we face a system gets stuck problem on heavy workload on local hard disk( virtual machine ) .Stuck how? One process? Can you log in while this is happening? Does it correct itself? How do you recover? What is the workload? I/O? Compute-bound?
We run Oracle VM 2.2.2 with virtual machines running on Oracle Linux 5.8Are these PVM, HVM, or PVHVM guests? Are the disks directly handled by the domU guest? How are the disks presented to the servers? HBA? iSCSI? USB?
If we create a heavy workload on the virtual machines by creating a archive with gz the hole system gets stuck for that time.On one guest? Multiple guests? Any guest? How do the other machines get "stuck".
I ´ve no idea where to find additional information.First, check the Oracle VM User's Guide. There is actually some information in there.
Does anyone can help to work this out?Not with the information we have so far.
Dude wrote:No, the xen variant in the 2.6.18 family is the PVM kernel for DomU guests. It's not the hypervisor (that's Xen itself).
The xen kernel is a type 1 hypervisor. You would use such kernel if you use e.g. Oracle VM Server for x86. It is not meant to be used as a VM guest kernel, e.g. to run Oracle database. The PVM drivers are supposed to increase throughput of I/O operations in guest operating systems on top of the Oracle VM Server hosts.
tbrinkmann wrote:Your Grid Infrastructure install is too old: you need 220.127.116.11 or higher for ACFS to work on 2.6.32-100, -200 and -300. Check My Oracle Support Doc ID 1369107.1 for more information.
Does anyone can help understand what I do wrong ?