Comware Based
1753695 Members
6059 Online
108799 Solutions
New Discussion

Re: Transition IRF ports from 10Gb ports to QSFP 40Gb ports

 
dsmj00
Occasional Visitor

Transition IRF ports from 10Gb ports to QSFP 40Gb ports

I have two HP 5900 SFP+ switches currently connected via IRF ports using 4 ten gigabit ports in each IRF port.  What I would like to do is gracefully transition from the 10Gb ports to the QSFP ports, ideally during production without disrupting communication.

What I was planning to do is to remove 2 of the physical 10Gb ports from the IRF ports, and then add the two QSFP ports to each IRF port.  After the two QSFP ports are added, remove the remaining 10Gb ports from the IRF ports.  While this logically makes sense to me, I don't know if it is recommended or supported to make changes to the IRF ports while it is in production?  Can someone tell me if this will work, not work, or what the caveats may be?  Below I've included the IRF configuration.

irf mac-address persistent timer
irf auto-update enable
undo irf link-delay

irf member 1 priority 1
irf member 2 priority 2

irf-port 1/1
port group interface Ten-GigabitEthernet1/0/45
port group interface Ten-GigabitEthernet1/0/46
port group interface Ten-GigabitEthernet1/0/47
port group interface Ten-GigabitEthernet1/0/48
#
irf-port 2/2
port group interface Ten-GigabitEthernet2/0/45
port group interface Ten-GigabitEthernet2/0/46
port group interface Ten-GigabitEthernet2/0/47
port group interface Ten-GigabitEthernet2/0/48

display irf configuration
MemberID NewID IRF-Port1 IRF-Port2
1 1 Ten-GigabitEthernet1/0/45 disable
Ten-GigabitEthernet1/0/46
Ten-GigabitEthernet1/0/47
Ten-GigabitEthernet1/0/48
2 2 disable Ten-GigabitEthernet2/0/45
Ten-GigabitEthernet2/0/46
Ten-GigabitEthernet2/0/47
Ten-GigabitEthernet2/0/48

display irf topology
Topology Info
-------------------------------------------------------------------------
IRF-Port1 IRF-Port2
MemberID Link neighbor Link neighbor Belong To
2 DIS --- UP 1 4431-9239-b5dc
1 UP 2 DIS --- 4431-9239-b5dc

 

1 REPLY 1
parnassus
Honored Contributor

Re: Transition IRF ports from 10Gb ports to QSFP 40Gb ports


dsmj00 wrote: What I was planning to do is to remove 2 of the physical 10Gb ports from the IRF ports, and then add the two QSFP ports to each IRF port.  After the two QSFP ports are added, remove the remaining 10Gb ports from the IRF ports.

Hello!

That's interesting...on the HPE 5920 & 5900 Switch Series IRF Configuration Guide (Chapter "Setting up an IRF fabric", Paragraph "Hardware compatibility") there are statements regarding restrictions in deploying a 40-GE QSFP+ port splitted into its four 10-GE breakout interfaces as members of a logical IRF port, the first restriction is:

"You must use all or none of the four 10-GE breakout interfaces as IRF physical interfaces. The four
breakout interfaces can be bound to different IRF ports."

This - I think - means that you, at the end of your transition, must have binded all four splitted ports of the 40-GE QSFP+ or none at all (and here you should be OK, this shouldn't be a stopper for your transition plan...since that is exactly what are you trying to achieve), then a second subsequent restriction is:

"Before you bind one 10-GE breakout interface to an IRF port or remove it from the IRF port, you must
shut down all the other 10-GE breakout interfaces. If any of the breakout interfaces are in up state,
the bind or remove action will fail."

So basically your plan should work, from what I understood, provided that *essentially* you bind all four 10-GE breakout interfaces of the 40-GE QSFP+ in just one initial step avoiding to do that sequentially in two separate steps as described (remove 2x10 GbE interfaces from IRF port, add 2x10-GE breakout interfaces of 40-GE QSFP+, remove remaining 2x10 GbE interfaces from IRF port, add remaining 2x10-GE breakout interfaces).

At least this is what I understood reading those restrictions applied to 40-GE QSFP+, if splitted...hope not be wrong.

If you're going to use the 40-GE QSFP+ port as a single link (not splitted) reported restrictions should not apply.


I'm not an HPE Employee
Kudos and Accepted Solution banner