Proxmox VE on newer SuperMicros with Intel "Patsburg" chipsets?

[Jim H]

All,

We are speccing out a brand-new server to run Proxmox. The current config would have a Supermicro X9DRW-3F motherboard with a "Patsburg" c606 chipset

This newer generation of chipsets is not listed on the Wiki under hardware/motherboards, but that doesn't mean it work. Before we spend money, though, can anyone attest to the magnificence of Proxmox running on a SuperMicro board with a model number starting with "X9D" (Patsburg-based boards)?

I promise to report success and/or failure to the boards ans Wiki

Jim

 

[briceburg]

Jim,

We ran intro trouble with the SAS controller on the motherboard (provided by the Intel "Patsburg" C606 chipset). The proxmox installer got to 99%, and then would fail to write to the MBR of the disk we were installing to. The error in console was SAS: errno -16 failed to reset device... Upon rebooting a blank screen comes up [ grub failed to write itself ].

We are installing to /dev/sda which was a SanDisk SSD attached to the first port of the Supermicro Backplane ( our system model is: SuperServer 1027R-WRFT+ ). I thought it was perhaps due to drive firmware, so tried installing to a "traditional" magnetic Toshiba SAS drive ( AL13SEB900 ) and got the same error. It hange on writing to the MBR. Checked for the latest Supermicro bios, and we have it ( 3.0 at the time of this writing ).

... so I suspect It's a Proxmox 3.1 driver issue with the C606 SAS controller.

I've gone ahead and purchased a real RAID controller and hope it will alleviate our problem. Will report back when they arrive.

~ Brice

 

[mir]

Before you buy a new controller. Have you tested with latest kernel proxmox-ve-2.6.32: 3.1-111 (running kernel: 2.6.32-24-pve) which was released yesterday?

 

[tom]

Before you buy a new controller. Have you tested with latest kernel proxmox-ve-2.6.32: 3.1-111 (running kernel: 2.6.32-24-pve) which was released yesterday?

Just to note, we plan to upload an updated 3.1 ISO installer with this kernel next week.

 

[briceburg]

I had used the old 3.1 ISO installer -- pve-kernel-2.6.32-23-pve (2.6.32-109), which is based on vzkernel-2.6.32-042stab079.5. I do not really have a means to test the new 3.1-111 until the ISO installer is updated! I don't have a base OS on these servers yet....

Of interest: There appear to be very little changes between vzkernel-2.6.32-042stab079.5 and vzkernel-2.6.32-042stab079.6 which the new proxmox 3.1-111 is based on. See:
http://openvz.org/Download/kernel/rhel6/042stab079.5

There do however, appear to be very many changes between vzkernel-2.6.32-042stab079.6 and the LATEST stable from openVZ, vzkernel-2.6.32-042stab081.3 See:
http://openvz.org/Download/kernel/rhel6/042stab081.3

In particular: "Rebase to new RHEL6.4 kernel 2.6.32-358.18.1.el6"

Proxmox 3.1 [old and latest -311] are based on: 2.6.32-358.14.1.el6. I found a interesting, and potentially related bugfix in the release notes for .18.1:

( https://access.redhat.com/site/docu..._Linux/6/html/6.4_Technical_Notes/kernel.html )
BZ#990470
A race condition in the abort task and SPP device task management path of the isci driver could, under certain circumstances, cause the driver to fail cleaning up timed-out I/O requests that were pending on an SAS disk device. As a consequence, the kernel removed such a device from the system. A patch applied to the isci driver fixes this problem by sending the task management function request to the SAS drive anytime the abort function is entered and the task has not completed. The driver now cleans up timed-out I/O requests as expected in this situation.


When an ISO installer is available, I'll try. Is there a way I can make an ISO installer based on -311 ?

Does the proxmox team intend to use vzkernel-2.6.32-042stab081.3 anytime soon? I personally think this is a more important and meaningful upgrade ... and hope it fixes my problem.

~ Brice

 

[briceburg]

^^ I updated my above post to include that I did indeed find a SAS related bug report in the OpenVZ 042stab081.3 kernel.