PowerOn Failed : Result - failed:errcode=20000, errmsg=No enough memory

Question

Hi,
When I tried to start a VM from ovm manager I get this message 'PowerOn Failed : Result - failed:errcode=20000, errmsg=No enough memory to run vm'
As per this document (https://communities.oracle.com/portal/server.pt/community/view_document/217?docID=198299) I have tried these commands to increase memory
[root@ovm-rac-exp01 ~]# xm info | grep -E '^free_memory'
free_memory : 15590
[root@ovm-rac-exp01 ~]# /opt/ovs-agent-2.3/db/db_kv.py srv ovm-rac-exp01 -t int -f mem_free 15590
old: ovm-rac-exp01 => 0
new: ovm-rac-exp01 => 15590
but after a few seconds mem_free returns to 0
How can I set the mem_free parameter ?
Thanks

Answer 1

I am experiencing the same behavior.
* The OVM server (2.2.2) in my test lab has 8GB physical memory with 563MB assigned to Dom0 per GRUB config: kernel /xen-64bit.gz dom0_mem=563M
* When attempting to start a single Windows 2008 (HVM) VM, allocated 1024MB, and with no other guest VM on the system I see the following in /var/log/ovs-agent/ovs_root.log:
"2011-10-18 12:35:28" DEBUG=> select_tgt_host: vm('/OVS/running_pool/W2K8_SVR01')=>memory=1024
"2011-10-18 12:35:28" ERROR=> select_tgt_host: faild. => errcode=20000, errmsg=No enough memory to run vm(/OVS/running_pool/W2K8_SVR01), required memory=1024.
StackTrace:
File "/opt/ovs-agent-2.3/OVSServerSelection.py", line 336, in select_tgt_host
raise OVSException(ERR_NO_MEMORY, {"vm_path": vm_path, "memory": memory})
"2011-10-18 12:35:28" ERROR=> start_vm: failed! vm('/OVS/running_pool/W2K8_SVR01') errcode=20000, errmsg=No enough memory to run vm(/OVS/running_pool/W2K8_SVR01), required memory=1024.
StackTrace:
File "/opt/ovs-agent-2.3/OVSSiteVM.py", line 124, in start_vm
raise e
"2011-10-18 12:35:28" INFO=> OVSPolicyServer.execute_policy(): vm('/OVS/running_pool/W2K8_SVR01'), action=start_vm action_params=(vm_path, ) => failed:errcode=20000, errmsg=No enough memory to run vm(/OVS/running_pool/W2K8_SVR01), required memory=1024.
StackTrace:
File "/opt/ovs-agent-2.3/OVSSiteVM.py", line 124, in start_vm
raise e
* There seems to be plenty of memory available:
# xm info | grep memory
total_memory : 8156
free_memory : 7490
node_to_memory : node0:7490
* Starting the VM manually using 'xm create' works:
# xm create /var/ovs/mount/C974E8A5DEFF4FD18CBD8D74BE9092D3/running_pool/W2K8_SVR01/vm.cfg
Using config file "/var/ovs/mount/C974E8A5DEFF4FD18CBD8D74BE9092D3/running_pool/W2K8_SVR01/vm.cfg".
Started domain W2K8_SVR01 (id=2)
* Memory usage seems as expected:
# xm list
Name ID Mem VCPUs State Time(s)
Domain-0 0 563 4 r..... 392.1
W2K8_SVR01 2 1024 1 ...... 1.2
# xm info | grep memory
total_memory : 8156
free_memory : 6453
node_to_memory : node0:6453
* Trying the procedure suggested here ( https://communities.oracle.com/portal/server.pt/community/view_document/217?docID=198299 ) is of no use as 'mem_free' reverts to value of zero within a few seconds. It seems the agent is re-reading what it thinks is total available memory (from where?) and updating the database. I tried stopping the agent, changing 'mem_free' value and then restarting the agent but it makes no difference. Not sure procedure in docID=198299 is supported anyway.
* Sorry I can't help much but at least you are not alone in experiencing this issue...