Comware

Hi guys,

I have a HP 5130 7.1.045, Release 3111P02 in a Switch Stack .

I get how to upgrade the Firmware, my only problem is how do i do this without connection loss of both members?

At the moment it looks like this

#Upload the Firrmware to switch
ftp 127.0.0.1
#activate binary mode
binary
get 5130EI-CMW710-R3208P03.ipe
get 5130ei-cmw710-packet-capture-r3208p03.bin


#upload ipe file to flash:/
#check startup software:
display boot-loader
#change startup software to newest build
boot-loader file flash:/5130EI-CMW710-R3208P03.ipe slot 2 main
boot-loader file flash:/5130EI-CMW710-R3208P03.ipe slot 1 main
#verify that new software is set as next startup software
display boot-loader
- disable software auto-update
undo irf auto-update enable
- save config
save
- verify upgrade modus -> incompatible
display version comp-matrix file system flash:/5130ei-cmw710-system-r3208p03.bin

- WORKAROUND: hard reboot (both members at the same time)
reboot
y

 My Question is how do I reboot 1 Member at a time to upgrade the Firmware?

I did see some guides where they use the ISSU command but this doesnt exist on this Switch. 

Does it actually work with 

reboot slot 1

reboot slot 2 

Thank you for your help.

Best Regards

Burak

---

wrote: I get how to upgrade the Firmware, my only problem is how do i do this without connection loss of both members?

Well...since ISSU on HPE 5130 is not an option you are forced to reboot sequentially your IRF Members to complete the software update procedure therefore a possible way to prevent (connectivity) loss for hosts connected to IRF stack is that those devices (that's primarily related to Servers hosts) are dual homed linked to both IRF Members at the same time (with dual homed linked I mean that an Host should be uplinked to both IRF Members concurrently using NIC ports teaming/bonding/aggregation - call it with the name you like more - using LACP...so one Host NIC link goes to IRF Member 1 and the other to IRF Member 2, on IRF you must define a LAG with LACP exactly like you do that on the Host's NIC to cope with the teaming defined on the Host or viceversa, that's just an example)...and, I would add, that your two members IRF stack adopts a sort of split brain management mechanism (MAD Multi Active Detection) to prevent/manage split brain scenarios that could happen.

Does that sound logical to you too?