Under OVMM, Server Pools-> , The status is showing "Running (Error)". A green right arrow with a Red 'X'.
Not sure where can get more informative logs to troubleshoot further. This is a pullback. Nothing has changed.
Your post is a little confusing.
Is the OVMM server physical or virtual?
Are you removing the VM server from the manager and then rediscovering it? Once a VM server is discovered the VM manager constantly maintains contact with the server to determine its status. You need to check the ovs agent log.
There is a known problem with 3.1.1 in which the rpm database gets corrupt for no apparent reason. When the agent goes to check status of the host it queries for installed packages and throws an error that causes the agent to believe the server is unavailable. Had it happen to me several times. You basically have to delete the rpm database and recreate it. There is MOS article on the problem. I can't remember the exact error you receive but it will be in the /var/log/ovs-agent log.
OVMM server is virtual. It is running on top of Virtualbox. I imported the Oracle VM Template.
I can't remove the VM Server as the server pool doesn't allow. It says the OVM Server is in error.
However I am able to rediscover it again and again.
Can I upgrade the OVM server to 3.2.2 then? Any good link for this pls?
I will check out in metalink meanwhile.
OK, I have checked in metalink. In the /var/log/ovs-agent.log I don't see any rpm db corrupted error. The log file seems ok.
However, I just rebooter OVM Server. the status is showing as Running(Error), two red color icon besides the OVM server name (ovm-svr.oow.com).
Seems not easy to troubleshoot.
I've never seen this before but it sounds like your VM server has the wrong uuid compared what is in the database. You need to get the uuid of what is referenced in the database and compare it to what is on the server....There is a way to fake the uuid if they are different.
If you go under the "Events" section of the VM server, what do you see?
Or. If you only have one server.... start over.
Just found out that one requires a CSI with Oracle VM contract prior able to use yum install for the Oracle Server. Why they allow OVMM upgrade using script downloaded from metalink but Oracle Server they use yum method which requires a registration?
I am using OVM for testing. And OVM Server 3.1.1 is not very stable. I strongly believe is a bug because I didn't change anything. Just shutdown my Host OS, and upon startup OVMM just won't talk to OVM Server despite it can discover OVM Server.
Why would the uuid changed? I need to verify this.
Anyway, I went to events --> Under Type
Summary: Server Pool Master Error
Okay. You have your answer. The server can't connect to the Master Pool.
My issue is pretty similar to what they encountered in this post
I tried to fakeuuid in my OVM server. But this doesn't make much different. In the first place, OVMM is able to discover OVM Server.
Like the event showed, it is because of the Master Pool. I went to Master Pool and clicked edit. I found that the UUID is different from the OVM Server UUID that OVMM discovered.
Question is why does the UUID changed? The explanation was it in the post One of my two VM Server 3.1 nodes is completely hosed. How do I recover it?
How can I fix the UUID in the Master Pool? Shall I recreate the master pool? And almost everything again. The OVM server and guest-VM?
Found that it is a bug that UUID gets changed in OVM 3.1.1. Really Oracle should update the VM template downloadable to a more stable release for a better experience. Really hampered testing using a buggy version. Much time spend to find this.
Notable bugs fixed in Oracle VM Server 3.2.2-520 and Oracle VM Manager 3.2.2-520 are:
- Fixes an upgrade issue to return the correct UUID when discovering an Oracle VM Server with an SMBIOS newer than version 2.6.
I've sen the UUID change from upgrading a BIOS version. The UUID is generated using information from the local hardware.