Hello
Can you please describe what is the nature of test you are trying to do ? as far as we understand you've installed the latest MLNX_OFED and FW
- Yes this is correct! I have installed the latest MLNX_OFED and FW on CentOS 6.5
My Problem is the Port Mapping on the Switches:
IP Information:
HP Bladeserver CentOS 6.5 with 2 x HP NC542m (Mellanox Network Card):
eth0: 192.168.10.170
eth1: 192.168.10.171
eth2: 192.168.10.172
eth3: 192.168.10.173
eth4: 192.168.10.174
eth5: 192.168.10.175
My privat windows Laptop:
192.168.10.200
I have 3 Switches on the HP Bladecenter C3000:
Bay1: Cisco Catalyst Blade Switch 3020 for HP <- Portmapping Internal Network Card of the Bladeserver (et0, eth1)
Bay2: Cisco Catalyst Blade Switch 3020 for HP <- PortmappingFlex HP NC542m (Mellanox Network Card) Megazinne Port 1 (eth2, eth3)
Bay3: HP BLc Flex-10 10Gb Virtual Connect Ethernet Modul (Firmware Version: 4.20) <- PortmappingFlex HP NC542m (Mellanox Network Card) Megazinne Port 2 (eth4, eth5)
When i make a power off on Bay 1: Cisco Catalyst Blade Switch 3020 all ports are down! I mean i can not ping from my laptop with the IP Adress: 192.168.10.200 this network ports: eth0 and eth1, eth2, eth3, eth4, eth5
This is not correct, because when i power off Bay 1: Cisco Catalyst Blade Switch 3020 the network ports:
eth0 and eth1
are down but not the another ports eth2, eth3, eth4, eth5! The another ports eth2, eth3, eth4, eth5! must be UP!! Because Bay2: Cisco Catalyst Blade Switch 3020 are running and Bay3: HP BLc Flex-10 10Gb Virtual Connect Ethernet Modul are running!
hope you understand me??? All Network Ports connect over 1 Switch!! But this is not correct..
I mean this must be a driver problem because under Windows Server 2008 R2 i dont have this problem!
Under Windows Server 2008 R2:
When i power off the Bay1: Cisco Catalyst Blade Switch 3020:
2 Network Ports are down! But the another ports are active or UP! I can ping the another ports but not the ports thats connect on Bay1: Cisco Catalyst Blade Switch 3020! And this is correct..
HP Level 2 Support says this:
Now it is confirmed that 542Mezz Nic’s in Windows 2008 works fine in all the Bays.
We still see issues with Cent OS and now this looks to be OS specific.
Here the results from the commands:
[root@localhost ~]# ethtool -i eth5
driver: mlx4_en
version: 2.2-1.0.0 (Jun 24 2014)
firmware-version: 2.9.1530
bus-info: 0000:09:00.0
supports-statistics: yes
supports-test: yes
supports-eeprom-access: no
supports-register-dump: no
supports-priv-flags: yes
[root@localhost ~]# ibv_devinfo
hca_id: mlx4_1
transport: InfiniBand (0)
fw_ver: 2.9.1530
node_guid: 78e7:d1ff:ff67:9b10
sys_image_guid: 78e7:d1ff:ff67:9b10
vendor_id: 0x02c9
vendor_part_id: 26468
hw_ver: 0xB0
board_id: HP_0150110010
phys_port_cnt: 2
port: 1
state: PORT_ACTIVE (4)
max_mtu: 4096 (5)
active_mtu: 1024 (3)
sm_lid: 0
port_lid: 0
port_lmc: 0x00
link_layer: Ethernet
port: 2
state: PORT_DOWN (1)
max_mtu: 4096 (5)
active_mtu: 1024 (3)
sm_lid: 0
port_lid: 0
port_lmc: 0x00
link_layer: Ethernet
hca_id: mlx4_0
transport: InfiniBand (0)
fw_ver: 2.9.1530
node_guid: d8d3:85ff:ff68:ac30
sys_image_guid: d8d3:85ff:ff68:ac30
vendor_id: 0x02c9
vendor_part_id: 26468
hw_ver: 0xB0
board_id: HP_0150110010
phys_port_cnt: 2
port: 1
state: PORT_ACTIVE (4)
max_mtu: 4096 (5)
active_mtu: 1024 (3)
sm_lid: 0
port_lid: 0
port_lmc: 0x00
link_layer: Ethernet
port: 2
state: PORT_ACTIVE (4)
max_mtu: 4096 (5)
active_mtu: 1024 (3)
sm_lid: 0
port_lid: 0
port_lmc: 0x00
link_layer: Ethernet