[CentOS] Centos 6.3 Network bnx2 Problem on HP DL360

Mon Mar 25 10:10:25 UTC 2013
Svavar Örn Eysteinsson <svavar at fiton.is>

After you have tried the pcie_aspm boot option, also try :

echo performance > /sys/module/pcie_aspm/parameters/policy
This will disable ASPM on PCIe and operate with maximum performance.

This is what I use today on the DL380 G7.




On 25.3.2013, at 09:06, Woehrle Hartmut SBB CFF FFS (Extern) wrote:

> Hello Svavar
> 
> This was the first time that this problem occurred - with 60 Servers and about half a year of Centos 6 (5 before).
> But because the interfaces have a permanent load - really 24x7 - problems with power management would be a disaster.
> I will try to switch off.
> 
> Thanks
> Hartmut
> 
>> How often are you getting these crashes ?
>> 
>> I had simular problem on my HP DL380 G7 server.
>> 
>> I disabled Active State PowerManagement on the PCI-E express.
>> 
>> Try it.
>> 
>> Add pcie_aspm=off as optional boot option.
>> 
>> 
>> Best regards,
>> 
>> Svavar O
>> Reykjavik - Iceland
> 
> 
> 
>> On 19.3.2013, at 15:32, Woehrle Hartmut SBB CFF FFS (Extern) wrote:
>> 
>> Hello Mailing List
>> 
>> I got a severe network error message at a HP DL360 Server.
>> The kernel log says:
>> 
>> ----------------------------------- /var/log/messages 
>> -----------------------------------------------------------------
>> Mar 19 15:45:06 server kernel: do_IRQ: 2.168 No irq handler for vector 
>> (irq -1) Mar 19 15:45:17 server kernel: bnx2 0000:02:00.1: eth1: 
>> DEBUG: intr_sem[0] PCI_CMD[00100446] Mar 19 15:45:17 server kernel: 
>> bnx2 0000:02:00.1: eth1: DEBUG: PCI_PM[19002108] 
>> PCI_MISC_CFG[92000088] Mar 19 15:45:17 server kernel: bnx2 
>> 0000:02:00.1: eth1: DEBUG: EMAC_TX_STATUS[00000008] 
>> EMAC_RX_STATUS[00000006] Mar 19 15:45:17 server kernel: bnx2 
>> 0000:02:00.1: eth1: DEBUG: RPM_MGMT_PKT_CTRL[40000088] Mar 19 15:45:17 
>> server kernel: bnx2 0000:02:00.1: eth1: DEBUG: 
>> HC_STATS_INTERRUPT_STATUS[017f0080]
>> Mar 19 15:45:17 server kernel: bnx2 0000:02:00.1: eth1: DEBUG: 
>> PBA[00000000] Mar 19 15:45:17 server kernel: bnx2 0000:02:00.1: eth1: 
>> <--- start MCP states dump ---> Mar 19 15:45:17 server kernel: bnx2 
>> 0000:02:00.1: eth1: DEBUG: MCP_STATE_P0[0003610e] 
>> MCP_STATE_P1[0003610e] Mar 19 15:45:17 server kernel: bnx2 
>> 0000:02:00.1: eth1: DEBUG: MCP mode[0000b880] state[80008000] evt_mask[00000500] Mar 19 15:45:17 server kernel: bnx2 0000:02:00.1: eth1: DEBUG: pc[0800adec] pc[0800aeb0] instr[8fb10014] Mar 19 15:45:17 server kernel: bnx2 0000:02:00.1: eth1: DEBUG: shmem states:
>> Mar 19 15:45:17 server kernel: bnx2 0000:02:00.1: eth1: DEBUG: 
>> drv_mb[0103000f] fw_mb[0000000f] link_status[0000006f] 
>> drv_pulse_mb[0000432b] Mar 19 15:45:17 server kernel: bnx2 
>> 0000:02:00.1: eth1: DEBUG: dev_info_signature[44564903] 
>> reset_type[01005254] condition[0003610e] Mar 19 15:45:17 server 
>> kernel: bnx2 0000:02:00.1: eth1: DEBUG: 000003cc: 44444444 44444444 
>> 44444444 00000a3c Mar 19 15:45:17 server kernel: bnx2 0000:02:00.1: 
>> eth1: DEBUG: 000003dc: 0ffeffff 0000ffff ffffffff ffffffff Mar 19 
>> 15:45:17 server kernel: bnx2 0000:02:00.1: eth1: DEBUG: 000003ec: 
>> 00000000 00000000 00000000 00000002 Mar 19 15:45:17 server kernel: 
>> bnx2 0000:02:00.1: eth1: DEBUG: 0x3fc[0000ffff] Mar 19 15:45:17 server 
>> kernel: bnx2 0000:02:00.1: eth1: <--- end MCP states dump ---> Mar 19 
>> 15:45:17 server kernel: bnx2 0000:02:00.1: eth1: NIC Copper Link is 
>> Down Mar 19 15:45:20 server kernel: bnx2 0000:02:00.1: eth1: NIC 
>> Copper Link is Up, 1000 Mbps full duplex
>> ----------------------------------------------------------------------
>> -------------------------------------------------------------
>> 
>> Does anyone know that problem?
>> 
>> System is Centos 6.3 Kernel
>> Linux server 2.6.32-279.5.2.el6.centos.plus.x86_64 #1 SMP Fri Aug 24 
>> 00:25:34 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux
>> 
>> 
>> Thanks
>> Hartmut
>> 
>> _______________________________________________
>> CentOS mailing list
>> CentOS at centos.org
>> http://lists.centos.org/mailman/listinfo/centos
> 
> 
> _______________________________________________
> CentOS mailing list
> CentOS at centos.org
> http://lists.centos.org/mailman/listinfo/centos
> _______________________________________________
> CentOS mailing list
> CentOS at centos.org
> http://lists.centos.org/mailman/listinfo/centos