Broadcom 57810S 10 Gigabit stops working after some time

[ramiro]

Hello

My Broadcom 57810S 10 Gigabit card stops to work after Unraids runs for some time , if i restart the Server everythings works again for some time , it was running fine for some months , but now i updated the Bios on PC that is directly connected to the Unraid Server and then this problem started , i think it usualy stops working when the PC that is directly connected goes into standby.

Any idea where the problem is comming from ?

 

Aug  6 21:37:13 ramiroserver kernel: bnx2x: [bnx2x_acquire_hw_lock:2016(eth3)]lock_status 0xffffffff  resource_bit 0x1
Aug  6 21:37:13 ramiroserver kernel: bnx2x 0000:09:00.1 eth3: MDC/MDIO access timeout
### [PREVIOUS LINE REPEATED 1 TIMES] ###
Aug  6 21:37:13 ramiroserver kernel: bnx2x: [bnx2x_state_wait:310(eth3)]timeout waiting for state 2
Aug  6 21:37:13 ramiroserver kernel: bnx2x: [bnx2x_func_stop:9107(eth3)]FUNC_STOP ramrod failed. Running a dry transaction
Aug  6 21:37:13 ramiroserver kernel: bnx2x: [bnx2x_igu_int_disable:895(eth3)]BUG! Proper val not read from IGU!
Aug  6 21:37:13 ramiroserver kernel: bnx2x: [bnx2x_func_hw_reset:6126(eth3)]Unknown reset_phase (0x0) from MCP
Aug  6 21:37:19 ramiroserver kernel: bnx2x: [bnx2x_fw_command:3047(eth3)]FW failed to respond!
Aug  6 21:37:19 ramiroserver kernel: bnx2x 0000:09:00.1 eth3: bc 7.13.54
Aug  6 21:37:19 ramiroserver kernel: bnx2x: [bnx2x_fw_dump_lvl:782(eth3)]3MCP PC at 0xffffffff
Aug  6 21:37:19 ramiroserver kernel: bnx2x: [bnx2x_fw_dump_lvl:794(eth3)]Unable to dump trace buffer (mark ffffffff)
Aug  6 21:37:19 ramiroserver kernel: bnx2x: [bnx2x_acquire_hw_lock:2016(eth3)]lock_status 0xffffffff  resource_bit 0x800

 

syslog.2.txt

[Frank1940]

I am not an expert in this area but I would be looking at the BIOS settings on the PC that you are allowing to go to sleep coupled with reading the MB manual.   You want to set it up so that the 10Gb card in that PC stays active.  That are often settings in the BIOS that control how 'deep' the sleep is that the system enters.  Most of the time, you can set the MB so that the NIC's to stay active so you can wake the system up from the LAN. 

 

As another suggestion, if you find that this is the issue and you can't resolve it, you might want to edit your first post to change the subject of this thread to more clearly indicate the exact nature of your problem.  Remember that Google can also be your friend when confronted with these types of issues.  I would think that this problem would also exist in the world outside of Unraid. 

 

EDIT:  I would imagine that you are not using a switch with this setup.  While this works fine for the most part, remember that by eliminating the switch, you have removed a layer of hardware which is intended to prevent these types of issues! 

[ramiro]

Hello

Already googled this issue just couldn't find anythig.

The network stays active on standby just at 100 mbit , that maybe confues the card.

But i found in the logs this error message

Aug  6 21:14:24 ramiroserver kernel: bnx2x 0000:09:00.1 eth3: NIC Link is Up, 10000 Mbps full duplex, Flow control: ON - receive & transmit
Aug  6 21:14:35 ramiroserver kernel: bnx2x: [bnx2x_timer:5796(eth3)]MFW seems hanged: drv_pulse (0x269b) != mcp_pulse (0x7fff)
Aug  6 21:14:35 ramiroserver kernel: bnx2x: [bnx2x_acquire_hw_lock:2016(eth3)]lock_status 0xffffffff  resource_bit 0x1
Aug  6 21:14:35 ramiroserver kernel: bnx2x: [bnx2x_hw_stats_update:869(eth3)]NIG timer max (4294967295)
Aug  6 21:14:35 ramiroserver kernel: bnx2x 0000:09:00.1 eth3: MDC/MDIO access timeout
### [PREVIOUS LINE REPEATED 1 TIMES] ###
Aug  6 21:14:35 ramiroserver kernel: bnx2x 0000:09:00.1 eth3: Error:  Power fault on Port 0 has been detected and the power to that SFP+ module has been removed to prevent failure of the card. Please remove the SFP+ module and restart the system to clear this error.

i just don't i have any idea what triggers it , it seems to happen when the other pc wakes up.

 

[Frank1940]

48 minutes ago, ramiro said:

Aug 6 21:14:35 ramiroserver kernel: bnx2x 0000:09:00.1 eth3: Error: Power fault on Port 0 has been detected and the power to that SFP+ module has been removed to prevent failure of the card. Please remove the SFP+ module and restart the system to clear this error.

Did you google this error?