Juniper_junos_software ISSU error for SRX

[stuartianbrown]

When attempting an ISSU using the module juniper_junos_software on an SRX550 an error is thrown for the number of routing engines:

An exception occurred during task execution. To see the full traceback, use -vvv. The error was: TypeError: ISSU/NSSU requires Multi RE setup
fatal: [LAB-SRX550-node0]: FAILED! => {"changed": false, "failed": true, "module_stderr": "Traceback (most recent call last):\n  File \"/tmp/ansible_lYr78N/ansible_module_juniper_junos_software.py\", line 788, in <module>\n    main()\n  File \"/tmp/ansible_lYr78N/ansible_module_juniper_junos_software.py\", line 701, in main\n    ok = junos_module.sw.install(**install_params)\n  File \"/usr/lib/python2.7/site-packages/jnpr/junos/utils/sw.py\", line 804, in install\n    'ISSU/NSSU requires Multi RE setup')\nTypeError: ISSU/NSSU requires Multi RE setup\n", "module_stdout": "", "msg": "MODULE FAILURE", "rc": 0}

However, running the command on the device itself works:

request system software in-service-upgrade junos-srxsme-12.3X48-D50.6-domestic.tgz no-sync

The playbook I'm using is:

---
- name: Install Junos OS
  hosts: LAB-SRX550-node0
  gather_facts: no
  roles:
    - Juniper.junos
  connection: local

  tasks:
    - name: Install Junos OS Package
      juniper_junos_software:
        provider: "{{ cli }}"
        version: 12.1X44-D40.2
        local_package: /etc/ansible/images/junos-srxsme-12.1X44-D40.2-domestic.tgz
        issu: yes
      register: response
    - name: Print the response
      debug:
        var: response

[prime001]

Did anyone ever resolve this?

[rahkumar651991]

@prime001 - We now support single-RE ISSU.
If still facing the issue, kindly share the netconf trace for it.

https://www.juniper.net/documentation/en_US/junos/topics/example/netconf-traceoptions-configuring.html

[mersl]

I hit the same issue today as I tried to ISSU upgrade a SRX345 cluster. I assume, I have quite recent modules installed (please check below python and ansible module versions).

Current running Junos: 18.4R3
Target Junos: 19.4R3-S3.3

I did set the traceoptions as advised at the earlier post and it stopped on VC information collection:

May 18 14:41:32 [NETCONF] - [62457] Incoming: <nc:rpc xmlns:nc="urn:ietf:params:xml:ns:netconf:base:1.0" message-id="urn:uuid:1468ae56-3e79-42ca-bd69-8079b22e69b7"></nc:rpc>]]>]]>
May 18 14:41:32 [NETCONF] - [62457] Outgoing:
May 18 14:41:32 [NETCONF] - [62457] Outgoing:
May 18 14:41:32 [NETCONF] - [62457] Outgoing: protocol
May 18 14:41:32 [NETCONF] - [62457] Outgoing: operation-failed
May 18 14:41:32 [NETCONF] - [62457] Outgoing: error
May 18 14:41:32 [NETCONF] - [62457] Outgoing: syntax error
May 18 14:41:32 [NETCONF] - [62457] Outgoing:
May 18 14:41:32 [NETCONF] - [62457] Outgoing: get-virtual-chassis-information
May 18 14:41:32 [NETCONF] - [62457] Outgoing:
May 18 14:41:32 [NETCONF] - [62457] Outgoing:
May 18 14:41:32 [NETCONF] - [62457] Debug: The last token parsed by mgd [62457] was [get-virtual-chassis-information] and gram data current token [get-virtual-chassis-information]
May 18 14:41:32 [NETCONF] - [62457] Outgoing:
May 18 14:41:32 [NETCONF] - [62457] Outgoing: ]]>]]>

BTW: the previous RPC call was - checking REs

May 18 14:41:30 [NETCONF] - [62457] Incoming: <nc:rpc xmlns:nc="urn:ietf:params:xml:ns:netconf:base:1.0" message-id="urn:uuid:7ef9f35b-d281-4de1-9cc7-187ab8ca5186"></nc:rpc>]]>]]>
May 18 14:41:30 [NETCONF] - [62457] Outgoing:
May 18 14:41:30 [NETCONF] - [62457] Outgoing:
May 18 14:41:30 [NETCONF] - [62457] Outgoing:
May 18 14:41:31 [NETCONF] - [62457] Outgoing:
...
May 18 14:41:31 [NETCONF] - [62457] Outgoing:
May 18 14:41:31 [NETCONF] - [62457] Outgoing:
May 18 14:41:31 [NETCONF] - [62457] Outgoing:
May 18 14:41:31 [NETCONF] - [62457] Outgoing:
May 18 14:41:31 [NETCONF] - [62457] Outgoing: ]]>]]>

also showing information (serial numbers, fpc etc etc) for both SRX cluster nodes as expected.

---

$ ansible-galaxy collection list

~/.ansible/collections/ansible_collections
Collection Version

---

ansible.netcommon 2.0.2
ansible.utils 2.1.0
community.general 3.0.2
juniper.device 1.0.0
junipernetworks.junos 2.1.0

$ python3 -m pip list
Package Version

---

ansible 3.3.0
ansible-base 2.10.9
bcrypt 3.2.0
cffi 1.14.5
cryptography 3.4.7
importlib-resources 5.1.2
Jinja2 2.11.3
jmespath 0.10.0
junos-eznc 2.6.0+3.g2358a3b
jxmlease 1.0.3
lxml 4.6.3
MarkupSafe 1.1.1
ncclient 0.6.9
netaddr 0.8.0
packaging 20.9
paramiko 2.7.2
pip 21.1.1
pycparser 2.20
PyNaCl 1.4.0
pyparsing 2.4.7
pyserial 3.5
PyYAML 5.4.1
scp 0.13.3
setuptools 39.2.0
six 1.16.0
transitions 0.8.8
wheel 0.36.2
xmltodict 0.12.0
yamlordereddictloader 0.4.0
zipp 3.4.1

[mersl]

I digged deeper in and it seems, the issue is close to here

if I check my facts gotten from the SRX Cluster, there is no element 'localre' included
the array is:
"current_re": [
"node0",
"master",
"fpc0",
"node",
"fwdd",
"member",
"pfem",
"re0",
"fpc0.pic0"
]

out of curiosity, I changed the localre to master and installation starts (kwargs: {"no-sync" : true ) is set as it's a ICU style ISSU on my 345s)

is there an issue in facts which causing a missing 'localre'? file show /etc/hosts.junos | match localre gives the expected result btw

[chidanandpujar]

Hi @mersl @prime001 @stuartianbrown
Thanks ,
Could you please verify the issue with latest PyEZ and share the netconf trace logs .
https://www.juniper.net/documentation/en_US/junos/topics/example/netconf-traceoptions-configuring.html

Thanks

[astritm]

Encountering exactly the same issue when using juniper.device.software collection, to upgrade a cluster of SRX300 boxes, with ISSU and setting "no-sync": true.

Have there been any solutions or workarounds identified?

Alternatively, I managed to perform the upgrade using the following task. However, it requires some adjustment as the RPC never receives a response, leading to eventual timeout. To address this, consider using ignore_errors: true in conjunction with a default timeout.

- name: "Upgrade JunOS {{ osversion }} on {{ inventory_hostname }}"
  juniper.device.rpc:
    rpcs:
      - request-package-in-service-upgrade
    kwargs:
      no-sync: true
      package-name: "/var/tmp/{{ ospackage }}"
    logfile: "{{ logdir }}/{{ inventory_hostname }}-upgrade.txt"
    port: "{{ ncf_port }}"
    timeout: "{{ wait }}"
  ignore_errors: true
  register: upgrade_response

My next objective is to retrieve the upgrade status for both nodes and present it to the user.
Thinking of getting the status from netconf traceoptions (via ansible) if nothing else works.