Comware

 View Only
Expand all | Collapse all

LACP trunk to PaloAlto FW

This thread has been viewed 2 times
  • 1.  LACP trunk to PaloAlto FW

    Posted Nov 22, 2016 02:53 AM

    Hi I need some help, finding a LACP-error.

    I have made some LACP trunks, all works fine, but one does not ...

    hawe do I find some error-log ?

    I have made a LACP :

    interface Bridge-Aggregation20
    port link-type trunk
    undo port trunk permit vlan 1
    port trunk permit vlan 3 to 4
    link-aggregation mode dynamic

    interface Ten-GigabitEthernet1/1/5
    port link-mode bridge
    port link-type trunk
    undo port trunk permit vlan 1
    port trunk permit vlan 3 to 4
    port link-aggregation group 20

    interface Ten-GigabitEthernet2/1/5
    port link-mode bridge
    port link-type trunk
    undo port trunk permit vlan 1
    port trunk permit vlan 3 to 4
    port link-aggregation group 20

    but, a simple ping-test have about 60% packet-loss :-(

    If I disable one interface, it works fine..

    any ideers ??

     



  • 2.  RE: LACP trunk to PaloAlto FW

    Posted Nov 22, 2016 03:23 AM

    What is the output of the display link-aggregation verbose Bridge-Aggregation 20 command?



  • 3.  RE: LACP trunk to PaloAlto FW

    Posted Nov 22, 2016 03:29 AM

    [dc2.core1]display link-aggregation verbose Bridge-Aggregation 20
    Loadsharing Type: Shar -- Loadsharing, NonS -- Non-Loadsharing
    Port Status: S -- Selected, U -- Unselected,
    I -- Individual, * -- Management port
    Flags: A -- LACP_Activity, B -- LACP_Timeout, C -- Aggregation,
    D -- Synchronization, E -- Collecting, F -- Distributing,
    G -- Defaulted, H -- Expired

    Aggregate Interface: Bridge-Aggregation20
    Aggregation Mode: Dynamic
    Loadsharing Type: Shar
    Management VLAN : None
    System ID: 0x8000, e8f7-242c-9537
    Local:
    Port Status Priority Oper-Key Flag
    --------------------------------------------------------------------------------
    XGE1/1/5 S 32768 1 {ACDEF}
    XGE2/1/5 S 32768 1 {ACDEF}
    Remote:
    Actor Partner Priority Oper-Key SystemID Flag
    --------------------------------------------------------------------------------
    XGE1/1/5 32 32768 48 0x8000, d4f4-be76-9401 {CDEF}
    XGE2/1/5 33 32768 48 0x8000, d4f4-be76-9401 {CDEF}



  • 4.  RE: LACP trunk to PaloAlto FW

    Posted Nov 22, 2016 06:43 AM

    Interesting.

    Note how the Remote partner (the Palo Alto Firewall with MAC Addressd 4f:4:be:76:94:01) is shown as {CDEF} flagged and not as {ACDEF} flagged as it should be (the A letter means "LACP Activity" as per command provided legend).

    Try to permit VLAN 1 (the Management VLAN) instead of setting it as not permitted either on interface Bridge-Aggregation20 and on its members (port trunk permit vlan 1).

    Eventually re-do the BAGG configuration from scratch (remove cables, reconfigure, re-connect cables) since in Comware based Switches the order of steps for Port Trunking configuration is important (to avoid mismatches between the Logical Interface BAGG20 and its members physical ports).

    Also provide configuration of LACP Port Trunking on the Palo Alto Firewall side <-- that could be the very culprit.

    Do these commands to start troubleshooting (Switch side):

    • display interface brief | include UP (limiting to copy and paste the relevant physical interfaces XGE1/1/5 and XGE2/1/5 and the logical interface BAGG20).
    • display lldp neighbor-information list
    • display link-aggregation summary
    • reset lacp statistics
    • display link-aggregation summary (again to see how zeroed statistics change, if any).

    Which Switch are you using (Model, Firmware version)?

    What is the Palo Alto Firewall configuration regarding its two ports (LACP) Port Trunking=



  • 5.  RE: LACP trunk to PaloAlto FW

    Posted Nov 22, 2016 08:57 AM

    hi

    yes, I have tryed to start over (removed all, and start over...)

    but the result is the same (allmost), my "ping-test" have a bit too high packet-loss :

    80 packets transmitted, 40 received, 50% packet loss, time 79044ms
    rtt min/avg/max/mdev = 0.536/0.605/0.775/0.056 ms

    but I did change a setting in the PaloAlto, witch gave this : (picture uploadet)

    display link-aggregation verbose Bridge-Aggregation 20
    Loadsharing Type: Shar -- Loadsharing, NonS -- Non-Loadsharing
    Port Status: S -- Selected, U -- Unselected,
    I -- Individual, * -- Management port
    Flags: A -- LACP_Activity, B -- LACP_Timeout, C -- Aggregation,
    D -- Synchronization, E -- Collecting, F -- Distributing,
    G -- Defaulted, H -- Expired

    Aggregate Interface: Bridge-Aggregation20
    Aggregation Mode: Dynamic
    Loadsharing Type: Shar
    Management VLAN : None
    System ID: 0x8000, e8f7-242c-9537
    Local:
    Port Status Priority Oper-Key Flag
    --------------------------------------------------------------------------------
    XGE1/1/5 S 32768 1 {ACDEF}
    XGE2/1/5 S 32768 1 {ACDEF}
    Remote:
    Actor Partner Priority Oper-Key SystemID Flag
    --------------------------------------------------------------------------------
    XGE1/1/5 32 32768 48 0x8000, d4f4-be76-9401 {ACDEF}
    XGE2/1/5 33 32768 48 0x8000, d4f4-be76-9401 {ACDEF}

     

    display interface brief | include UP :


    BAGG20 UP 20G(a) F(a) T 1 Uplink til dc2.fw1
    XGE1/1/5 UP 10G(a) F(a) T 1
    XGE2/1/5 UP 10G(a) F(a) T 1

    display lldp neighbor-information list
    Chassis ID : * -- -- Nearest nontpmr bridge neighbor
    # -- -- Nearest customer bridge neighbor
    Default -- -- Nearest bridge neighbor
    System Name Local Interface Chassis ID Port ID
    dc2.man XGE1/1/23 e8f7-2446-2282 Ten-GigabitEthernet1/0/49
    Multi 7.0 XGE1/1/24 0017-a4b6-c200 200
    dc2.core1 XGE1/2/23 e8f7-242c-9537 Ten-GigabitEthernet2/2/24
    dc2.core1 XGE1/2/24 e8f7-242c-9537 Ten-GigabitEthernet2/2/23
    dc2.man M-GE0/0/0 e8f7-2446-2282 GigabitEthernet1/0/2
    dc2.man XGE2/1/23 e8f7-2446-2282 Ten-GigabitEthernet1/0/50
    dc2.core1 XGE2/2/23 e8f7-242c-9537 Ten-GigabitEthernet1/2/24
    dc2.core1 XGE2/2/24 e8f7-242c-9537 Ten-GigabitEthernet1/2/23
    [dc2.core1]display link-aggregation summary
    Aggregation Interface Type:
    BAGG -- Bridge-Aggregation, BLAGG -- Blade-Aggregation, RAGG -- Route-Aggregation
    Aggregation Mode: S -- Static, D -- Dynamic
    Loadsharing Type: Shar -- Loadsharing, NonS -- Non-Loadsharing
    Actor System ID: 0x8000, e8f7-242c-9537

    AGG AGG Partner ID Selected Unselected Individual Share
    Interface Mode Ports Ports Ports Type
    --------------------------------------------------------------------------------
    BAGG1 D 0x8000, 0000-0000-0000 0 2 0 Shar
    BAGG5 D 0x8000, e8f7-2446-2282 2 0 0 Shar
    BAGG20 D 0x8000, d4f4-be76-9401 2 0 0 Shar
    BAGG21 D 0x8000, 0000-0000-0000 0 2 0 Shar

    it is running on at HPE FF 5930-4Slot Switch, Software Version 7.1.045, Release 2422P01

    the normal setup for a PaloAlto : https://live.paloaltonetworks.com/t5/Configuration-Articles/How-to-Configure-LACP/ta-p/65837 



  • 6.  RE: LACP trunk to PaloAlto FW

    Posted Nov 22, 2016 10:20 AM

    OK, better...now the run show lacp aggregate-ethernet all on the Palo Alto Firewall what output produces?



  • 7.  RE: LACP trunk to PaloAlto FW

    Posted Nov 22, 2016 10:27 AM

    fw1(active)> show lacp aggregate-ethernet all

    LACP:

    **********************************************************************************
    AE group: ae1
    Members: Bndl Rx state Mux state Sel state
    ethernet1/17 yes Current Tx_Rx Selected
    ethernet1/18 yes Current Tx_Rx Selected
    Status: Enabled
    Mode: Active
    Rate: Slow
    Max-port: 8
    Fast-failover: Disabled
    Pre-negotiation: Disabled
    Local: System Priority: 32768
    System MAC: d4:f4:be:76:94:01
    Key: 48
    Partner: System Priority: 32768
    System MAC: e8:f7:24:2c:95:37
    Key: 1
    Port State
    --------------------------------------------------------------------------------
    Interface Port
    Number Priority Mode Rate Key State
    --------------------------------------------------------------------------------
    ethernet1/17 32 32768 Active Slow 48 0x3D
    Partner 5 32768 Active Slow 1 0x3D

    ethernet1/18 33 32768 Active Slow 48 0x3D
    Partner 210 32768 Active Slow 1 0x3D

    Port Counters
    --------------------------------------------------------------------------------
    Interface LACPDUs Marker Marker Response Error
    Sent Recv Sent Recv Sent Recv Unknown Illegal
    --------------------------------------------------------------------------------
    ethernet1/17 531 491 0 0 0 0 0 0
    ethernet1/18 529 489 0 0 0 0 0 0



  • 8.  RE: LACP trunk to PaloAlto FW

    Posted Nov 22, 2016 10:53 AM

    Where exactly are you executing the Ping command (source of the ping: a VM on the ESXi, the ESXi itself, the 5930 Switch itself or what else)? and what's about the destination of the ping command?

    Are you working on a IRF Stack of 5930 switches or the 5930 is just a single unit?



  • 9.  RE: LACP trunk to PaloAlto FW

    Posted Nov 23, 2016 02:57 AM

    the ping-test is all running from our central FW (two switches "away", from this setup). I have two destinations (in this setup) boath configured as a lacp trunks direct connectet to this 5930-IRF stack... one set (other produckt) running just fine and the other (this PA firewalls) a bit of packet-loss...

    So all testing are from and to real network aplaince-boxes (no vm)



  • 10.  RE: LACP trunk to PaloAlto FW

    Posted Nov 23, 2016 08:19 AM

    The output of the command show lacp aggregate-ethernet all on the PA looks good (the exchange of LACPDUs between both Trunk's ends happens and also the "ae1" LACP Trunk is shown as Active and as Enabled...against the HPE switch, as the Partner MAC Address shows).

    What's about checking "ae1" Aggregate Ethernet Interface's Link Settings (Link Speed, Link Duplex and Link State) for any possible mismatch?

    What's about performing the ICMP Ping from the HPE switch against the PA Firewall? do you obtain a similar or identical pattern (50% packet loss) or what?

    On both Switch and Firewall sides: do all involved physical interfaces report link flapping on transmission errors?

    Sorry, what do you mean with "two switches "away", from this setup"? Isn't the HPE 5930 directly trunked to the PA Firewall (so why you pointed out "two switches away"?) as we expected?

    What I understood is that the PA Firewall is at one LACP Trunk's end and another Switch is at another LACP Trunk's end, The HPE 5930 is just sitting at the center...separately providing these two independent LACP Trunks. Is that right?



  • 11.  RE: LACP trunk to PaloAlto FW

    Posted Nov 23, 2016 10:05 AM

    FYI... the best practice here is to only change Bridge-Aggregation parameters from the Bridge-Aggregation interface.  Comware will take those settings and apply them to all of the members of the Bridge Aggregation group.  That said, under certain conditions, the settings are not migrated down.  You need to remove the differences from the members and the Bridge interface and then re-apply them at the bridge interface.  That should bring them all into sync.

    David



  • 12.  RE: LACP trunk to PaloAlto FW

    Posted Nov 24, 2016 03:34 AM

    every interface setting is "auto".

    The switch only have ip on out of band mangement, we use the switch as a layer-2 an not routing/ip/layer-3... thats is the fw´s job.. so pinging from the switch is not posible.

    the frase "two switches away" is to my oher internal firewall (the one I use to ping from) I can understand it is confusing, I am usin an old firewall to ping a new firewall :-)

    the PA firewall is direckt connectet to this 5930-irf-switches, and with both interfaces active = pacet loss.

    the same 5930-irf switch, I have connectet a set of F5 loadballancers also in LACP aggregate - trunks .. = works perfecktly



  • 13.  RE: LACP trunk to PaloAlto FW

    Posted Nov 24, 2016 10:54 AM
    Which exact PAN Firewall appliance model and PAN-OS version are you using? I'm curious about its XGE (10Gbps) physical interfaces...