There is xenfs entry in fstab.
Remount is not successful:
# mount -a
mount: mount point /proc/xen does not exist
The is no xenfs message in /var/log/messages, but a lot of:
/sbin/mingetty: hvc0: No such file or directory
Looks like you discovered an ISO file bug.
Your quickest path to resolution is a dom0 rebuild, or Oracle support :-)
It would be great if you tried the 3.2.6 ISO for the rebuild and check back with us to see if the fresh install is ok, if it has the same bug, next do a 3.2.2 build and patch it using the public latest repo to get up-to 3.2.6..
I can confirm that the xend problem does not occur with 3.1 and 3.2 to 3.2.6 yum ugrades.
And could you tell me, what is the best way to rebuild dom0, not to lost my repository and VM's?
If I choose Reinstall system after boot iso, it wants to remove partitions.
Do I have to backup vm images and then import them to reinstalled VM server or the repository will be ok (it's on different storage - SAS storage used as local disk).
Hi, I am facing the same exact problem after upgrading from 3.2.2 to 3.2.6 using the ISO.
Let me try doing a fresh install of 3.2.6 for one of the vm servers and see if this fixes the issue.
Actually on reinstalling the vm server with 3.2.6 all services including xend and xencommons are running normally. unfortunately i am not able to remove the previously upgraded (from 3.2.2 to 3.2.6) virtual machine servers from the pool.
Something is wrong with the ISO upgrade process. It breaks the system.
Is there any other way i can cleanly remove the ISO upgraded (3.2.2 to 3.2.6) servers from the pool ?
I have tried to remove them them using vmanager and it is not successful . I keep getting error messages related to zend not running on the hosts.
Q: Is there any other way i can cleanly remove the ISO upgraded (3.2.2 to 3.2.6) servers from the pool ?
A: Unfortunately, there is now way to revert that I am aware of. I would be interested to see what O support says about "a way to revert an ISO upgrade".
Q; I have tried to remove them them using vmanager and it is not successful .
A: You will need to reset the Oracle VM Manager DB and rediscover :-) LOL still not fixed with 3.2.6 grgrgrgrgrgrg!
If you have the Oracle VM 3.2.4 ISO with a 3.2.4 yum repo, I would recommend moving forward with 3.2.4, its a GREAT release, or revert back to 188.8.131.525, its ROCK solid!
Unfortunately post Oracle VM Release 3.2.1, Oracle stopped maintaining the Oracle VM ULN repositories and desired to standardize on Oracle VM Latest. Using latest is always VERY exciting, albeit not very productive :-) Lets hope that we see the missing Oracle VM repos soon!
Thank you Roddy, i greatly appreciate your efforts to assist.
I did not get hold of Oracle VM 3.2.4 ISO, it was not made available in the Software Delivery cloud.
For now, i may have to reset the Oracle VM Manager DB and rediscover.
I am glad I could provide assistance!
I missed this post: "Actually on reinstalling the vm server with 3.2.6 all services including xend and xencommons are running normally. unfortunately i am not able to remove the previously upgraded (from 3.2.2 to 3.2.6) virtual machine servers from the pool."
During the re-install, the installer will pop up a Warning screen that will display each disk, partition and/or multipath device. When you see the Warning screen, STOP! You need to know which disk and/or partition is being used for the OS/dom0, and "only" that disk should be initialized, all the other disks, partitions and/or multipath devices should "NOT" be initialized. initialize = delete! As long as you do not initialize the disks, partitions, and/or mpath devices with the VMs, the VMs on the partitions can be re-used.
Note: Since Oracle stopped maintaining the Oracle VM ULN repositories, i.e. post 3.2.1, it necessary for us to maintain local RPM repositories for each release, (3.2.2, 3.2.4, etc....), or risk being forced to use / upgrade to the latest release, which can be very exciting but not a good use of our time. Try to always stay on the latest stable Oracle VM release, at least one dot version / full release back.
Noted. Thank you.
I need to ask you something not related to OVM. I hope it is ok with you.
Do you know of someone reliable to work with inside Oracle on product related matters ? If you do and it is OK with you to share their contact, please assist me with a contact.
Your best bet is to request access to the desired oracle product manager (PM) via your sales rep, or ask your sales rep for your sales engineer's (SE) contact details. PMs and SEs should be able to provide the info/support your looking for. If your organization is international, or has multiple locations, try to get local sales consulting contacts from each geo. I hope this help!
There is one that i keep getting referred to in Nairobi and he never responds to email nor does he call back after missing a call.
I am located in Uganda and we are supposed to be supported by the Nairobi office but getting any kind of help is close to impossible and yet the potential for business here is high.
The option Reinstall system from the same iso is working.
Now I have running VM server and I try to recover previous repository, network, etc...
Is there a simple solution or I have to create everything from VM Manager again?
You will need to do everything from VM Manager. If your using local disks, be VERY careful not to overwrite/recreate the repository wich will delete all the VMs :-) With local disks, post unclustered pool setup in Oracle VM Manager, refresh from the repositories tab, and the local repo and VMs should be there.
Note: When we recovering single servers installs, before the re-install backup the /etc/sysconfig/network-scripts directory, post install (before discovery) restore the /etc/sysconfig/network-scripts directory, restart networking or reboot, and when you discover the host in Oracle VM Manager the networking will setup.
This blog post has recovery details: Oracle VM Manager Installation and Recovery