cloudstack-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Anthony Xu <Xuefei...@citrix.com>
Subject RE: Deploying using ISOs
Date Thu, 12 Jul 2012 00:39:44 GMT
We removed HVM from UI intentionally due to following reasons,
1. most users may not know what's HVM, it is confusing.
2. for vmware, it doesn't care about HVM, it can boot any OS on any x86_64 host.
3. for kvm, HVM support is required. 
So it is hard for user to configure hvm for template/iso.
And 
4. most of hosts support HVM.


Only XenServer cares about HVM, and even for XenServer , HVM is OS specific, for example,
XenServer only supports PV linux, no HVM linux.
Luckily, XenServer will complain if you want to start a HVM on a non-HVM host.


So I think we can add a global configuration to enable/disable all HVM checks in CloudStack


Anthony







> -----Original Message-----
> From: Nitin Mehta [mailto:Nitin.Mehta@citrix.com]
> Sent: Wednesday, July 11, 2012 5:07 PM
> To: cloudstack-users@incubator.apache.org
> Subject: RE: Deploying using ISOs
> 
> Celeb - could you please open a UI bug for this @
> http://bugs.cloudstack.org/secure/Dashboard.jspa ? The UI should have
> an option to mark/unmark the template/iso hvm while registering it. The
> api already provides that.
> 
> Thanks,
> -Nitin
> -----Original Message-----
> From: Caleb Call [mailto:calebcall@me.com]
> Sent: Wednesday, July 11, 2012 11:37 AM
> To: cloudstack-users@incubator.apache.org
> Subject: Re: Deploying using ISOs
> 
> Great, thanks, that's exactly what it was.  VM started fine this time.
> 
> On Jul 11, 2012, at 12:29 PM, Edison Su wrote:
> 
> > Yah, that's the problem. All the templates/ISOs uploaded through UI,
> are marked as HVM enabled by default, while your hypervisor doesn't
> support HVM, then you can't create VM.
> > You can hack the DB, by "update vm_template set hvm=0", then try to
> create the VM again. But you need to make sure, your ISO does support
> PV.
> >
> >> -----Original Message-----
> >> From: Caleb Call [mailto:calebcall@me.com]
> >> Sent: Wednesday, July 11, 2012 11:18 AM
> >> To: cloudstack-users@incubator.apache.org
> >> Subject: Re: Deploying using ISOs
> >>
> >> Guest OS is CentOS, I've tried setting it to CentOS 6 (what it is),
> >> CentOS 5.6, and CentOS 5.3 (what the CentOS 5.6 template uses) and
> it
> >> still does not work.  My hypervisor is Xenserver 6 without HVM
> >> support (test environment, using older hardware).
> >>
> >>
> >> On Jul 11, 2012, at 12:13 PM, Edison Su wrote:
> >>
> >>> What's the guest os type of your ISO? How about change the guest OS
> >> type to CentOS 5.6 also? Will it work?
> >>> And what's your hypervisor? Does your hypervisor host support HVM?
> >>>
> >>>> -----Original Message-----
> >>>> From: Caleb Call [mailto:calebcall@me.com]
> >>>> Sent: Wednesday, July 11, 2012 2:28 AM
> >>>> To: cloudstack-users@incubator.apache.org
> >>>> Subject: Deploying using ISOs
> >>>>
> >>>> I'm trying to create VMs using ISOs I've uploaded.  The ISOs all
> >>>> say they are ready to be used and they show up fine during
> instance
> >>>> creation.  However, once I create the instance, I get an error
> that
> >>>> says "Unable to create a deployment for VM[User|<vm id>]".
> Looking
> >>>> through the logs, it seems like the way it's determining resources
> >> is
> >>>> different for ISO based and Template based.  I can create a VM
> >>>> using the included CentOS 5.6 template with the exact same compute
> >> offering.
> >>>> In the logs I get this when creating a VM from a template:
> >>>>
> >>>> -------------
> >>>> 2012-07-11 03:08:36,893 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-1:job-51) DeploymentPlanner allocation algorithm:
> firstfit
> >>>> 2012-07-11 03:08:36,893 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-1:job-51) Trying to allocate a host and storage pools
> from
> >>>> dc:2, pod:null,cluster:null, requested cpu: 500, requested ram:
> >>>> 536870912
> >>>> 2012-07-11 03:08:36,893 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-1:job-51) Is ROOT volume READY (pool already allocated)?:
> >> No
> >>>> 2012-07-11 03:08:36,893 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-1:job-51) Searching all possible resources under this
> Zone:
> >> 2
> >>>> 2012-07-11 03:08:36,898 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-1:job-51) Listing clusters in order of aggregate capacity,
> >>>> that have (atleast one host with) enough CPU and RAM capacity
> under
> >>>> this Zone: 2
> >>>> 2012-07-11 03:08:36,900 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-1:job-51) CPUOverprovisioningFactor considered: 3.0
> >>>> 2012-07-11 03:08:36,921 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-1:job-51) Checking resources in Cluster: 3 under Pod: 2
> >>>> 2012-07-11 03:08:36,922 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-1:job-51) Calling HostAllocators to find suitable hosts
> >>>> 2012-07-11 03:08:36,922 DEBUG [allocator.impl.FirstFitAllocator]
> >> (Job-
> >>>> Executor-1:job-51 FirstFitRoutingAllocator) Looking for hosts in
> dc:
> >> 2
> >>>> pod:2  cluster:3
> >>>> 2012-07-11 03:08:36,928 DEBUG [allocator.impl.FirstFitAllocator]
> >> (Job-
> >>>> Executor-1:job-51 FirstFitRoutingAllocator) FirstFitAllocator has
> 2
> >>>> hosts to check for allocation: [Host[-1-Routing], Host[-5-Routing]]
> >>>> 2012-07-11 03:08:36,949 DEBUG [allocator.impl.FirstFitAllocator]
> >> (Job-
> >>>> Executor-1:job-51 FirstFitRoutingAllocator) Found 2 hosts for
> >>>> allocation after prioritization: [Host[-5-Routing],
> >>>> Host[-1-Routing]]
> >>>> 2012-07-11 03:08:36,949 DEBUG [allocator.impl.FirstFitAllocator]
> >> (Job-
> >>>> Executor-1:job-51 FirstFitRoutingAllocator) Looking for
> >>>> speed=500Mhz,
> >>>> Ram=512
> >>>> 2012-07-11 03:08:36,972 DEBUG [cloud.capacity.CapacityManagerImpl]
> >>>> (Job-Executor-1:job-51 FirstFitRoutingAllocator) Checking if host:
> >>>> 5 has enough capacity for requested CPU: 500 and requested RAM:
> >>>> 536870912 , cpuOverprovisioningFactor
> >>>> : 3.0
> >>>> 2012-07-11 03:08:36,986 DEBUG [cloud.capacity.CapacityManagerImpl]
> >>>> (Job-Executor-1:job-51 FirstFitRoutingAllocator) Hosts's actual
> >> total
> >>>> CPU: 11172 and CPU after applying overprovisioning: 33516
> >>>> 2012-07-11 03:08:36,986 DEBUG [cloud.capacity.CapacityManagerImpl]
> >>>> (Job-Executor-1:job-51 FirstFitRoutingAllocator) Free CPU: 30516 ,
> >>>> Requested CPU: 500
> >>>> 2012-07-11 03:08:36,994 DEBUG [cloud.capacity.CapacityManagerImpl]
> >>>> (Job-Executor-1:job-51 FirstFitRoutingAllocator) Free RAM:
> >> 4830291968 ,
> >>>> Requested RAM: 536870912
> >>>> 2012-07-11 03:08:36,994 DEBUG [cloud.capacity.CapacityManagerImpl]
> >>>> (Job-Executor-1:job-51 FirstFitRoutingAllocator) Host has enough
> >>>> CPU and RAM available
> >>>> -------------
> >>>>
> >>>> However, using the exact same compute offering, when I try to
> >>>> create
> >> a
> >>>> VM using an ISO, I get the following in the logs:
> >>>>
> >>>> -------------
> >>>> 2012-07-11 03:13:58,534 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-2:job-52) DeploymentPlanner allocation algorithm:
> firstfit
> >>>> 2012-07-11 03:13:58,535 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-2:job-52) Trying to allocate a host and storage pools
> from
> >>>> dc:2, pod:null,cluster:null, requested cpu: 500, requested ram:
> >>>> 536870912
> >>>> 2012-07-11 03:13:58,535 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-2:job-52) Is ROOT volume READY (pool already allocated)?:
> >> No
> >>>> 2012-07-11 03:13:58,535 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-2:job-52) Searching all possible resources under this
> Zone:
> >> 2
> >>>> 2012-07-11 03:13:58,539 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-2:job-52) Listing clusters in order of aggregate capacity,
> >>>> that have (atleast one host with) enough CPU and RAM capacity
> under
> >>>> this Zone: 2
> >>>> 2012-07-11 03:13:58,541 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-2:job-52) CPUOverprovisioningFactor considered: 3.0
> >>>> 2012-07-11 03:13:58,558 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-2:job-52) Checking resources in Cluster: 3 under Pod: 2
> >>>> 2012-07-11 03:13:58,558 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-2:job-52) Calling HostAllocators to find suitable hosts
> >>>> 2012-07-11 03:13:58,558 DEBUG [allocator.impl.FirstFitAllocator]
> >> (Job-
> >>>> Executor-2:job-52 FirstFitRoutingAllocator) Looking for hosts in
> dc:
> >> 2
> >>>> pod:2  cluster:3
> >>>> 2012-07-11 03:13:58,564 DEBUG [allocator.impl.FirstFitAllocator]
> >> (Job-
> >>>> Executor-2:job-52 FirstFitRoutingAllocator) FirstFitAllocator has
> 2
> >>>> hosts to check for allocation: [Host[-1-Routing], Host[-5-Routing]]
> >>>> 2012-07-11 03:13:58,567 DEBUG [allocator.impl.FirstFitAllocator]
> >> (Job-
> >>>> Executor-2:job-52 FirstFitRoutingAllocator) Found 0 hosts for
> >>>> allocation after prioritization: []
> >>>> 2012-07-11 03:13:58,568 DEBUG [allocator.impl.FirstFitAllocator]
> >> (Job-
> >>>> Executor-2:job-52 FirstFitRoutingAllocator) Looking for
> >>>> speed=500Mhz,
> >>>> Ram=512
> >>>> 2012-07-11 03:13:58,568 DEBUG [allocator.impl.FirstFitAllocator]
> >> (Job-
> >>>> Executor-2:job-52 FirstFitRoutingAllocator) Host Allocator
> >>>> returning
> >> 0
> >>>> suitable hosts
> >>>> 2012-07-11 03:13:58,568 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-2:job-52) No suitable hosts found
> >>>> 2012-07-11 03:13:58,568 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-2:job-52) No suitable hosts found under this Cluster: 3
> >>>> 2012-07-11 03:13:58,568 DEBUG [cloud.deploy.FirstFitPlanner] (Job-
> >>>> Executor-2:job-52) Could not find suitable Deployment Destination
> >> for
> >>>> this VM under any clusters, returning.
> >>>> 2012-07-11 03:13:58,654 DEBUG [cloud.capacity.CapacityManagerImpl]
> >>>> (Job-Executor-2:job-52) VM state transitted from :Starting to
> >> Stopped
> >>>> with event: OperationFailedvm's original host id: null new host id:
> >>>> null host id before state tran
> >>>> sition: null
> >>>> 2012-07-11 03:13:58,754 DEBUG [cloud.capacity.CapacityManagerImpl]
> >>>> (Job-Executor-2:job-52) VM state transitted from :Stopped to Error
> >> with
> >>>> event: OperationFailedToErrorvm's original host id: null new host
> id:
> >>>> null host id before state
> >>>> transition: null
> >>>> 2012-07-11 03:13:58,800 DEBUG [agent.manager.AgentManagerImpl]
> >>>> (AgentManager-Handler-15:null) SeqA 3-187783: Processing Seq 3-
> >> 187783:
> >>>> { Cmd , MgmtId: -1, via: 3, Ver: v1, Flags: 11,
> >>>> [{"ConsoleProxyLoadReportCommand":{"_proxyVmId":2,"_
> >>>> loadInfo":"{\n  \"connections\": [\n    {\n      \"id\": 524,\n
> >>>> \"clientInfo\": \"\",\n      \"host\": \"10.1.11.3\",\n
> >> \"port\":
> >>>> 5906,\n      \"tag\": \"\",\n      \"createTime\":
> 1341997999371,\n
> >>>> \"lastUsedTime\": 1341
> >>>> 997999371\n    },\n    {\n      \"id\": 8,\n      \"clientInfo\":
> >>>> \"\",\n      \"host\": \"10.1.11.4\",\n      \"port\": 5902,\n
> >>>> \"tag\": \"c43d83ff-35e0-4eb6-87b4-d56f7a2929f9\",\n
> >>>> \"createTime\": 1341904048748,\n      \"las
> >>>> tUsedTime\": 1341998035188\n    }\n  ]\n}","wait":0}}] }
> >>>> 2012-07-11 03:13:58,857 DEBUG [agent.manager.AgentManagerImpl]
> >>>> (AgentManager-Handler-15:null) SeqA 3-187783: Sending Seq 3-187783:
> >>>> { Ans: , MgmtId: 74142025330, via: 3, Ver: v1, Flags: 100010,
> >>>> [{"AgentControlAnswer":{"result":true,"wa
> >>>> it":0}}] }
> >>>> 2012-07-11 03:13:59,133 INFO  [api.commands.DeployVMCmd] (Job-
> >> Executor-
> >>>> 2:job-52) com.cloud.exception.InsufficientServerCapacityException:
> >>>> Unable to create a deployment for VM[User|i-2-18-
> VM]Scope=interface
> >>>> com.cloud.dc.DataCenter; id=2
> >>>> 2012-07-11 03:13:59,134 WARN  [cloud.api.ApiDispatcher] (Job-
> >> Executor-
> >>>> 2:job-52) class com.cloud.api.ServerApiException : Unable to
> create
> >> a
> >>>> deployment for VM[User|i-2-18-VM]
> >>>> -------------
> >>>>
> >>>>
> >>>> So after looking for the exact same thing "Looking for
> >>>> speed=500Mhz, Ram=512" in both cases, the template passes (or
> >>>> atleast proceeds to
> >> do
> >>>> further checks, and eventually passing) and the ISO immediately
> >> fails.
> >>>> Is there something I'm doing wrong?  Is this a bug that needs to
> be
> >>>> patched or looked in to?  Or, is there another way to create a
> >> template
> >>>> other than deploying with an ISO and creating a template off that
> >>>> deployment?
> >>>>
> >>>> Environment info:
> >>>>
> >>>> Cloudstack 3.0.2
> >>>> ISO Tried - CentOS 6.2 minimal (both x86 and x64), Ubuntu Server
> >> 12.04,
> >>>> CentOS 5.8 Netinstall
> >>>>
> >>>> Thanks,
> >>>> Caleb
> >>>
> >


Mime
View raw message