I was having some issues upgrading from beta9 to beta10 then beta10a.  So copied off my flash drive contents and formatted the drive to a clean beta10a.  I don't know if that is useful or not.  But I do notice the following in my syslog:

Oct  4 09:13:18 Tower2 kernel: IOMMU 0 0xfed90000: using Queued invalidation

Oct  4 09:13:18 Tower2 kernel: IOMMU 1 0xfed91000: using Queued invalidation

Oct  4 09:13:18 Tower2 kernel: IOMMU: Setting RMRR:

Oct  4 09:13:18 Tower2 kernel: IOMMU: Setting identity map for device 0000:00:02.0 [0xbf800000 - 0xcf9fffff]

Oct  4 09:13:18 Tower2 kernel: IOMMU: Setting identity map for device 0000:00:14.0 [0xbd5a5000 - 0xbd5d6fff]

Oct  4 09:13:18 Tower2 kernel: IOMMU: Setting identity map for device 0000:00:1a.0 [0xbd5a5000 - 0xbd5d6fff]

Oct  4 09:13:18 Tower2 kernel: IOMMU: Setting identity map for device 0000:00:1d.0 [0xbd5a5000 - 0xbd5d6fff]

Oct  4 09:13:18 Tower2 kernel: IOMMU: Prepare 0-16MiB unity mapping for LPC

Oct  4 09:13:18 Tower2 kernel: IOMMU: Setting identity map for device 0000:00:1f.0 [0x0 - 0xffffff]

Oct  4 09:13:18 Tower2 kernel: PCI-DMA: Intel® Virtualization Technology for Directed I/O

Oct  4 09:13:18 Tower2 kernel: ------------[ cut here ]------------

Oct  4 09:13:18 Tower2 kernel: WARNING: CPU: 5 PID: 1 at drivers/pci/search.c:131 pci_find_upstream_pcie_bridge+0x51/0x63() (Minor Issues)

Oct  4 09:13:18 Tower2 kernel: Modules linked in:

Oct  4 09:13:18 Tower2 kernel: CPU: 5 PID: 1 Comm: swapper/0 Not tainted 3.16.3-unRAID #3 (Errors)

Oct  4 09:13:18 Tower2 kernel: Hardware name: To Be Filled By O.E.M. To Be Filled By O.E.M./Z77 Pro4, BIOS P1.80 07/11/2013

Oct  4 09:13:18 Tower2 kernel: 0000000000000000 ffff880419073d30 ffffffff815d9532 0000000000000000

Oct  4 09:13:18 Tower2 kernel: ffff880419073d68 ffffffff81040687 ffffffff8137f232 0000000000000000

Oct  4 09:13:18 Tower2 kernel: ffff880418f6b000 ffff880418f6b098 0000000000000000 ffff880419073d78

Oct  4 09:13:18 Tower2 kernel: Call Trace: (Errors)

Oct  4 09:13:18 Tower2 kernel: [<ffffffff815d9532>] dump_stack+0x45/0x56

Oct  4 09:13:18 Tower2 kernel: [<ffffffff81040687>] warn_slowpath_common+0x75/0x8e

Oct  4 09:13:18 Tower2 kernel: [<ffffffff8137f232>] ? pci_find_upstream_pcie_bridge+0x51/0x63

Oct  4 09:13:18 Tower2 kernel: [<ffffffff8104073f>] warn_slowpath_null+0x15/0x17

Oct  4 09:13:18 Tower2 kernel: [<ffffffff8137f232>] pci_find_upstream_pcie_bridge+0x51/0x63

Oct  4 09:13:18 Tower2 kernel: [<ffffffff814e65d3>] intel_iommu_add_device+0x37/0x1a9

Oct  4 09:13:18 Tower2 kernel: [<ffffffff814df2e7>] ? bus_set_iommu+0x4a/0x4a

Oct  4 09:13:18 Tower2 kernel: [<ffffffff814df31c>] add_iommu_group+0x35/0x43

Oct  4 09:13:18 Tower2 kernel: [<ffffffff8140ea3f>] bus_for_each_dev+0x78/0x82

Oct  4 09:13:18 Tower2 kernel: [<ffffffff814df2e0>] bus_set_iommu+0x43/0x4a

Oct  4 09:13:18 Tower2 kernel: [<ffffffff818c5706>] intel_iommu_init+0x51b/0x565

Oct  4 09:13:18 Tower2 kernel: [<ffffffff81894cff>] pci_iommu_init+0x12/0x3c

Oct  4 09:13:18 Tower2 kernel: [<ffffffff81894ced>] ? memblock_find_dma_reserve+0x146/0x146

Oct  4 09:13:18 Tower2 kernel: [<ffffffff81002146>] do_one_initcall+0xea/0x17c

Oct  4 09:13:18 Tower2 kernel: [<ffffffff81057208>] ? parse_args+0x19e/0x278

Oct  4 09:13:18 Tower2 kernel: [<ffffffff8188cffe>] kernel_init_freeable+0x17a/0x200

Oct  4 09:13:18 Tower2 kernel: [<ffffffff8188c805>] ? initcall_blacklist+0x9f/0x9f

Oct  4 09:13:18 Tower2 kernel: [<ffffffff815cce50>] ? rest_init+0x74/0x74

Oct  4 09:13:18 Tower2 kernel: [<ffffffff815cce59>] kernel_init+0x9/0xd5

Oct  4 09:13:18 Tower2 kernel: [<ffffffff815df3bc>] ret_from_fork+0x7c/0xb0

Oct  4 09:13:18 Tower2 kernel: [<ffffffff815cce50>] ? rest_init+0x74/0x74

Oct  4 09:13:18 Tower2 kernel: ---[ end trace 100dfd329ac6f065 ]---

Oct  4 09:13:18 Tower2 kernel: RAPL PMU detected, hw unit 2^-16 Joules, API unit is 2^-32 Joules, 3 fixed counters 163840 ms ovfl timer

Oct  4 09:13:18 Tower2 kernel: futex hash table entries: 2048 (order: 5, 131072 bytes)

Oct  4 09:13:18 Tower2 kernel: Installing knfsd (copyright © 1996 [email protected]).

 

The boot continues, and I can interact with unRAID (it seems normally).

Try checking for an update for your BIOS.  If none, you may need to wait for either a kernel update or a BIOS update, or turn off the virtualization support in your BIOS.  However sometimes this kind of issue can be harmless, if it only affects some subsystem which you aren't actually using.  Does the system appear to run fine?

I checked for a BIOS update, there aren't any.  I'm seeing any issue where I get a BIOS error and the system won't boot if I try to restart from the unRAID menu (I get "MACHINE CHECK ERROR" in big letters).  But power off, then power on seems to work without issue.

 

I am using KVM and seem to have GPU passthrough working, although I need to fix some other things to really test that out.

From what I read, "machine check errors" are notoriously difficult to fix.  I would make sure there is no overclocking set, then make sure that the CPU and RAM and I/O chips are at their most stable settings, perhaps even slower.  If that doesn't help, this machine may not be reliable for you, for what you are trying to do.

I checked for a BIOS update, there aren't any.  I'm seeing any issue where I get a BIOS error and the system won't boot if I try to restart from the unRAID menu (I get "MACHINE CHECK ERROR" in big letters).  But power off, then power on seems to work without issue.

 

I am using KVM and seem to have GPU passthrough working, although I need to fix some other things to really test that out.

 

I saw this issue at one point in an earlier beta and discovered it was due to overheating on my CPU.  I resolved the overheating with a new fan / better job of mounting it / better thermal paste.  It seems odd because you probably haven't changed anything like that from the old to the new version of unRAID, but if you weren't using GPU pass through before, the heat you were generating inside the chassis would have been less.  Curious to hear your thoughts on this possibility. As Rob indicated, Machine Check Errors are notoriously difficult to resolve as well as diagnose their root cause.

Hmm, I will take a look at that.  This is a new CPU, so it is the factory thermal paste they put on the CPU fan.  Haven't messed with it since the initial install.

 

The funny thing is the difference between restart and power off.  When I power off, I pretty much immediately power back on.  And that works.  For reboot, it seems to go into a loop of up/down, up/down, with the error message showing up after the bios screen for a second and then retrying.  Seems like that would eventually cool and work, but it doesn't seem to.

 

I'm not sure if there is a checksum on the BIOS or some validity check.  I was thinking maybe something messed up the bios and re-installing the current (newest) bios might help.

Hmm, I will take a look at that.  This is a new CPU, so it is the factory thermal paste they put on the CPU fan.  Haven't messed with it since the initial install.

 

The funny thing is the difference between restart and power off.  When I power off, I pretty much immediately power back on.  And that works.  For reboot, it seems to go into a loop of up/down, up/down, with the error message showing up after the bios screen for a second and then retrying.  Seems like that would eventually cool and work, but it doesn't seem to.

 

I'm not sure if there is a checksum on the BIOS or some validity check.  I was thinking maybe something messed up the bios and re-installing the current (newest) bios might help.

I will tell you that is EXACTLY what my scenario was when I bought a new CPU (Intel 4990K).  The thermal paste I received with the CPU was bad. It seemed way to runny when we applied it.  I highly recommend folks use arctic silver thermal paste as it is much better.

 

After reapplying good paste and reseating, my machine check errors went away. I was also having symptoms of weird rebooting after a power down.