ONTAP Cluster Switch: BES-53248 3.8.0.2 to 3.11.1.2 Upgrade

on

Some preparatory notes from BES-53248 upgrade from 3.8.0.2 to 3.11.1.2.

1) Check this page - NetApp Support Site - BES-53248 EFOS Software Download - and the switch compatibility matrix to verify your EFOS and RCF file versions.


Currently, EFOS 3.11.1.2 and ClusterRCF_v1.10 are the recommended EFOS and RCF (Reference Configuration File) versions.

2) Download the EFOS version and RCF.

Download the RCF from here: https://mysupport.netapp.com/site/products/all/details/broadcom-cluster-switches/downloads-tab


Note: You will need a HTTP server (other protocols allowed) which the switch can reach, to download the software and RCF files.

3) Review the latest NetApp BES-53248 documentation:


It's also worth checking on kb.netapp.com if there are any KBs relating to the EFOS version you want to go to (as of writing there is nothing for 3.11.1.2.)

Walkthrough

Note 1: Using the above documentation and my previous post BES-53248 Upgrade from 3.4.4.6 to 3.10.0.3 (cosonok.com).
Note 2: Assuming you're not using the FIPS compliant version and are not changing to a FIPS compliant version (which came out with EFOS 3.7.) That is "show fips status" says "Invalid input" and not "System running in FIPS mode."
Note 3: ONTAP version is 9.8 or greater.

1) Prepare

CLU::> set adv
CLU::*> cluster ping-cluster -node ANODENAME
CLU::*> network device-discovery show
CLU::*> system switch ethernet show
CLU::*> network port show -ipspace Cluster
CLU::*> network interface show -vserver Cluster
CLU::*> cluster show
CLU::*> system node autosupport invoke -node \* -type all -message MAINT=4h

2) Install the EFOS software


Firstly on switch 2:

CS2# ping HTTP_SERVER_IP
CS2# show bootvar
CS2# show version
CS2# copy http://HTTP_SERVER_IP/EFOS-3.11.1.2.stk active
CS2# show bootvar
CS2# write mem
CS2# reload
CS2# show version

Verify cluster ports are healthy (LIFs should have returned home):

CLU::*> network port show -ipspace Cluster
CLU::*> network interface show -vserver Cluster
CLU::*> cluster show

Then repeat on switch 1:

CS1# ping HTTP_SERVER_IP
CS1# show bootvar
CS1# show version
CS1# copy http://HTTP_SERVER_IP/EFOS-3.11.1.2.stk active
CS1# show bootvar
CS1# write mem
CS1# reload
CS1# show version

Verify cluster ports are healthy (LIFs should have returned home):

CLU::*> network port show -ipspace Cluster
CLU::*> network interface show -vserver Cluster
CLU::*> cluster show

3) Install the RCF


CLU::*> network interface modify -vserver Cluster -lif * -auto-revert false

Firstly we do cluster switch 2.

CS2# config t
CS2(Config)# interface 0/1-0/16
CS2(Interface 0/1-0/16)# shutdown
CS2(Interface 0/1-0/16)# exit
CS2(Config)# exit

CLU::*> network interface show -vserver Cluster
CLU::*> cluster show

CS2# show running-config

Then to remove the previous configuration (could use the previous "clear config" method if have onsite console access): Copy and paste the following commands to remove the previous RCF configuration (depending on the previous RCF version used, some commands might generate an error if a particular setting is not present):

NOTE: This procedure did not work for me. Since I was on site I was able to use the "clear config" method instead. I have escalated to NetApp.

clear config interface 0/1-0/56
y
clear config interface lag 1
y
configure
deleteport 1/1 all
no policy-map CLUSTER
no policy-map WRED_25G
no policy-map WRED_100G
no class-map CLUSTER
no class-map HA
no class-map RDMA
no classofservice dot1p-mapping
no random-detect queue-parms 0
no random-detect queue-parms 1
no random-detect queue-parms 2
no random-detect queue-parms 3
no random-detect queue-parms 4
no random-detect queue-parms 5
no random-detect queue-parms 6
no random-detect queue-parms 7
no cos-queue min-bandwidth
no cos-queue random-detect 0
no cos-queue random-detect 1
no cos-queue random-detect 2
no cos-queue random-detect 3
no cos-queue random-detect 4
no cos-queue random-detect 5
no cos-queue random-detect 6
no cos-queue random-detect 7
exit
vlan database
no vlan 17
no vlan 18
exit

CS2# write memory
CS2# reload

Note: If additional port licenses have been installed on the switch, you must modify the RCF to configure the additional licensed ports. See Activate newly licensed ports for details.

CS2# copy http://HTTP_SERVER_IP/BES-53248-RCF-v1.10-Cluster-HA.txt nvram:script BES-53248-RCF-v1.10-Cluster-HA.scr
CS2# script list
CS2# script apply BES-53248-RCF-v1.10-Cluster-HA.scr
CS2# show clibanner
CS2# show port all | exclude Detach
CS2# show running-config
CS2# write memory
CS2# reload

CS2# config t
CS2(Config)# interface 0/1-0/16
CS2(Interface 0/1-0/16)# no shutdown
CS2(Interface 0/1-0/16)# exit
CS2(Config)# exit

CLU::*> network port show -ipspace Cluster
CLU::*> network interface show -vserver Cluster
CLU::*> cluster show

The repeat the above for cluster switch 1.

Once switch 2 and switch 1 are complete.

CLU::*> system switch ethernet show
CLU::*> network interface modify -vserver Cluster -lif * -auto-revert true

Finally, verify all is good:

CSW# show interfaces status all | exclude Detach
CSW# show port-channel 1/1

CLU::*> network port show -ipspace Cluster
CLU::*> network interface show -vserver Cluster
CLU::*> cluster show
CLU::*> system node autosupport invoke -node * -type all -message MAINT=END

Comments

Post a Comment