nv_edge_upgrade.exp results in loosing rack0 interfaces config

[olegch]

Hello All,

We have a cluster of two 9010 with 5.1.0. and 36x10G in each node.

After Rack-By-Rack upgrade to 5.1.1 we get cluster but without any configuration on rack0 interfaces (Mgmt, IRL, bundle members). ICL looks good. IRLs from rack0 side are in err-disable state.

Script logs attached. Also I attached all related commands issued after upgrade. And diff of configs after/before upgrade.

Have somebody a glue what happend wrong ?

[Eddie Chami]

Hi Oleg,

Can you please run the following commands in global and admin plane "cfs check". Could you also remove the logging disable line from the script, so i see where the config is failing at bootup? Send the bootlog of Rack0.

Regards,

Eddie.

[olegch]

Hi echami,

Thank you for your response.

I've just made "load tftp://ip/saved.cfg ; commit replace;" so I can't get appropriate bootlog from Rack0 now. But I still want to recreate the issue. So I tried to make a rollback to previous 5.1.0 state but it failed. Can you suggest something. Commands log attached.

PS: ASR9k_Upgrade_Downgrade_Procedure_IOSXR_Rel_511.pdf mentions downgrade SMU CSCui99165 . As far as I understand I should install this SMU to downgrade from 5.1.1 to 5.1.0 , right ? But I can't find it in download area.

[Eddie Chami]

Oleg,

Sorry for the delay, been working on recreating this, and i finally can, just give me sometime i will get back to you with some further information.

Regards

Eddie.

[olegch]

Hi echami,

Great news. Glad to hear you can recreate that. I am looking forward for further information.

We are not using our cluster in production environment yet, so we have plenty of time to play with it.

Thank you for your attention.

[Eddie Chami]

Could you private message me your ftp/tftp server, so i can upload a SMU for you?

[Eddie Chami]

Another workaround you can use to downgrade is:

#admin install activate *5.1.0*.

So i don't need the ftp/tftp. We will fix the issue with the downgrade though, and we will upgrade the 5.1.1 upgrade/downgrade instructions

Regards

Eddie.

[olegch]

Hello Eddie,

We' ve downgraded cluster successfully to 5.1.0 (without loosing rack0 interfaces config). Then we started script again (logs are attached) with commented "log console disable" as you requested. We've get the same results (all rack0 interfaces config lost). Some commands results from 0/RSP0 after upgrade attached. Then we "reload rack0" ( console logs from 0/RPS0 and 1/RSP0 also attached).