[Condor-users] Cannot start A VM

[Daniele Fetoni <daniele.fetoni@xxxxxxxxxx>]

Hi,

I am trying to start a VM with condor 7.5.2 using kvm as Hypervysor.
After ome troubles making VM universe start, I manage to submit successfully a VM job, but the VM doesn't start.
I found this error in StarterLog.slot1

******************************************************
06/24/10 09:40:04 ** condor_starter (CONDOR_STARTER) STARTING UP
06/24/10 09:40:04 ** /opt/condor-7.5.2/sbin/condor_starter
06/24/10 09:40:04 ** SubsystemInfo: name=STARTER type=STARTER(8) class=DAEMON(1)
06/24/10 09:40:04 ** Configuration: subsystem:STARTER local:<NONE> class:DAEMON
06/24/10 09:40:04 ** $CondorVersion: 7.5.2 Apr 20 2010 BuildID: 232940 $
06/24/10 09:40:04 ** $CondorPlatform: X86_64-LINUX_RHEL5 $
06/24/10 09:40:04 ** PID = 5451
06/24/10 09:40:04 ** Log last touched 6/24 09:38:31
06/24/10 09:40:04 ******************************************************
06/24/10 09:40:04 Using config source: /opt/condor-7.5.2/etc/condor_config
06/24/10 09:40:04 Using local config sources:
06/24/10 09:40:04    /opt/condor-7.5.2/local.Black/condor_config.local
06/24/10 09:40:04 DaemonCore: command socket at <192.168.1.37:34071>
06/24/10 09:40:04 Done setting resource limits
06/24/10 09:40:04 Communicating with shadow <192.168.1.37:35928>
06/24/10 09:40:04 Submitting machine is "Black"
06/24/10 09:40:04 setting the orig job name in starter
06/24/10 09:40:04 setting the orig job iwd in starter
06/24/10 09:40:04 File transfer completed successfully.
06/24/10 09:40:05 Job 1.0 set to execute immediately
06/24/10 09:40:05 Starting a VM universe job with ID: 1.0
06/24/10 09:40:05 About to start new VM
06/24/10 09:40:25 About to exec /opt/condor-7.5.2/sbin/condor_vm-gahp -f -M 3
06/24/10 09:40:25 VMGAHP server pid=5466
06/24/10 09:40:32 VMGAHP write line(RESULTS) Error
06/24/10 09:40:32 Failed to create a new VM
06/24/10 09:40:33 VMGAHP write line(QUIT) Error
06/24/10 09:40:34 Failed to start job, exiting
06/24/10 09:40:34 ShutdownFast all jobs.
06/24/10 09:40:34 **** condor_starter (condor_STARTER) pid 5451 EXITING WITH STATUS 0

And this is ShadowLog

06/24/10 10:04:29 ******************************************************
06/24/10 10:04:29 ** condor_shadow (CONDOR_SHADOW) STARTING UP
06/24/10 10:04:29 ** /opt/condor-7.5.2/sbin/condor_shadow
06/24/10 10:04:29 ** SubsystemInfo: name=SHADOW type=SHADOW(6) class=DAEMON(1)
06/24/10 10:04:29 ** Configuration: subsystem:SHADOW local:<NONE> class:DAEMON
06/24/10 10:04:29 ** $CondorVersion: 7.5.2 Apr 20 2010 BuildID: 232940 $
06/24/10 10:04:29 ** $CondorPlatform: X86_64-LINUX_RHEL5 $
06/24/10 10:04:29 ** PID = 7144
06/24/10 10:04:29 ** Log last touched 6/24 10:04:29
06/24/10 10:04:29 ******************************************************
06/24/10 10:04:29 Using config source: /opt/condor-7.5.2/etc/condor_config
06/24/10 10:04:29 Using local config sources:
06/24/10 10:04:29    /opt/condor-7.5.2/local.Black/condor_config.local
06/24/10 10:04:29 DaemonCore: command socket at <192.168.1.37:44994>
06/24/10 10:04:29 Initializing a VM shadow for job 2.0
06/24/10 10:04:29 (2.0) (7144): Request to run on slot2@Black <192.168.1.37:57569> was REFUSED
06/24/10 10:04:29 (2.0) (7144): Job 2.0 is being evicted from slot2@Black
06/24/10 10:04:29 (1.0) (7143): **** condor_shadow (condor_SHADOW) pid 7143 EXITING WITH STATUS 108
06/24/10 10:04:29 (2.0) (7144): logEvictEvent with unknown reason (108), aborting
06/24/10 10:04:29 (2.0) (7144): **** condor_shadow (condor_SHADOW) pid 7144 EXITING WITH STATUS 108


I cannot unterstand where is the error in VMGAHP; moreover if I run the command /opt/condor-7.5.2/sbin/condor_vm-gahp -f -M 0 vmtype "kvm", it works properly.
I can start a VM via qemu-kvm, so there should not be problems with hypervisor.

What could be the problem? Any suggestion?

Thanks in advance

Daniele

---

Messenger Radio. La radio che si fa cliccare!