Bad NIC on Supermicro H8DME-2 (cheap 24bay server)?

Hi everyone,

I've been using unraid for the past couple of years, and have outgrown my 12 bay tower and just got one of the $300 24bay servers that have been discussed in this forum and over on avsforum. However there appears to be a problem with the nic or the driver when booting as it's not getting an ip address and has no network connectivity.

This is the motherboard: http://www.supermicro.com/aplus/motherboard/opteron2000/mcp55/h8dme-2.cfm

Here's a section of the syslog on the new Supermicro server (complete attached):

Jun  8 21:43:35 Tower ifplugd(eth0)[1421]: Using interface eth0/00:30:48:C7:1D:50 with driver <forcedeth> (version: 0.64)

Jun  8 21:43:35 Tower ifplugd(eth0)[1421]: Using detection mode: SIOCETHTOOL

Jun  8 21:43:35 Tower ifplugd(eth0)[1421]: Initialization complete, link beat detected.

Jun  8 21:43:35 Tower kernel: forcedeth 0000:00:08.0: irq 28 for MSI/MSI-X

Jun  8 21:43:35 Tower ifplugd(eth0)[1421]: Executing '/etc/ifplugd/ifplugd.action eth0 up'.

Jun  8 21:43:35 Tower ifplugd(eth0)[1421]: client: Polling for DHCP server on interface eth0:

Jun  8 21:43:36 Tower logger: /etc/rc.d/rc.inet1:  /sbin/dhcpcd -d -t 30 -h Tower eth0

Jun  8 21:43:36 Tower kernel: forcedeth 0000:00:08.0: irq 28 for MSI/MSI-X

Jun  8 21:43:36 Tower dhcpcd[1449]: broadcasting DHCP_DISCOVER

Jun  8 21:44:06 Tower dhcpcd[1449]: timed out waiting for a valid DHCP server response

Jun  8 21:44:06 Tower ifplugd(eth0)[1421]: client: dhcpcd: MAC address = 00:30:48:c7:1d:50

Jun  8 21:44:06 Tower ifplugd(eth0)[1421]: Program executed successfully.

Jun  8 21:44:06 Tower kernel: forcedeth 0000:00:08.0: irq 28 for MSI/MSI-X

This is comparison to the syslog when I boot on the tower that I'm trying to replace (complete attached):

Jun  9 03:04:56 Tower ifplugd(eth0)[1267]: Using interface eth0/10:78:D2:D9:9A:07 with driver <r8169> (version: 2.3LK-NAPI)

Jun  9 03:04:56 Tower ifplugd(eth0)[1267]: Using detection mode: SIOCETHTOOL

Jun  9 03:04:56 Tower ifplugd(eth0)[1267]: Initialization complete, link beat detected.

Jun  9 03:04:56 Tower kernel: r8169: eth0: link up

Jun  9 03:04:56 Tower kernel: r8169: eth0: link up

Jun  9 03:04:56 Tower ifplugd(eth0)[1267]: Executing '/etc/ifplugd/ifplugd.action eth0 up'.

Jun  9 03:04:56 Tower ifplugd(eth0)[1267]: client: Polling for DHCP server on interface eth0:

Jun  9 03:04:57 Tower logger: /etc/rc.d/rc.inet1:  /sbin/dhcpcd -d -t 30 -h Tower eth0

Jun  9 03:04:57 Tower dhcpcd[1295]: broadcasting DHCP_DISCOVER

Jun  9 03:04:57 Tower kernel: r8169: eth0: link up

Jun  9 03:04:57 Tower dhcpcd[1295]: broadcastAddr option is missing in DHCP server response. Assuming 10.0.0.255

Jun  9 03:04:57 Tower dhcpcd[1295]: dhcpIPaddrLeaseTime=604800 in DHCP server response.

Jun  9 03:04:57 Tower dhcpcd[1295]: dhcpT1value is missing in DHCP server response. Assuming 302400 sec

Jun  9 03:04:57 Tower dhcpcd[1295]: dhcpT2value is missing in DHCP server response. Assuming 529200 sec

Jun  9 03:04:57 Tower dhcpcd[1295]: DHCP_OFFER received from  (10.0.0.1)

Jun  9 03:04:57 Tower dhcpcd[1295]: broadcasting DHCP_REQUEST for 10.0.0.4

Jun  9 03:04:57 Tower dhcpcd[1295]: dhcpIPaddrLeaseTime=604800 in DHCP server response.

Jun  9 03:04:57 Tower dhcpcd[1295]: dhcpT1value is missing in DHCP server response. Assuming 302400 sec

Jun  9 03:04:57 Tower dhcpcd[1295]: dhcpT2value is missing in DHCP server response. Assuming 529200 sec

Jun  9 03:04:57 Tower dhcpcd[1295]: DHCP_ACK received from  (10.0.0.1)

Jun  9 03:04:57 Tower ifplugd(eth0)[1267]: client: dhcpcd: MAC address = 10:78:d2:d9:9a:07

Jun  9 03:04:57 Tower ifplugd(eth0)[1267]: client: dhcpcd: your IP address = 10.0.0.4

Jun  9 03:04:57 Tower ifplugd(eth0)[1267]: client: dhcpcd: MAC address = 10:78:d2:d9:9a:07

Jun  9 03:04:57 Tower ifplugd(eth0)[1267]: client: dhcpcd: your IP address = 10.0.0.4

Jun  9 03:04:57 Tower ifplugd(eth0)[1267]: Program executed successfully

I have used several network cables, used two different routers, tried using static ip addresses (which worked on the existing tower, not on the new 24bay server, no connection to rest of network and not detected by routers), reset bios settings, and tried stock boots of 4.7, 5.12a, and 5.13. I consider myself to be pretty good with computers and networking (had cisco classes growing up at my high school), but I'm out of ideas here. I might go pickup a pci-e intel nic to bypass the internal nics that are supposed to work just fine on this server (lots of other people using it), but wanted to check here first if anyone had any ideas.

syslog_4.7_24bay.txt

syslog_4.7_tower.txt

First, disable one of the NICs in the BIOS to see if there's some issue with the dual NICs having some conflict with UnRAID.

If that doesn't resolve it, I'd so exactly as you planned => pick up an Intel NIC card and disable the internal NICs.  That will almost certainly fix it.

I had a similar issue with my SM server. I could access the server via my LAN, but it would not access the web. This is after LAN port 1 didn't function at all. So, on port 2, I could access the gui, but no WAN access. All I did was swap the disks, PSU, and USB drive from the previous working machine. It was set to DHCP=OFF, and had a static IP (.185) outside of my router's range. The router also had the specific IP (.185) set as static in my Tomato driven router.

I had to set DHCP=ON and remove the saved static route in the router. All WAN activity began at that point.

:shrugs: