi guys,


next one:

I had this problem with 4.15.0 and also with 4.15.1 SNAPSHOT

I have successfully done the configuration wizard and was able to add 
the first host to my cluster. After finishing it i wanted to add more 
hosts. So i clicked on Infrastructure -> Hosts and get the error message 
popup:


For input string 0,58

The number varies. See attached screenshot.

I am able to add additional hosts with the cloudstack-setup-agent CLI 
but via WebUI.



2021-06-01 11:36:12,942 DEBUG [c.c.a.ApiServlet] 
(qtp1630678941-19:ctx-4b7e9914) (logid:7aca0057) ===START=== 
192.168.7.244 -- GET 
listall=true=routing=1=20=listHostsMetrics=json
2021-06-01 11:36:12,948 DEBUG [c.c.a.ApiServer] 
(qtp1630678941-19:ctx-4b7e9914 ctx-a4662b81) (logid:7aca0057) CIDRs from 
which account 'Acct[a7e89cef-c2ca-11eb-8de2-525400f5bc6b-admin]' is 
allowed to perform API calls: 0.0.0.0/0,::/0
2021-06-01 11:36:12,959 WARN [c.c.a.d.ParamGenericValidationWorker] 
(qtp1630678941-19:ctx-4b7e9914 ctx-a4662b81) (logid:7aca0057) Received 
unknown parameters for command listHostsMetrics. Unknown parameters : 
listall
2021-06-01 11:36:12,960 DEBUG [c.c.a.q.QueryManagerImpl] 
(qtp1630678941-19:ctx-4b7e9914 ctx-a4662b81) (logid:7aca0057) 
>>>Searching for hosts>>>
2021-06-01 11:36:12,971 DEBUG [c.c.a.q.QueryManagerImpl] 
(qtp1630678941-19:ctx-4b7e9914 ctx-a4662b81) (logid:7aca0057) 
>>>Generating Response>>>
2021-06-01 11:36:13,009 INFO  [c.c.a.ApiServer] 
(qtp1630678941-19:ctx-4b7e9914 ctx-a4662b81) (logid:7aca0057) For input 
string: "1,55"

4.15.1 SNAPSHOT: cant create users to account

[Udo Müller]

Hi all,

The following was working with 4.15.0.


Status quo: 4.15.1 SNAPSHOT, database new initialized. Configuration 
done with the same settings as 4.15.0



i am trying to add a user to the account admin. So i click on

Accounts -> admin -> View Users -> Add User

Popup Window opens fine.


- When i enter all required fields (username, password 2x, email, first 
and last name) and hit ok, i get the error, that the parameter "account" 
is missing.


- When i do like above and then click on the drop-down field Domain it 
is empty. Same for drop-down Account.



Means i cant create users to accounts which is absolutely needed for my 
application.



management.log

2021-06-01 15:31:27,334 DEBUG [c.c.a.ApiServlet] 
(qtp1630678941-3995:ctx-cff2deda) (logid:0a4b2d51) ===START=== 
192.168.7.244 -- POST  command=createUser=json
2021-06-01 15:31:27,339 DEBUG [c.c.a.ApiServer] 
(qtp1630678941-3995:ctx-cff2deda ctx-f77a5abe) (logid:0a4b2d51) CIDRs 
from which account 'Acct[a7e89cef-c2ca-11eb-8de2-525400f5bc6b-admin]' is 
allowed to perform API calls: 0.0.0.0/0,::/0
2021-06-01 15:31:27,341 INFO  [c.c.a.ApiServer] 
(qtp1630678941-3995:ctx-cff2deda ctx-f77a5abe) (logid:0a4b2d51) Unable 
to execute API command createuser due to missing parameter account


apilog.log

2021-06-01 15:31:27,342 INFO  [a.c.c.a.ApiServer] 
(qtp1630678941-3995:ctx-cff2deda ctx-f77a5abe) (logid:0a4b2d51) 
(userId=2 accountId=2 sessionId=node0cacgcc2ki0zr1kjau5c1p59w511) 
192.168.7.244 -- POST command=createUser=json 431 Unable to 
execute API command createuser due to missing parameter account

Re: 4.15.1 Snapshot UEFI Legacy: Unable to start VM due to Unsupported Configuration: IDE controllers are unsupported for this QEMU binary or machine type

[Udo Müller]

Hi Slavka,

Deployment is done via WebUI. In this case i choose the target host 
(since only this one has debug enabled), then choose service, disk and 
network offering, activate advanced boot options and check UEFI Legacy.


After setting the name the vm is started.

The XML of the VM is attached.

There you see



But the VM will work with sata instead of ide. I tested it by defining a 
VM via virt-manager and choosing sata for the bus type.


See also:

https://bugs.launchpad.net/nova/+bug/1831538

Regards Udo


Am 01.06.21 um 16:16 schrieb Slavka Peleva:

Hi Udo,

Could you please share more information about the deployment of a VM?
Are you creating a new instance from ISO or a template? Also, can you set
the DEBUG level at the agent, and is it possible to share the XML that
CloudStack sends to libvirt (it's in the
/var/log/cloudstack/agent/agent.log)?

One more question did you update and the management with the 4.15.1.0
version?

Kind regards,
Slavka

On Tue, Jun 1, 2021 at 4:33 PM Udo Müller  wrote:


Hi guys,


I played around with cloudstack 4.15 and UEFI boot until someone here
pointed me to a PR which fixes the problem with the missing q35 chipset.

Thats the reason why I just installed the 4.15.1 SNAPSHOT.

Now I get the following problem:

Unable to start VM on Host... due to Unsupported Configuration: IDE
controllers
are unsupported for this QEMU binary or machine type

I am using cloudstack on CentOS 7.9 with

# rpm -qa | egrep "(qemu|libvirtd|OVMF|cloudstack)"
*cloudstack*-agent-4.15.1.0-SNAPSHOT.20210601.el7.x86_64
*qemu*-kvm-ev-2.12.0-44.1.el7_8.1.x86_64
ipxe-roms-*qemu*-20180825-3.git133f4c.el7.noarch
libvirt-daemon-driver-*qemu*-4.5.0-36.el7.x86_64
*OVMF*-20180508-6.gitee3198e672e2.el7.noarch
*qemu*-img-ev-2.12.0-44.1.el7_8.1.x86_64
*cloudstack*-common-4.15.1.0-SNAPSHOT.20210601.el7.x86_64
*qemu*-kvm-common-ev-2.12.0-44.1.el7_8.1.x86_64

1) Is the PR regarding the q35 chipset already included in the SNAPSHOT?
I guess yes, because with 4.15.0 and i440 chipset the VM could be
created and was running.

2) What do I have to change inside my settings to switch from IDE to
SCSI or SATA?

Regards Udo

  [175/1177]
i-4-10-VM
95a92efb-d5c5-4b27-ac52-8c9cb7bdba3e
Other (32-bit)


Apache Software Foundation
CloudStack KVM Hypervisor
95a92efb-d5c5-4b27-ac52-8c9cb7bdba3e



hvm
/usr/share/OVMF/OVMF_CODE.secboot.fd
/var/lib/libvirt/qemu/nvram/95a92efb-d5c5-4b27-ac52-8c9cb7bdba3e.fd



restart
destroy
destroy
524288



1

  [148/1177]




500




/usr/libexec/qemu-kvm








dd1a64aecc3c455eac6f

hide network from lower domains?

[Yordan Kostov]

Hey everyone,

I am playing around with shared networks and I noticed that 
when network is created for the ROOT user the lower level domains see it too.
Is there a way to hide that network from the lower level domain 
users?

Here is my setup:

  *   Domains
 *   ROOT
 *   ROOT/DEVELOPERS
  *   Accounts - there is a separate account ACC-DEV for ROOT/DEVELOPERS domain

One thing I tried was to create a network with offering that is not public but 
the result is the same.

Best regards,
Jordan

Re: Building a CloudStack environment with CloudStack+VMware+Nutanix

[Daan Hoogland]

ESXI servers are supposed to be reachable on the ACS mgmt net, but most
communications go through the vcenter machine. I'm not sure, but I don't
think the ACS mgmt net has to be the same as vshpere mgmt net.

On Tue, Jun 1, 2021 at 4:32 AM 小林 美佳子  wrote:

> Hi.
>
> Does anyone have a track record of building a CloudStack environment using
> a combination of CloudStack+VMware+Nutanix?
> If you have, what segment did you use for CloudStack Management Network,
> vCenter IP address and ESXi Management IP address? Are they all in the same
> segment, or is only ESXi in a different segment?
>
> It seems that Nutanix requirements require the ESXi management IP address
> to be set to a different segment than the CloudStack management network and
> vCenter IP address. I am concerned if all the functions will work fine when
> configured this way.
>
>

-- 
Daan

Re: 4.15.1 Snapshot UEFI Legacy: Unable to start VM due to Unsupported Configuration: IDE controllers are unsupported for this QEMU binary or machine type

[Slavka Peleva]

Hi Udo,

Could you please share more information about the deployment of a VM?
Are you creating a new instance from ISO or a template? Also, can you set
the DEBUG level at the agent, and is it possible to share the XML that
CloudStack sends to libvirt (it's in the
/var/log/cloudstack/agent/agent.log)?

One more question did you update and the management with the 4.15.1.0
version?

Kind regards,
Slavka

On Tue, Jun 1, 2021 at 4:33 PM Udo Müller  wrote:

> Hi guys,
>
>
> I played around with cloudstack 4.15 and UEFI boot until someone here
> pointed me to a PR which fixes the problem with the missing q35 chipset.
>
> Thats the reason why I just installed the 4.15.1 SNAPSHOT.
>
> Now I get the following problem:
>
> Unable to start VM on Host... due to Unsupported Configuration: IDE
> controllers
> are unsupported for this QEMU binary or machine type
>
> I am using cloudstack on CentOS 7.9 with
>
> # rpm -qa | egrep "(qemu|libvirtd|OVMF|cloudstack)"
> *cloudstack*-agent-4.15.1.0-SNAPSHOT.20210601.el7.x86_64
> *qemu*-kvm-ev-2.12.0-44.1.el7_8.1.x86_64
> ipxe-roms-*qemu*-20180825-3.git133f4c.el7.noarch
> libvirt-daemon-driver-*qemu*-4.5.0-36.el7.x86_64
> *OVMF*-20180508-6.gitee3198e672e2.el7.noarch
> *qemu*-img-ev-2.12.0-44.1.el7_8.1.x86_64
> *cloudstack*-common-4.15.1.0-SNAPSHOT.20210601.el7.x86_64
> *qemu*-kvm-common-ev-2.12.0-44.1.el7_8.1.x86_64
>
> 1) Is the PR regarding the q35 chipset already included in the SNAPSHOT?
> I guess yes, because with 4.15.0 and i440 chipset the VM could be
> created and was running.
>
> 2) What do I have to change inside my settings to switch from IDE to
> SCSI or SATA?
>
> Regards Udo
>

4.15.1 Snapshot UEFI Legacy: Unable to start VM due to Unsupported Configuration: IDE controllers are unsupported for this QEMU binary or machine type

[Udo Müller]

Hi guys,


I played around with cloudstack 4.15 and UEFI boot until someone here 
pointed me to a PR which fixes the problem with the missing q35 chipset.


Thats the reason why I just installed the 4.15.1 SNAPSHOT.

Now I get the following problem:

Unable to start VM on Host... due to Unsupported Configuration: IDE 
controllers

are unsupported for this QEMU binary or machine type

I am using cloudstack on CentOS 7.9 with

# rpm -qa | egrep "(qemu|libvirtd|OVMF|cloudstack)"
*cloudstack*-agent-4.15.1.0-SNAPSHOT.20210601.el7.x86_64
*qemu*-kvm-ev-2.12.0-44.1.el7_8.1.x86_64
ipxe-roms-*qemu*-20180825-3.git133f4c.el7.noarch
libvirt-daemon-driver-*qemu*-4.5.0-36.el7.x86_64
*OVMF*-20180508-6.gitee3198e672e2.el7.noarch
*qemu*-img-ev-2.12.0-44.1.el7_8.1.x86_64
*cloudstack*-common-4.15.1.0-SNAPSHOT.20210601.el7.x86_64
*qemu*-kvm-common-ev-2.12.0-44.1.el7_8.1.x86_64

1) Is the PR regarding the q35 chipset already included in the SNAPSHOT?
I guess yes, because with 4.15.0 and i440 chipset the VM could be 
created and was running.


2) What do I have to change inside my settings to switch from IDE to 
SCSI or SATA?


Regards Udo

RE: VMware Instance Error

[Corey, Mike]

Thanks for the reply Abhishek!

1- ACS deploys the VM from template; however, it fails and shows Error in 
ACS-Instances UI.  So the VM remains in vCenter - that VM has the CPU/RAM as 
the template NOT what the ACS service offering (Small Instance) I selected.  I 
suppose ACS is not finishing the configuration because the deployment fails.

3 - This is the issue I need help to troubleshoot.  ACS IS NOT creating any 
settings on the Windows 10 and CentOS template I've uploaded into ACS.  Both 
template OVF files have the hardware descriptions used by the template VM.  
However, NO SETTINGS are being added to the ACS template.  

How can this be troubleshot?  Obviously, ACS is NOT populating the settings on 
the template as expected - what could be the issue?  What commands occur when 
the template is downloaded and populate the settings of the template (so I can 
search for errors)?  Is something supposed to be written to ACS but is not due 
to permissions or file doesn't exist?

Item 1 & 4 - Yes, the VM in vCenter is named the same as the ones that fail 
deployment in ACS (i-2-##-vm) - the VM has resources of the template (not 
service offering) and powers ON and loads OS.   

5 - snippet from logs below - diskcontroller is being identified by OVF - 
nothing is set by ACS on template.

INFO  [c.c.h.v.r.VmwareResource] (DirectAgent-496:ctx-f077199e 
usphlmvesxt01.phl.global.corp.sap, job-164/job-166, cmd: StartCommand) 
(logid:6c707559) Executing resource StartCommand: 
{"vm":{"id":"26","name":"i-2-26-VM","bootloader":"HVM","state":"Starting","type":"User","cpus":"1","minSpeed":"166","maxSpeed":"500","minRam":"(512.00
 MB) 536870912","maxRam":"(512.00 MB) 
536870912","hostName":"June1-01","arch":"x86_64","os":"Windows 10 
(64-bit)","platformEmulator":"windows9_64Guest","bootArgs":"","enableHA":"false","limitCpuUse":"false","enableDynamicallyScaleVm":"false","vncPassword":"X1k9AVje8WIPzXb_8P_88Q","params":{"deployvm":"true","dataDiskController":"scsi","memoryOvercommitRatio":"1.0","Message.ReservedCapacityFreed.Flag":"false","nestedVirtualizationFlag":"false","cpuOvercommitRatio":"3.0","vmware.reserve.mem":"false","vmware.reserve.cpu":"false","nicAdapter":"E1000","rootDiskController":"lsilogicsas"},"uuid":"dafd18a5-04a9-4bda-b701-0d24e4b1e61d","enterHardwareSetup":"false","disks":[{"data":{"org.apache.cloudstack.storage.to.VolumeObjectTO":{"uuid":"973a1795-4569-4769-afcb-93e3911be12e","volumeType":"ROOT","dataStore":{"org.apache.cloudstack.storage.to.PrimaryDataStoreTO":{"uuid":"9df43ec0-1e52-3cc3-8e5a-bf2313bbbd8d","id":"1","poolType":"NetworkFilesystem","host":"usphlcv302.phl.global.corp.sap","path":"/usphlmvesx_lab_cspri01","port":"2049","url":"NetworkFilesystem:///usphlmvesx_lab_cspri01/?ROLE=Primary=9df43ec0-1e52-3cc3-8e5a-bf2313bbbd8d","isManaged":"false"}},"name":"ROOT-26-0","size":"(80.00
 GB) 
85899345920","volumeId":"17","vmName":"i-2-26-VM","accountId":"2","format":"OVA","provisioningType":"THIN","id":"17","deviceId":"0","bytesReadRate":"(0
 bytes) 0","bytesWriteRate":"(0 bytes) 0","iopsReadRate":"(0 bytes) 
0","iopsWriteRate":"(0 bytes) 
0","hypervisorType":"VMware","directDownload":"false","deployAsIs":"true"}},"diskSeq":"0","type":"ROOT","_details":{"storageHost":"usphlcv302.phl.global.corp.sap","managed":"false","storagePort":"2049","volumeSize":"(80.00
 GB) 
85899345920"}},{"data":{"org.apache.cloudstack.storage.to.TemplateObjectTO":{"id":"0","format":"ISO","accountId":"0","hvm":"false","bootable":"false","directDownload":"false","deployAsIs":"false"}},"diskSeq":"3","type":"ISO"}],"nics":[{"deviceId":"0","networkRateMbps":"200","defaultNic":"true","pxeDisable":"false","nicUuid":"7587229e-bbb3-4e65-b093-fa9ff690bb43","details":{"MacAddressChanges":"true","ForgedTransmits":"true","PromiscuousMode":"false"},"dpdkEnabled":"false","uuid":"62bfbed1-d07c-4886-b4f4-bf594543bb28","ip":"10.4.80.238","netmask":"255.255.240.0","gateway":"10.4.80.1","mac":"1e:00:9d:00:00:4d","dns1":,"dns2":"","broadcastType":"Vlan","type":"Guest","broadcastUri":"vlan://
 ","isolationUri":"vlan:// 
","isSecurityGroupEnabled":"false","name":"vds3-phl-csm02-vm,138,vmwaredvs"}],"guestOsDetails":{},"extraConfig":{},"deployAsIsInfo":{"nicAdapterMap":{}}},"hostIp":"","executeInSequence":"true","wait":"0"}
2021-06-01 08:12:41,346 WARN  [c.c.a.m.DirectAgentAttache] 
(DirectAgent-496:ctx-f077199e) (logid:6c707559) Seq 1-7962082666214335044: 
Throwable caught while executing command
com.cloud.utils.exception.CloudRuntimeException: Invalid root disk controller 
detected : none

Many thanks!

Mike


-Original Message-
From: Abhishek Kumar  
Sent: Monday, May 31, 2021 8:46 AM
To: users@cloudstack.apache.org
Subject: Re: VMware Instance Error

Hi Mike,

1 - The CPU/RAM  and network on the VMware side should match the selected 
config from ACS end. Make sure you are checking the same ACS and vCenter VM. 
You may match the internal name.
2 - Mentioned logs are DEBUG logs. Wrt deployment behaviour, ACS first tries to 
allocate 

RE: Quickstart question

[Corey, Mike]

Jeremy,

Check this global setting.  I've set it to 3 as we use NFS 3 in our environment.

secstorage.nfs.version  Enforces specific NFS version when mounting Secondary 
Storage. If NULL default selection is performed

-Original Message-
From: Harikrishna Patnala  
Sent: Sunday, May 30, 2021 10:01 PM
To: users@cloudstack.apache.org
Subject: Re: Quickstart question

Hi Jeremy,

May I know the NFS version number that you are using in your environment?

Also, check the status of the default template by clicking on the template and 
then into the zones tab, to check if it is downloaded properly or not.

Regards,
Harikrishna




 

From: Jeremy Hansen
Sent: Sunday, May 30, 2021 4:51 PM
To: users@cloudstack.apache.org
Subject: Re: Quickstart question

Went over the process about 20 times and made progress.  I'm able to launch a 
VM as long as the template is a direct download bypassing secondary storage.  I 
see this in the logs:

2021-05-30 04:01:57,148 DEBUG [c.c.a.t.Request] (AgentManager-Handler-12:null) 
(logid:) Seq 2-4381439487478071309: Processing:  { Ans: , MgmtId: 
198690285003419, via: 2, Ver: v1, Flags: 10, 
[{"com.cloud.agent.api.Answer":{"result":"false","details":"com.cloud.utils.exception.CloudRuntimeException:
 GetRootDir for nfs://192.168.10.60/volume1/cloudstack-dev/secondary failed due 
to com.cloud.utils.exception.CloudRuntimeException: Unable to mount 
192.168.10.60:/volume1/cloudstack-dev/secondary at 
/mnt/SecStorage/4683f710-43e8-359d-bc9a-041363aa77b9 due to mount.nfs: parsing 
error on 'vers=' option

I'm not sure how to troubleshoot this issue.  Do I have to specify the version 
somewhere for NFS?

Thanks

> On May 30, 2021, at 12:30 AM, Jeremy Hansen  wrote:
>
> At the very end of the configuration wizard when it's doing the Host vm 
> install, I get this error and I can't get past it:
>
> Something went wrong; please correct the following:
> TypeError: Cannot read property 'value' of undefined
>
> Thanks
> -jeremy
>
>
>> On May 29, 2021, at 11:27 PM, Jeremy Hansen  wrote:
>>
>> New to Cloudstack.  I'm trying to get the Quickstart going.
>>
>> It feels like an issue with secondary storage but I'm not quite sure how to 
>> troubleshoot.
>>
>> CentOS 8, kvm.  Created a bridge per the quick install docs.
>>
>> 
>>
>> I see no data for templates and isos.  I did do the template install.  The 
>> secondary storage VM is running and connected:
>>
>> 
>>
>>
>> 
>>
>>
>> 
>>
>> I attempted to upload an iso and it failed.
>>
>> Also, how do I get back to the initial configuration wizard.  I dismissed it 
>> but I'm not sure how to get it back.
>>
>> Thanks!
>> -jeremy
>>
>>
>>
>>
>>
>>
>>
>>
>>
>

Re: Issue with live migrations

[Alessandro Caviglione]

No, but if you're using Advanced Networking you've to manage the trunk
configuration on the switch ports.
Every Virtual Router has a public IP on a specific interface and a specific
VLAN and a private IP on another interface in another vlan that acts as a
gateway for the instance.
So, if VR and instance are in the same host, packets do not exit from host
and instance are reachable, but if they run on different hosts the packets
must go through a switch that should be configured to allow traffic on that
private VLAN.
So, in fact, you can allow a single switch port to forward traffic on
multiple VLAN, this is called Trunk.

On Tue, Jun 1, 2021 at 12:31 PM Jeremy Hansen  wrote:

> Ahh, so I just noticed that I could also migrate the router and when I do
> that, everything works as expected again.  So is this how migrates work?
> Do routers always have to move with the instance?
>
> -jeremy
>
>
> On Jun 1, 2021, at 3:26 AM, Jeremy Hansen  wrote:
>
> One thing I noticed, and I don’t know if this is expected, but the virtual
> router my instance is using remains on the original VM host.  If I log in
> to the guest instance while it’s on the second VM host, I can no longer
> ping the private net gateway:
>
> 
>
> and the virtual router can’t ping the private IP
>
> 
>
> If I migrate the instance back, it ping reachable again:
>
> 
>
>
> Thanks
> -jeremy
>
>
>
> On Jun 1, 2021, at 3:07 AM, Jeremy Hansen  wrote:
>
> Hmm. VM hosts are on the same vlan. Port configuration on both hosts is
> exactly the same. I’m not saying it isn’t vlan related but I would expect
> neither host to work if it was a vlan issue?
>
> Thanks
>
> On Jun 1, 2021, at 2:47 AM, Alessandro Caviglione 
> wrote:
>
> I think that your switch doesn't have ports tagged with all required
> VLANs...
>
> On Tue, Jun 1, 2021 at 11:20 AM Jeremy Hansen  wrote:
>
>
> I set up another VM host so now I have two VM hosts.  I’m experimenting
> with live migrations.  It looks like the VM migrates, but I can no longer
> SSH to the guest once it’s migrated.  I’m using an isolated network.  I
> setup firewall rules to allow all, along with port forwarding rules for
> ssh.  I’m using the CentOS 5.5 template.  When I migrate to the second VM
> host, ping/icmp still works, but I’m unable to ssh to the VM guest.  If I
> migrate back to the original VM host, everything is fine again.  Any clues
> what I’m doing wrong on the second VM host or why rules would change on the
> migration?
>
> Thanks
> -jeremy
>
>
>
>
>
>

Re: CSEUG Virtual - Session Recordings

[Slavka Peleva]

Thanks for sharing, Ivet,

The event was great! Many thanks to all that participated in it! :)

Kind regards,
Slavka

On Tue, Jun 1, 2021 at 11:27 AM Ivet Petrova 
wrote:

> Hi All and Happy Tuesday :)
>
> If you have missed CSEUG Virtual or some talks out of it, I am happy to
> share that this week we will publish the recordings on the official Apache
> CloudStack YouTube channel.
> The videos will come live this and next week on an instant premier on the
> channel.
>
> If you still have not subscribed to ACS channel on YouTube, now is the
> time to do so, to be notified when the videos come live:
>
> Watch the videos here:
> https://www.youtube.com/watch?v=Ph-NQv80j0k=PLnIKk7GjgFlbG2W3lRW9Evjy2Vtg_HKHX
>
> Kind regards,
>
>
>
>
>

Re: Issue with live migrations

[Jeremy Hansen]

Ahh, so I just noticed that I could also migrate the router and when I do that, 
everything works as expected again.  So is this how migrates work?  Do routers 
always have to move with the instance?

-jeremy


> On Jun 1, 2021, at 3:26 AM, Jeremy Hansen  wrote:
> 
> One thing I noticed, and I don’t know if this is expected, but the virtual 
> router my instance is using remains on the original VM host.  If I log in to 
> the guest instance while it’s on the second VM host, I can no longer ping the 
> private net gateway:
> 
> 
> 
> and the virtual router can’t ping the private IP
> 
> 
> 
> If I migrate the instance back, it ping reachable again:
> 
> 
> 
> 
> Thanks
> -jeremy
> 
> 
> 
>> On Jun 1, 2021, at 3:07 AM, Jeremy Hansen > > wrote:
>> 
>> Hmm. VM hosts are on the same vlan. Port configuration on both hosts is 
>> exactly the same. I’m not saying it isn’t vlan related but I would expect 
>> neither host to work if it was a vlan issue?
>> 
>> Thanks
>> 
>>> On Jun 1, 2021, at 2:47 AM, Alessandro Caviglione >> > wrote:
>>> 
>>> I think that your switch doesn't have ports tagged with all required
>>> VLANs...
>>> 
 On Tue, Jun 1, 2021 at 11:20 AM Jeremy Hansen >>> > wrote:
 
 
 I set up another VM host so now I have two VM hosts.  I’m experimenting
 with live migrations.  It looks like the VM migrates, but I can no longer
 SSH to the guest once it’s migrated.  I’m using an isolated network.  I
 setup firewall rules to allow all, along with port forwarding rules for
 ssh.  I’m using the CentOS 5.5 template.  When I migrate to the second VM
 host, ping/icmp still works, but I’m unable to ssh to the VM guest.  If I
 migrate back to the original VM host, everything is fine again.  Any clues
 what I’m doing wrong on the second VM host or why rules would change on the
 migration?
 
 Thanks
 -jeremy
 
 
>> 
> 



signature.asc
Description: Message signed with OpenPGP

Re: Issue with live migrations

[Jeremy Hansen]

Hmm. VM hosts are on the same vlan. Port configuration on both hosts is exactly 
the same. I’m not saying it isn’t vlan related but I would expect neither host 
to work if it was a vlan issue?

Thanks

> On Jun 1, 2021, at 2:47 AM, Alessandro Caviglione  
> wrote:
> 
> I think that your switch doesn't have ports tagged with all required
> VLANs...
> 
>> On Tue, Jun 1, 2021 at 11:20 AM Jeremy Hansen  wrote:
>> 
>> 
>> I set up another VM host so now I have two VM hosts.  I’m experimenting
>> with live migrations.  It looks like the VM migrates, but I can no longer
>> SSH to the guest once it’s migrated.  I’m using an isolated network.  I
>> setup firewall rules to allow all, along with port forwarding rules for
>> ssh.  I’m using the CentOS 5.5 template.  When I migrate to the second VM
>> host, ping/icmp still works, but I’m unable to ssh to the VM guest.  If I
>> migrate back to the original VM host, everything is fine again.  Any clues
>> what I’m doing wrong on the second VM host or why rules would change on the
>> migration?
>> 
>> Thanks
>> -jeremy
>> 
>> 

Re: Issue with live migrations

[Alessandro Caviglione]

I think that your switch doesn't have ports tagged with all required
VLANs...

On Tue, Jun 1, 2021 at 11:20 AM Jeremy Hansen  wrote:

>
> I set up another VM host so now I have two VM hosts.  I’m experimenting
> with live migrations.  It looks like the VM migrates, but I can no longer
> SSH to the guest once it’s migrated.  I’m using an isolated network.  I
> setup firewall rules to allow all, along with port forwarding rules for
> ssh.  I’m using the CentOS 5.5 template.  When I migrate to the second VM
> host, ping/icmp still works, but I’m unable to ssh to the VM guest.  If I
> migrate back to the original VM host, everything is fine again.  Any clues
> what I’m doing wrong on the second VM host or why rules would change on the
> migration?
>
> Thanks
> -jeremy
>
>

Issue with live migrations

[Jeremy Hansen]

I set up another VM host so now I have two VM hosts.  I’m experimenting with 
live migrations.  It looks like the VM migrates, but I can no longer SSH to the 
guest once it’s migrated.  I’m using an isolated network.  I setup firewall 
rules to allow all, along with port forwarding rules for ssh.  I’m using the 
CentOS 5.5 template.  When I migrate to the second VM host, ping/icmp still 
works, but I’m unable to ssh to the VM guest.  If I migrate back to the 
original VM host, everything is fine again.  Any clues what I’m doing wrong on 
the second VM host or why rules would change on the migration?

Thanks
-jeremy



signature.asc
Description: Message signed with OpenPGP

CSEUG Virtual - Session Recordings

[Ivet Petrova]

Hi All and Happy Tuesday :)

If you have missed CSEUG Virtual or some talks out of it, I am happy to share 
that this week we will publish the recordings on the official Apache CloudStack 
YouTube channel.
The videos will come live this and next week on an instant premier on the 
channel.

If you still have not subscribed to ACS channel on YouTube, now is the time to 
do so, to be notified when the videos come live:

Watch the videos here: 
https://www.youtube.com/watch?v=Ph-NQv80j0k=PLnIKk7GjgFlbG2W3lRW9Evjy2Vtg_HKHX

Kind regards,


 

Re: cloudstack 4.15: UEFI boot, chipset q35

[Slavka Peleva]

Hi Udo,

Unfortunately, I don't know the exact date. I guess it will be announced in
a few days.
Here is the thread of the discussion of its release
https://www.mail-archive.com/dev@cloudstack.apache.org/msg95593.html

Regards,
Slavka

On Mon, May 31, 2021 at 10:59 PM Udo Müller  wrote:

> Hi Slavka,
>
> Thx for your answer. I found that already and thought it is included in my
> download of 4.15.0
>
> Is a date for 4.15.1 known already?
>
> Regards Udo
>
> Gruß Udo Müller
>
> > Am 31.05.2021 um 18:42 schrieb Slavka Peleva
> :
> >
> > Hi Udo,
> >
> > There is a fix of this problem with UEFI legacy boot mode, which is
> > included in version 4.14.1.0, and will also be included in 4.15.1.0
> > PR 4576
> >
> > 
> > Best regards,
> > Slavka
> > 
> >
> >> On Mon, May 31, 2021 at 5:58 PM Udo Müller  wrote:
> >>
> >> Hi guys,
> >>
> >> i have set up a cloudstack 4.15 with a KVM cluster and i try to get UEFI
> >> legacy boots running.
> >>
> >> The VM itself boots fine but i cant get a console via cloudstack or
> >> virt-manager
> >>
> >> Guest has not initialized the graphics
> >>
> >> Checking the XML of the VM i can see:
> >>
> >> # virsh dumpxml 48 | egrep "(domain type|arch|loader|nvram)"
> >> 
> >> hvm
> >>  >> type='pflash'>/usr/share/OVMF/OVMF_CODE.secboot.fd
> >>  >>
> >>
> template='/usr/share/OVMF/OVMF_VARS.fd'>/var/lib/libvirt/qemu/nvram/c8e95c63-34d5-40f9-88ca-ad37f87db345.fd
> >>
> >>
> >> But this is not what i expect:
> >>
> >> hvm
> >>
> >> should be sth like
> >>
> >> hvm
> >>
> >>
> >> It looks like cloudstack isnt generating the correct XML. What is needed
> >> to make cloudstack manager generate the correct XML containing the
> >> supported chipset?
> >>
> >>
> >> System: CentOS 7.9
> >>
> >> # rpm -qa | grep cloudstack
> >> cloudstack-common-4.15.0.0-1.el7.x86_64
> >> cloudstack-management-4.15.0.0-1.el7.x86_64
> >> cloudstack-cli-4.15.0.0-1.el7.x86_64
> >>
> >>
> >> Agent: CentOS 7.9
> >>
> >> # rpm -qa | grep cloudstack
> >> cloudstack-common-4.15.0.0-1.el7.x86_64
> >> cloudstack-management-4.15.0.0-1.el7.x86_64
> >> cloudstack-cli-4.15.0.0-1.el7.x86_64
> >>
> >>
> >>
> >> The needed
> >>
> >>
>
>