Oracle® VM Manager User's Guide Release 2.1 Part Number E10901-04 |
|
|
View PDF |
This appendix describes some problems you may encounter when using Oracle VM Manager, and explains how to resolve them. It includes the following topics:
Find additional information on the following Oracle support-oriented Web sites:
Oracle MetaLink, available at http://metalink.oracle.com
Oracle Virtualization Forum, available at http://forums.oracle.com/forums/forum.jspa?forumID=482
Oracle VM Manager log files are stored in the directory:
/var/log/ovm-manager/
Table D-1, "Log Files" lists the log files it contains.
Table D-1 Log Files
Log File | Description |
---|---|
ovm-manager.log |
The Oracle VM Manager installation log. |
db.log |
The Oracle Database log. When you install Oracle VM Manager on an existing database, the log information is stored here. |
oc4j.log |
The Oracle Containers for J2EE (OC4J) installation log. When oc4j.log exceeds 10 MB, a new log file oc4j.log.1. is generated to store the logs in oc4j.log. Subsequently, the oc4j.log is cleared to record new log information. |
upgrade_oldversion_newversion.log.log |
The Oracle VM Manager upgrading log. |
It may take a very long time to log in, or it is stuck in the Login page. This may be caused by lack of memory for OC4J.
To solve this problem:
Log in to the computer where Oracle VM Manager is installed, and run the following command to check the log information:
# cat /var/log/ovm-manager/oc4j.log | grep "heap"
If OC4J runs out of memory, the following information is displayed:
Internal Exception: java.lang.OutOfMemoryError: Java heap space
Run the following command to restart OC4J:
# service oc4j stop # service oc4j start
If OC4J runs out of memory, you need to increase the OC4J memory size. Follow these steps to do it:
View the OC4J configuration information:
# vi /opt/oc4j/bin/oc4j
Locate the following line, and increase the memory size to an appropriate value, such as 512:
OC4J_JVM_ARGS="-XX:PermSize=256m -XX:MaxPermSize=512m"
Restart OC4J:
# service oc4j stop # service oc4j start
The following message is displayed: "Error: There is no server supporting hardware virtualization in the selected server pool. "
To solve this problem, ensure that the Virtual Machine Server supports hardware virtualized machine (HVM).
Follow these steps to check:
Run the following command to check if HVM is supported by the CPU:
# cat /proc/cpuinfo |grep -E 'vmx|smx'
If any information that contains vmx
or smx
is displayed, it means that the CPU supports HVM. Here is an example of the returned message:
flags : fpu tsc msr pae mce cx8 apic mtrr mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx lm constant_tsc pni monitor ds_cpl vmx est tm2 cx16 xtpr lahf_lm
Ensure that you have enabled HVM in the BIOS.
Run the following command to check if the operating system supports HVM:
# xm info |grep hvm
The following is an example of the returned message:
xen_caps : xen-3.0-x86_64 xen-3.0-x86_32p hvm-3.0-x86_32 hvm-3.0-x
If the CPU does not support HVM, use the paravirtualized method to create the virtual machine. See Creating a New Virtual Machine from Installation Media.
A message similar to the following is displayed:"Error: The largest virtual disk that can be allocated is: 4815 MB."
To solve this problem, decrease the disk size, or add a new repository.
Adding a New Repository
Before adding a new repository, ensure that all the Oracle VM Servers in the same server pool share the same repository, and mount it to the /OVS folder on each server.
To add a new repository:
Run the following command on Virtual Machine Server:
# /usr/lib/ovs/ovs-makerepo source shared description
The source
parameter is the block device or NFS path to the file system to be added. The shared
parameter sets whether the file system is to be shared between computers. Enter a 1
to have the file system shared, or a 0
to make it non-shared. The description
parameter is the text displayed in Oracle VM Manager.
For example, run the following command:
# /usr/lib/ovs/ovs-makerepo /dev/sdc1 1 A new repository
You will get one repository in /etc/ovs/repositories:
# cat /etc/ovs/repositories # This configuration file was generated by ovs-makerepo # DO NOT EDIT C860A37B7E4D437A93CA4116A79BD9C8 /dev/sdc1
The makerepo
script identifies the file system or shared virtual disk as a repository, and updates the repository configuration to enable it.
Use the df command to review the usage and mount point:
# df
The following is an example of the returned information:
Filesystem 1K-blocks Used Available Use% Mounted on /dev/sda1 9920592 848376 8560148 10% / tmpfs 262232 0 262232 0% /dev/shm /dev/sdb1 234380284 161920232 72460052 70% /OVS /dev/sdc1 292977372 5167844 287809528 2% /OVS/C860A37B7E4D437A93CA4116A79BD9C8
The /OVS/seed_pool, /OVS/running_pool, and /OVS/iso_pool directories will be created automatically under the newly added repository.
Repeat the preceding steps on all Virtual Machine Servers to make sure the added repository has mounted correctly on each server.
When you create a virtual machine, Oracle VM Manager will choose a Virtual Machine Server with the maximum available disk size from /OVS or /OVS/uuid to run it.
For Oracle VM Manager Release 2.1.1 and Release 2.1.2 Users
To solve this problem, reset the status to Powered Off, and then delete the virtual machine directly. See Deleting a Virtual Machine Stuck in a Certain Status for more information.
For Oracle VM Manager Release 2.1 Users
To solve this problem, change the status manually as follows:
Make sure that the virtual machine in the target server has been shut down. Use xm list or virsh list to check the virtual machine name and ID.
If the virtual machine is running, use xm shutdown vm_id or virsh shutdown vm_id to shut it down, where vm_id refers to the virtual machine name or ID.
Use xm list or virsh list to check if the virtual machine has been shut down. If it is still running, use xm destroy vm_id, or virsh destroy vm_id to shut it down.
Log in as oracle (or root first, then su - oracle) to the computer where Oracle VM Manager is installed, and run the following commands to change the virtual machine status to Powered off
in Oracle XE database:
export ORACLE_HOME='/usr/lib/oracle/xe/app/oracle/product/10.2.0/server'
export ORACLE_SID=XE
$ORACLE_HOME/bin/sqlplus / as sysdba
SQL> update ovs.ovs_vm_img t set t.status='Powered Off' where t.img_name like 'vm_name';
SQL> commit;
SQL> quit
vm_name refers to the virtual machine name.
Retry your operation, or delete the virtual machine directly.
To view more details, click the Virtual Machines tab, and then click the Show link of the virtual machine. Click the number after the Log item to display the error log. The error log information includes: operation, operation details, start time and status.
The following are some error log items and solutions:
Error: HVM guest support is unavailable: is VT/AMD-V supported by your CPU and enabled in your BIOS?
Failed:<Exception: cannot find host server for vm('/OVS/running_pool/vm_name')>
Failed:<Exception: Could not find '/OVS/running_pool/vm_name' in any storage repository.>
Failed:<Exception: vm('/OVS/running_pool/vm_name') status ('RUNNING') not in ('DOWN', 'ERROR')>
Failed:<Exception: no server selected to run vm('/OVS/running_pool/vm_name') memory=512>
Possible Cause
The Virtual Machine Server does not support HVM.
Solution
Verify that the Virtual Machine Server supports HVM, and has HVM enabled in BIOS. See Appendix D, "Cannot Create a Virtual Machine from Installation Media".
If the Virtual Machine Server does not support HVM, you can create the virtual machine from installation media.
Possible Cause
No Virtual Machine Server available to run the virtual machine.
Insufficient memory on the Virtual Machine Server.
Solution
Ensure that at least one Virtual Machine Server is available to run the virtual machine.
Ensure that the Virtual Machine Server has sufficient memory.
To release server memory, you can:
Shut down other running virtual machines.
Decrease the virtual machine memory, and restart the virtual machine. See Editing Virtual Machine Configuration.
Possible Cause
The virtual machine has been removed from the Virtual Machine Server.
Solution
Remove the virtual machine in Oracle VM Manager, and create a new one.
Possible Cause
You have entered an incorrect user name or password when creating the Utility Server.
Solution
Enter correct user name and password for the Utility Server. See Creating a Server Pool.
Possible Cause
The virtual machine has been started already.
Solution
Use the command xm list to check the virtual machine status. Refresh the virtual machine periodically.
Possible Cause
The status in Oracle VM Manager has not been synchronized with the status data in the database.
Solution
Shut down the virtual machine, and then start it.
Possible Cause
There is insufficient memory on this Virtual Machine Server.
Solution
Decrease the virtual machine memory, and then restart the virtual machine. Or shut down other running virtual machines to release more memory.
Possible Cause
No bridge for this virtual machine.
Solution
Select a bridge for it. See Network.
Possible Cause
Invalid VIF type for the virtual machine.
Solution
When starting a hardware virtualized machine through PXE, the VIF type must be Fully Virtualized (ioemu). Select a valid VIF type for the virtual machine. See Network.
Possible Cause
The resource location you entered is incorrect.
Solution
Delete the virtual machine, and create a new one. Make sure you enter a correct resource location when creating a paravirtualized machine.
If your operating system is Linux, and you are using Mozilla Firefox to access the virtual machine, download and install the Console Plug-in. See Installing the Console Plug-In.
If Oracle VM Manager is upgraded from the Release 2.1, rather than a fresh installation, you may encounter the following error when accessing the virtual machine console:
java.lang.ClassNotFoundException: VncViewer.class
This is caused by the delay in cache refresh. To solve this problem, shut down all the Web pages, and open a new one to access the virtual machine console.
The following message is displayed:
Error: Server is not ready for live migration.
To solve this problem, you must use identical computers to perform live migrations, that is, the computer make and model number must be identical.
To change the CD in a virtual machine:
Unmount the first CD:
# umount mount-point
Select the second ISO file, and click Change CD.
Mount the second CD:
# mount /dev/cdrom mount-point
Cannot Import Due to Invalid URL
The following information is displayed: "The URL is invalid", or "The proxy URL is invalid".
To solve this problem, log in to the Virtual Machine Server, and run the following command to test the network connection:
# wget url
Here, url refers to the link for downloading the resource.
If the resource is from the Internet, you may need a proxy before running the wget command. Contact your network administrator to get the proxy, an then run the following commands to configure the proxy:
# export http_proxy=http://host:port/ # export ftp_proxy=http://host:port/
Here, host refers to the host name or IP address of the proxy server, and port refers to the port number of the proxy.
Cannot Import Due to NFS Disk
The following information is displayed:
Cannot obtain memory size from vm.cfg.
This error may occur when there are two or more Utility Servers in a server pool, and the disk is shared by the Network File System (NFS) mode.
To solve this problem, click the Next button periodically.
If you reinstall the Oracle VM Server software, or change the root user's password of an Oracle VM Server in a server pool, communication from the server pool to that server may fail.
If you reinstall the Oracle VM Server software (not upgraded), the RSA key is likely to change, even if you use the same IP address and host name. You may see an error similar to:
REMOTE HOST IDENTIFICATION HAS CHANGED
or
Host key verification failed
Workaround: Remove the Oracle VM Server from the server pool, and add the Oracle VM Server again.