cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Announcements

678
Views
0
Helpful
0
Replies
Robert Burns
Cisco Employee

Ever wonder about deterministic failover with VPC-HM?

I've heard a few people asking about setting failover preferences for mac-pinning and figured it would be a great opportunity to highlight a great feature introduced in version 1.4a.

As of version 1.4a and later you can now set the "order" of failover sub groups.

Ex.  Let's say you want to do some traffic engineering and keep your Service Console/Management traffic & Vmotion traffic pinned to separate uplinks.  We know we can use sticky pinning to force these virtual interfaces to separate sub group uplink interface, but what happens when there is a failure with one of these links?  Previously, when a sub group member failed with pinned interfaces it would be dynamically re-pinned in a round-robin fashion to the remaining operational members.  New functionality was added to force the order of which backup sub groups will be utilized.

To specify multiple sub-groups for failover you can use the following pinning command in either the vEth interface or Port profile level.

1000v(config-if) pinning id <primary pinning SGID> backup <list of backup pinning SGIDs>

So let's say we had four uplink vmnics 0 through 3.  We'd like to keep our management and vmotion traffic separated if possilbe.

vmnic0 = subgroup ID 0 (primary for management)

vmnic1 = subgroup ID 1 (secondary for mangement, tertiary for vmotion)

vmnic2 = subgroup ID 2 (secondary for vmotion, tertiary for management)

vmnic3 = subgroup ID 3 (primary for vmotion)

Our port profiles might look like this.

port-profile type ethernet system-uplink

  vmware port-group

  switchport trunk allowed vlan 1-3967,4048-4093

  switchport mode trunk

  channel-group auto mode on mac-pinning

  no shutdown

  system vlan 10, 3001-3002

  state enabled

port-profile type vethernet management

  vmware port-group

  no shutdown

  switchport mode access

  switchport access vlan 10

  system vlan 10

  state enabled

  pinning id 0 backup 1 2

port-profile type vethernet vmotion

  vmware port-group

  no shutdown

  switchport mode access

  switchport access vlan 99

  state enabled

  pinning id 3 backup 2 1

>>> How to verify your current pinning?  Which uplink is my virtual interfaces using to exit the VEM?

See the vmnic mapping to "SG_ID", then look at your virtual interfaces and the "Pinned_SGID".

n1000v-BL1(config-port-prof)# module vem 3 execute vemcmd show port-old

  LTL    IfIndex   Vlan    Bndl  SG_ID Pinned_SGID  Type  Admin State  CBL Mode   Name

    6          0      1 T     0     32          32  VIRT     UP    UP    1  Trunk vns

    8          0   3969       0     32          32  VIRT     UP    UP    1 Access

    9          0   3969       0     32          32  VIRT     UP    UP    1 Access

   10          0   3001       0     32           0  VIRT     UP    UP    1 Access

   11          0   3968       0     32          32  VIRT     UP    UP    1 Access

   12          0   3002       0     32           1  VIRT     UP    UP    1 Access

   13          0      1       0     32          32  VIRT     UP    UP    0 Access

   14          0   3971       0     32          32  VIRT     UP    UP    1 Access

   15          0   3971       0     32          32  VIRT     UP    UP    1 Access

   16          0      1 T     0     32          32  VIRT     UP    UP    1  Trunk arp

   17   25008000      1 T   305      0          32  PHYS     UP    UP    1  Trunk vmnic0

   18   25008040      1 T   305      1          32  PHYS     UP    UP    1  Trunk vmnic1

   18   25008060      1 T   305      2          32  PHYS     UP    UP    1  Trunk vmnic2

   18   25008070      1 T   305      3          32  PHYS     UP    UP    1  Trunk vmnic3

   49   1c000030     19       0     32           0  VIRT     UP    UP    1 Access vmk0

   50   1c0000a0     99       0     32           1  VIRT     UP    UP    1 Access vmk1

   51   1c000000      1       0     32           2  VIRT     UP    UP    1 Access vmk2

   52   1c000100     25       0     32           3  VIRT     UP    UP    1 Access ucs-dashboard-v1.0.eth0

   53   1c000090   3002       0     32           0  VIRT     UP    UP    1 Access n1000v-BL1-Primary.eth2

   54   1c000080     19       0     32           1  VIRT     UP    UP    1 Access n1000v-BL1-Primary.eth1

   55   1c000070   3001       0     32           2  VIRT     UP    UP    1 Access n1000v-BL1-Primary.eth0

   56   1c000180   3002       0     32           3  VIRT     UP    UP    1 Access n1000v-BL1-Secondary.eth2

   57   1c000170     19       0     32           0  VIRT     UP    UP    1 Access n1000v-BL1-Secondary.eth1

   58   1c000120   3001       0     32           1  VIRT     UP    UP    1 Access n1000v-BL1-Secondary.eth0

   60   1c000040     25       0     32           2  VIRT     UP    UP    1 Access Sharepoint-CAE ethernet0

   61   1c000010     19       0     32           3  VIRT     UP    UP    1 Access SV Team Sharepoint ethernet1

  305   16000001      1 T     0     32          32  CHAN     UP    UP    1  Trunk

So what does this do?  Well it's going to give us a predictive behavior as far as failover is concerned.  It's a nice feature to be able to engineer traffic even in the event of a link failure.

Any questions, fire away.

Regards,

Robert

0 REPLIES 0
Content for Community-Ad