Description of problem: I was trying submit VM (kvm) job from cumin UI and it doesn't work. Job got into held state with reason: HoldReasonError from slot1@server-name: VMGAHP_ERR_JOBCLASSAD_XEN_NO_DISK_PARAM This is log from cumin's web.log: 15786 2011-12-09 15:37:23,652 INFO Request GET /index.update?session=index.html%3Fframe%3Dmain.grid%3Bmain.m%3Dgrid%3Bmain.grid.id%3D5%3Bmain.grid.view.body.m%3Dpool_submissions;widget=main.tasks; widget=main.grid.view.body.pool_submissions.table 15786 2011-12-09 15:37:23,690 INFO Response 200 OK 15786 2011-12-09 15:37:23,691 DEBUG Response headers: 15786 2011-12-09 15:37:23,692 DEBUG Content-Length 6753 15786 2011-12-09 15:37:23,692 DEBUG Content-Type text/xml 15786 2011-12-09 15:37:23,693 DEBUG Cache-Control no-cache 15786 2011-12-09 15:37:23,994 INFO Request POST /form.html? 15786 2011-12-09 15:37:24,005 DEBUG Validating cumin.grid.submission.VmJobSubmitForm('modes.VmJobSubmit') 15786 2011-12-09 15:37:24,015 DEBUG Starting cumin.grid.submission.VmJobSubmit 15786 2011-12-09 15:37:24,016 INFO Started cumin.grid.submission.VmJobSubmit 15786 2011-12-09 15:37:24,017 DEBUG Job ad: 15786 2011-12-09 15:37:24,017 DEBUG !!descriptors {'RequestMemory': 'com.redhat.grid.Expression', 'Requirements': 'com.redhat.grid.Expression'} 15786 2011-12-09 15:37:24,018 DEBUG Cmd '/var/lib/libvirt/images/testvm.img' 15786 2011-12-09 15:37:24,018 DEBUG DiskUsage 0 15786 2011-12-09 15:37:24,019 DEBUG Iwd '/tmp' 15786 2011-12-09 15:37:24,019 DEBUG JobUniverse 13 15786 2011-12-09 15:37:24,020 DEBUG JobVMCheckpoint False 15786 2011-12-09 15:37:24,021 DEBUG JobVMMemory 512 15786 2011-12-09 15:37:24,021 DEBUG JobVMNetworking False 15786 2011-12-09 15:37:24,022 DEBUG JobVMType 'kvm' 15786 2011-12-09 15:37:24,022 DEBUG JobVM_VCPUS 1 15786 2011-12-09 15:37:24,023 DEBUG Owner 'cumin' 15786 2011-12-09 15:37:24,023 DEBUG RequestDisk 5242880 15786 2011-12-09 15:37:24,024 DEBUG RequestMemory 'ceiling(ifThenElse(JobVMMemory =!= undefined,JobVMMemory, ImageSize / 1024.000000))' 15786 2011-12-09 15:37:24,024 DEBUG Requirements 'VM_Type == "KVM" && Arch == "X86_64" && HasVM && VM_AvailNum > 0 && TotalDisk >= DiskUsage && TotalMemory >= 512 && VM_Memory >= 512' 15786 2011-12-09 15:37:24,025 DEBUG ShouldTransferFiles 'NEVER' 15786 2011-12-09 15:37:24,025 DEBUG Submission 'pokus01' 15786 2011-12-09 15:37:24,026 DEBUG VMPARAM_Kvm_Disk '/var/lib/libvirt/images/testvm.img:vda:w' 15786 2011-12-09 15:37:24,595 DEBUG Exiting cumin.grid.submission.VmJobSubmit 15786 2011-12-09 15:37:24,596 INFO Exited cumin.grid.submission.VmJobSubmit 15786 2011-12-09 15:37:24,597 INFO Response 303 See Other 15786 2011-12-09 15:37:24,598 DEBUG Response headers: 15786 2011-12-09 15:37:24,598 DEBUG Location index.html?frame=main.grid;main.m=grid;main.grid.id=5;main.grid.view.body.m=pool_submissions 15786 2011-12-09 15:37:24,663 DEBUG Method response for request 1323441368 received from Broker connected at: server-name:5672 15786 2011-12-09 15:37:24,664 DEBUG Response: OK (0) - {u'Id': 'server-name#13.0'} 15786 2011-12-09 15:37:24,664 DEBUG Ending cumin.grid.submission.VmJobSubmit 15786 2011-12-09 15:37:24,665 INFO Ended cumin.grid.submission.VmJobSubmit Version-Release number of selected component (if applicable): cumin-0.1.5098-2.el5.noarch How reproducible: 100% Steps to Reproduce: 1. Add broker configuration into cumin.conf and start cumin 2. Try to submit VM job with provided VM path from server administrator 3. See job status/details Actual results: Job isn't running Expected results: Job is running Additional info:
VMPARAM_Kvm_Disk was renamed as VMPARAM_vm_Disk at sometime before 2.0.
I think I need some more info to go on here. I have not yet tried anything with a VM job. Is there any special condor setup that I'll need to reproduce this properly? Any chance that anyone has a file that I could run via condor_submit? Where can I find the img file described in /var/lib/libvirt/images/testvm.img? Better yet, what steps would I need to do to create an img of my own? Based on Luigi's comment #1, I think this may wind-up being a small fix, but it seems as though I will have some learning to do to get to the point of fixing it. Thanks
(In reply to comment #2) > I think I need some more info to go on here. I have not yet tried anything > with a VM job. Is there any special condor setup that I'll need to reproduce > this properly? Basic configuration for virtualization support as documented here: http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_MRG/2/html/Grid_User_Guide/chap-Grid_User_Guide-The_Virtual_Machine_Universe.html > Any chance that anyone has a file that I could run via condor_submit? Something like: -------------------- Universe=vm Log=log.$(cluster) Executable=testvm VM_TYPE=kvm VM_MEMORY=512 VM_DISK=/var/lib/libvirt/images/testvm.job:vda:w Queue -------------------- > Where can I find the img file described in /var/lib/libvirt/images/testvm.img? > Better yet, what steps would I need to do to create an img of my own? Any valid (raw or qcow2) image will work. virt-manager or simply virt-install can help to install a machine.
Fixed in revision 5178 on trunk.
Technical note added. If any revisions are required, please edit the "Technical Notes" field accordingly. All revisions will be proofread by the Engineering Content Services team. New Contents: Consequence: Submitting a VM job from within cumin was appearing to succeed in cumin, but failing to start the VM. Cause: The job classad produced by cumin was slightly outdated and still using VMPARAM_Kvm_Disk instead of VMPARAM_vm_Disk. Fix: Cumin now builds the job classad with VMPARAM_vm_Disk Result: VM jobs submitted from within cumin are working again.
I was able to test cumin-0.1.5184-1.el6.noarch condor-vm-gahp-7.6.5-0.11.el6.x86_64 And it was working with submitting KVM guest job from cumin without extra parameters. Need to test on rhel 5/6, i386/x86_64 + XEN+KVM where supported.
RHEL5 i386, cumin-0.1.5184-1.el5.noarch KVM job OK notes: I filled only job description and vm image location: /var/lib/libvirt/images/testvm.img XEN job OK notes: job description and vm image location: /var/lib/xen/images/testvm.img Probably I hit Bug 765894 So I needed to add extra params: Requirements = True JobVMType = xen VMPARAM_Xen_Kernel = included VMPARAM_vm_Disk = /var/lib/xen/images/testvm.img:xvda:w
RHEL5 x86_64, cumin-0.1.5184-1.el5.noarch RHEL6 i386, cumin-0.1.5184-1.el6.noarch RHEL6 x86_64, cumin-0.1.5184-1.el6.noarch KVM job OK notes: I filled only job description and vm image location: /var/lib/libvirt/images/testvm.img XEN job OK notes: job description and vm image location: /var/lib/xen/images/testvm.img Probably I hit Bug 765894 So I needed to add extra params: Requirements = True JobVMType = xen VMPARAM_Xen_Kernel = included VMPARAM_vm_Disk = /var/lib/xen/images/testvm.img:xvda:w I also have a lot of "errors" in cumin web.log - Bug 782839 There were apparently only information from cumin that there were changes in the condor node. Web interface wasn't affected.
(In reply to comment #10) > I was able to test > cumin-0.1.5184-1.el6.noarch > condor-vm-gahp-7.6.5-0.11.el6.x86_64 > > And it was working with submitting KVM guest job from cumin without extra > parameters. > > Need to test on rhel 5/6, i386/x86_64 + XEN+KVM where supported. I hit during this test on my laptop Bug 782054 . I did proper verification ( comment 11 and comment 15 ) and there was no such issue.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. http://rhn.redhat.com/errata/RHSA-2012-0100.html