Booting OVM server from multipath SAN

Hi,
I read in the release notes of OVM 3.0.3 that "booting from multipath SAN is not supported". I have a limitation that forces me to use SAN boot.
What is the best practise?
1. Go for local disks for OVM server
or
2. Is there a workaround which will enable use of device-mapper multipathing.
Thanks in advance.

Local disks on the servers.

Similar Messages

Cluster Not Available error when booting OVM server DOM0 from USB

Hi all,
I have been wanting to run OVM 2.2 from a USB stick for a while now. I ran through the install CD today and saw there was a 'minimal USB install' option on the installation screen. So I plugged in a 4GB USB stick, rebooted, and ran the USB install.
After a while, the install completed, OVM booted up perfectly it seemed, and I was happy.
The server had no local storage apart from this USB stick as we are going to be hooking it up to an iSCSI target for guest VM storage.
This was a new lab setup, so I got the VM Manager running on another system, connected to the USB booted OVM server, and went to create a pool.
Running the 'test connection' that worked fine.
Trying to add the pool - failed. I got the error 'Cluster Not Available'
On the OVM server I did a repos.py -l and nothing gets listed.
I understand that for HA you need a repository with shared storage. However, before I have used these same install CDs with a singe local hard disk, to get OVM server running, and it did so fine.
So what's different with the USB disk install that means you don't get a repository?
Do I have to create a special repository LUN, manually connect the OVM server to this, and then use repos.py to create it?
Cheers,
Jeff

Ok I have read, re-read related parts of the underground manual and have hit a brick wall with the 'simple' task of setting up a NIC for dedicated iSCSI traffic. The manual does not cover this task in the iSCSI SAN section.
So I'm in a catch 22.
My OVM server boots up, with it's single management NIC configured, but once booted - I don't want to connect to the iSCSI target via this network - I need it completely closed off.
I want to use another NIC (or bond of two) for dedicated iSCSI traffic to and from the OVM Server.
However, the underground manual strangely only seems to cover a single NIC configuration! It looks like he assumes you are going to be running iSCSI on your management NIC, which I can only think is incredibly bad practice.
I have installed OVM with the linux asknetwork - this did not give me the option to configure additional interfaces.
I then manually edited /etc/sysconfig/network-scripts/ifcfg-eth1 on the OVMServer, restarted network, restarted the entire box - I can ping my new NIC / IP locally from within the OVM server, but I can not connect to the iSCSI target, nor can I ping the iSCSI NIC on the OVM server from the target.
This is such a simple thing to try to achieve - super frustrating...
Jeff

Can't remove OVM server from server pool

Hi!
I tried to remove Oracle VM server using OVM. Both have version 3.0.3
I use Add/remove wizard from server pool. Selected my single server from list and pressed OK. Nothing happend. Server still in pool.
I tried to do so when server was online, offline. Results are the same. I reinstalled OVS, but now i'm can't discover it again even i configured different name and IP address. MAC is the same and it's the reason.
So, is it any way to force delete OVS using any tool or i have to reinstall OVM and it's repository?

I did it but i don't know how :)
I tried to discover re-installed server with new IP and name. This operation was failed.
I changed IP address on re-installed server to old one. And re-discovered server with old name. After that it appeared in unassigned machines list. So, i was able to delete it and also delete old server pool

Install OVM server 3.2.1 on iSCSI LUN

Hello friends,
I have a Cisco Blade and would like to boot OVM server from an iSCSI lun. I create a LUN from a SAN and presented the lun to the blade. When the server boots up, it sees the iSCSI lun just fine but when I tried to install OVM server 3.2.1, it did not detect that LUN. I tried to install Oracle Linux 6.3 and it sees the LUN ok. Is there a way to make OVM server to see that LUN.
thanks,
TD

I have done to many oel and ovs install lately such that they are blurring together...
On the OVM 3.2.1 install watch closely for a place to add support for non local storage. I remember seeing a small prompt in some of the installs but it may have been from some of the OEL installs (I have also been doing OEL 4, 5, & 6) & not OVM.
Sorry I can't remember right now. If I get a moment and try to run the install process I will and report back if no one else does.

Unable to Boot OVM Manager ISO from iLOM

Hi all,
I am trying to boot the OVM Manager ISO on a Fire X4150 using the console redirection but all that comes up is "Grub" and allows no further interaction. I've tried downloading the ISO multiple times, using an older version of the OVM installation ISO, and upgraded the iLOM firmware to the latest and greatest. When I use the OVM Server ISO, it boots up normally and starts the installation process. The machine that I am running the remote console from and has the ISO image is in the same rack as the server where I would want to install OVM Manager.
Getting physical access to pop a DVD in the drive is limited so really trying to figure out a way I can do this remotely.
Any one else have this issue and knows of a workaround?
Thanks,
Peter

FYI, update the FIRWARE on your Sun Fire X4150s. I was on a project a couple weeks ago and we where unable to use the ILOM to mount ISO files. After 10 hours of FUN, we discovered that its was a BUG! Update each server to the latest FIRMWARE, i.e. download the FIRWARE, and update each node from its web portal.
Sun x86 is FUN!
Roddy

OVM Installed From Template Doesn't Show Server Pool

I was wondering if someone might be able to help me with this before I go messing things up trying to fix it on my own.
I just installed OVS 2.2.1 and OVM 2.2.0 from Template (Ver. 2.4). The install went well and the OVM server is up and running. However when I log in to the OVM web administration page I don't see the Server Pool or the OVM_EL5U4_X86_OVM_MANAGER_PVM virtual machine listed. I haven't restarted the systems yet because I thought I might ask first.
Does anyone know how I might get the server pool I specified during the template installation script to show up in the OVM manager?

Nevermind. I forgot to VNC in to the console to continue the installation.

VM template image missing from OVM Server 2.2.0

Hi,
We have configured couple of linux OVM images on OVM server 2.20. These images were running successfully from 1 month , suddenly we found 1 of the VM image missing from the /OVS/running_pool directory. Can you let me know if there is any bug in OVM 2.20 as we are running 100 of VM's and VM's image missing could be big issue in future.

Welcome to the forums !
The scripts should be available in the companion directory of the CD
http://download.oracle.com/docs/cd/B19306_01/server.102/b14198/installation.htm#I4367
HTH
Srini

Failed OVM Server Upgrade from 3.0.2 to 3.1.1

I was trying to upgrade my OVM Server environment from *3.0.2* to *3.1.1* using the CD. I got the following error message after choosing the Upgrade option: Oracle VM server 3.0.2 (cciss/c0d0p2). The other option was to Reinstall System.
+"Error mounting device sda2 as /:+
+Device or resource busy+
+This most likely means this+
+partition hast not been formatted.+
+Press OK to reboot your system."+
My hardware is a HP BL465c G7 server, which is certified by Oracle to run the OVM 3.1.
http://linux.oracle.com/pls/apex/f?p=117:2:3396212307219555::NO::P2_CERT_ID:1871
Any advice that can help me find the way to upgrade my servers?
Thanks in advance,
Andrés L.

Hi,
I tried "yum update -y" and after reboot I encountered a kernel panic after the following message: switchroot: mount failed: No such file or directory.
So I got to reinstall ovm 3.0.3 from scratch.
Maybe the hard disk drivers in 3.1.1 kernel version (uek2) are not compatible with my server disks.
If someone has an accurate explanation to this issue, could be helpfull.
Best Regards,
Andres Loaiza.

Tcp/7002 needed from OVM Server to OVM Manager?

Hi,
When i take a look at the diagram on 3.3.2.6. Firewall Configuration i see that the Oracle VM Server only need tcp/7001 and udp/123. When i check the logging of an OVM Server i see the following logs:
[2013-10-23 12:24:20 7761] DEBUG (base:109) url: https://0.0.0.0:7002/ovm/core/bindings/python_3.2.2.520.zip option: None
[2013-10-23 12:24:20 7761] DEBUG (base:112) new url: https://10.100.0.38:7002/ovm/core/bindings/python_3.2.2.520.zip
On 3.3.2.6. Firewall Configuration i read: "The ports used by the Oracle VM Servers to connect to Oracle VM Manager are: 7001, 7002 and 54321". Can someone of Oracle update the diagram? The diagram is not complete.
Regards,
Alwin.

This blog post will provide all of the Oracle VM intra-communication details your looking for: Oracle VM Manager Installation

Lost connection between ovm-manager and ovm-server (ovm 3.1)

Hi,
I have a manager as a VM on my laptop and a physical server as an ovm-server. Everything is fine when i boot the server but after some time, the manager stop to show the information from the server. in the ovs-agent log file the following messages are repeated regulary
[2013-08-29 13:06:19 6160] DEBUG (notification:289) Trying to connect to manager.
[2013-08-29 13:06:20 6160] DEBUG (notification:291) Connected to manager.
[2013-08-29 13:06:20 6160] ERROR (notification:316) No manager Core API server object for 10:60:4b:88:bc:24:10:60:4b:88:bc:25:fe:ff:ff:ff.
[2013-08-29 13:06:34 6171] ERROR (notification:64) Unable to send notification: (111, 'Connection refused')
I did not reboot neither the server nor the manager. No IP change. Rebooting the manager has no effect.
The only way to restore communication is to restart the ovs-agent on the server, but this restart disable the network card. I must be physical on the server to restart the network.
any help will be very helpful
best regards
Jean-Marc

Hi,
The are many steps to verify:
Verify with your firewall on the Oracle VM Manager system (service iptables stop).
Verify with a ping between OVM Manager and OVM Server using the IP Address and using the hostname for each one (from OVM Manager : ping ovm-server ; and from OVM Server: ping ovm-manager).
I hope this can help you
Best Regards

OVM Server 2.1.2 does not prompt for Networking

OVM Server 2.1.2 does not prompt for a network setup (ie DHCP or static IPs).
- I tried re-installing this several times
This same machine works with native Enterprise Linux 5.2 x86-64 networking.
The motherboard has two onboard Realtek 8111C chips (10/100/1000 Mbit)
http://www.gigabyte.com.tw/Products/Motherboard/Products_Overview.aspx?ProductID=2842
Will setting up networking by hand work with OVM Server 2.1.2?

OK I finaly made my Realtek Semiconductor Co., Ltd. RTL8111/8168B PCI Express Gigabit Ethernet controller to work.
First install an Ethernet card in your box that OVM supports. This to make the OVM installation run smoothly.
When OVM is up and running follow the steps in [Mini-HOWTO] RPMs needed to compile a kernel-module on / for VM Server to get a working environment to compile modules.
Now get the driver for the unsupported card from ftp://61.56.86.122/cn/nic/r8168-8.008.00.tar.bz2 .
Unpack the file by running;:
tar -jxvf r8168-8.008.00.tar.bz2
Now follow the steps in the included readme file to compile and install the driver.
After this I edited the file /etc/modules.conf and changed the line
alias eth0 xxx
to
alias eth0 r8168
I then turned of the machine and removed my temporary network card.
When booted I ran system-config-network to configure the network once more to the original values I entered when I installed.

OVM Server 3.2.1 install reboots server

I have been trying to install OVM Server on an IBM xSeries 336 server. OVM Server 2.2.1 was previously installed on this server.
Trying to install 3.2.1 now and I get to the splash screen booting from the CD. I press [Enter] to begin the installation. Once the process reaches the point of 'Loading SCSI driver', the server reboots. I have tested the media on other machines without issue. I have also tested the server with OVM Server 2.2.1 media and the install progresses normally to the end.
I suspect there are missing drivers causing the install to fail. Is there a way to load the missing drivers during the install process?
Thanks in advance,
UPDATE: I have successfully installed OVM Server 3.0.2... Not exactly what I wanted but....
Edited by: SBaugher on Feb 20, 2013 11:04 AM

Oracle VM 3 is entirely different from 2.2.1.
take a look at the release notes for 3.2.1 There are few instances in which you have to specify different boot parameters for the installation to work.
http://docs.oracle.com/cd/E35328_01/E35329/E35329.pdf
You might have a MEGARAID product.
If you are installing Oracle VM Server on a Sun Fire X4800, you must provide extra parameters when
booting from the installation media (CDROM or ISO file), or when using a kickstart installation. These
parameters allow the megaraid_sas driver to load correctly.
If booting from the installation media, press F2 when the initial boot screen is displayed and provide the
following additional parameters as part of the boot command:
mboot.c32 xen.gz extra_guest_irqs=64,2048 nr_irqs=2048 --- vmlinuz --- initrd.img
If using a kickstart installation, add the additional kernel parameters to the PXE configuration file.

Problem with OVM Server 3.2.6 - xend not running

Hi, just now I upgraded my server to OVM Server 3.2.6 version and server boots up, cluster is up but I have message that Hypervisor is running in UNDETERMINED bit mode (WARNING: XEND IS PROBABLY NOT RUNNING). While boot I saw message that xend failed to start because "mount point /proc/xen does not exist". Also when it tries to start xend, there is message "xencommons should be started first". Reinstalling OVMS did not help. Does anyone have this problem?
Thanks.

For me, it looks like the same.
1 - Yes
2 - B
3 - Not yet, I've tried to downgrade to 3.2.4, but it is not possible.
4.1 - service xend status:
xc: error: Could not obtain handle on privileged command interface (2 = No such file or directory): Internal error
Traceback (most recent call last):
File "/usr/sbin/xend", line 36, in ?
    from xen.xend.server import SrvDaemon
File "/usr/lib64/python2.4/site-packages/xen/xend/server/SrvDaemon.py", line 26, in ?
    import relocate
File "/usr/lib64/python2.4/site-packages/xen/xend/server/relocate.py", line 28, in ?
    from xen.xend import XendDomain
File "/usr/lib64/python2.4/site-packages/xen/xend/XendDomain.py", line 36, in ?
    from xen.xend import XendOptions, XendCheckpoint, XendDomainInfo
File "/usr/lib64/python2.4/site-packages/xen/xend/XendCheckpoint.py", line 20, in ?
    from xen.xend import balloon, sxp, image
File "/usr/lib64/python2.4/site-packages/xen/xend/image.py", line 46, in ?
    xc = xen.lowlevel.xc.xc()
xen.lowlevel.xc.Error: (1, 'Internal error', 'xc_interface_open failed: No such file or directory')
xend daemon is stopped
4.2 - service ovs-agent status:
log server (pid 4460) is running...
notification server (pid 4489) is running...
remaster server (pid 4492) is running...
monitor server (pid 4494) is running...
ha server (pid 4496) is running...
stats server (pid 4498) is running...
xmlrpc server (pid 4499) is running...

OVM Server 3.2.8 - LUN Larger Than Allowed By The Host Adapter

Heya,
We have 2 blades that where recently updated to OVM Server 3.2.8. We had 2x100GB LUN's assigned to those servers ... however, when trying to to discover them we get the following :
Oct 27 14:34:43 XXXXXXXXXX kernel: scsi: host 0 channel 3 id 0 lun4194304 has a LUN larger than allowed by the host adapter
Oct 27 14:34:43 XXXXXXXXXX kernel: scsi: host 0 channel 3 id 0 lun4194560 has a LUN larger than allowed by the host adapter
These blades where previously on 3.1.1 and where updated along with the OVM Manager. Everything is working fine on the blades and VM's there. We have had storage assigned previously from the SAN but when on 3.1.1.
Any idea what needs to be changed ?
Regards

What make of HBA is this? This LUN numbers are actually unusually high. And what storage HW provides these LUNs?
Cheers,
budy

Unable to delete server from pool

We are in the proces of upgrading oure OVM farm from 3.2.8 >> 3.3.1 (10 servers)
The management server is allready upgraded to 3.3.1 + MySQL and we started on the servers last week.
After reinstalling the first server we tried to add it to the pool. This job was completed succesfully.
But the server suddenly showed as no owned / not in ownership status. We tried the "edit server" setting and marked the "Take Ownership of Server" box and used the OVM Agent password.
This job runs and the server is showing as owned and the job results are succefull!!
This status stay's for about 5 seconds and then the server is "locked" again and the status changed to "not owned" status (blue server icon) again however there is no mention of this job in the logs??
We tried to remove the server from the pool again. This won't work because the server is not owned by the pool / cluster anymore.
But for some reason now the server is still in the pool / cluster and is showing as owned by the cluster (green icon) but it is locked. Any job for this server will fail with the following error,
GEN_000002:No object of type Server with id null(34:35:39:35:32:39:5a:43:4a:32:32:31:30:4c:34:53), is currently in the db.
We can't do anything with this server anymore, we can't delete it either.
Because if this issue we halted the 3.2.8 >> 3.3.1 server migrations.
Really stuck now on how to fix this one.
Logging on the ovmm server and the node server don't show at all why it can't delete this server other then error "No object of type Server with id null(34:35:39:35:32:39:5a:43:4a:32:32:31:30:4c:34:53), is currently in the db"

Quick update,
Last weekend we tried to upgrade our non production OVM cluster (8 nodes)
The OVM manager upgrade (3.2.9 to 3.3.1) went fine the OVM servers on the other hand.....
We used the ./UpgradeServers.py script and repository's on the OVM management server.
The upgrade of the OVM server itself went fine. After the reboot the server comes back online and then switches to UNOWNED status with an cluster status of UNKNOWN.
There is no way set ownership of the server as happened before. Also doing anything at all with the server is not possible because the server node is UNOWNED.
When re-installing the server node from scratch with OVM 3.3.1 it works BUT the server must not be removed from the cluster.
Best solution I think is,
1: Shutodown ALL vm's in the cluster.
2: Unpresent ALL storage to cluster nodes
3: Remove ALL cluster nodes (ovm servers) from cluster.
4: Clean install OVM 3.3.x on ALL servers
5: Add servers to cluster
6: Present ALL storage to cluster nodes again
7: Start VM's
We are going to test this solution in the next 2 weeks.
I'll post any updates

Booting OVM server from multipath SAN

Similar Messages

Maybe you are looking for