Guru Meditation 1103 (VINF_EM_DBG_HYPER_ASSERTION)
Posted: 23. Apr 2012, 16:42
Hi. My first post here. I have read the forum guide, and searched.
I have repeated VM failures, with the above error. This is VB 4.1.12. Host is Oracle Enterprise Linux x86_64 5.5, guests are OEL x86 5.7. I can't use 64bit guests because the CPUs are the older AMD Opterons, that don't have hardware virtualization. I have set --hwvirtex off in all the VMs. There is plenty of CPU and RAM. The failures occur unpredictably, but usually within ten minutes of starting the VMs: I lose one, then another a couple of minutes later.
I am trying to run three VMs, with the objective of setting up an Oracle RAC cluster, using Oracle Grid Infrastructure release 11.2.0.3.
The first VM is small (500M RAM) and doesn't do much: it is just the DHCP and DNS server. This never crashes.
Then two more VMs that I have configured as the RAC nodes, with 2G RAM each. I've now reduced the RAM to 1.5G each in case VirtualBox had problems allocating that much. I have the Grid Infrastructure 11.2.0.3 installed and working on the two RAC nodes, the failures usually occur shortly after booting them and starting the Grid services. So these two VMs are working, but not very hard, at the time of failure. It is possible I suppose that they are still building up to full RAM usage.
All VMs have 4 NICs, using host-only networking.
Each VM has it's own VDI disc to boot off, and then there are some shareable VDIs too.
If I start only one of the 1.5G VMs, I still get the same problem: it runs for a few minutes: sometimes long enough to get all the Grid services going before it fails, sometimes not.
I'll upload the logs and the .vbox files
Any insight into this issue will be very gratefully received.
John.
I have repeated VM failures, with the above error. This is VB 4.1.12. Host is Oracle Enterprise Linux x86_64 5.5, guests are OEL x86 5.7. I can't use 64bit guests because the CPUs are the older AMD Opterons, that don't have hardware virtualization. I have set --hwvirtex off in all the VMs. There is plenty of CPU and RAM. The failures occur unpredictably, but usually within ten minutes of starting the VMs: I lose one, then another a couple of minutes later.
I am trying to run three VMs, with the objective of setting up an Oracle RAC cluster, using Oracle Grid Infrastructure release 11.2.0.3.
The first VM is small (500M RAM) and doesn't do much: it is just the DHCP and DNS server. This never crashes.
Then two more VMs that I have configured as the RAC nodes, with 2G RAM each. I've now reduced the RAM to 1.5G each in case VirtualBox had problems allocating that much. I have the Grid Infrastructure 11.2.0.3 installed and working on the two RAC nodes, the failures usually occur shortly after booting them and starting the Grid services. So these two VMs are working, but not very hard, at the time of failure. It is possible I suppose that they are still building up to full RAM usage.
All VMs have 4 NICs, using host-only networking.
Each VM has it's own VDI disc to boot off, and then there are some shareable VDIs too.
If I start only one of the 1.5G VMs, I still get the same problem: it runs for a few minutes: sometimes long enough to get all the Grid services going before it fails, sometimes not.
I'll upload the logs and the .vbox files
Any insight into this issue will be very gratefully received.
John.