Hello,
Using the Bigip_ucs_fetch module to pull a UCS backup. I keep getting not found, since in the source code of the module its appending /mgmt/tm/sys/ucs to the url path: https://IP:Port/mgmt/tm/sys/ucs.
What is your f5 configuration doesn’t use that path? Has anyone found a workaround?
Using F5 version 16.1.3.
Regards,
Kyle
Hi,
I don’t know much about this module, or even F5 BigIP product API, but this path seems to be the default REST API uri path for ucs : https://clouddocs.f5.com/api/icontrol-rest/APIRef_tm_sys_ucs.html.
This path is indeed hardcoded in bigip_ucs_fetch module ($ansibleCollectionsInstallPath/f5networks/f5_modules/plugins/modules/bigip_ucs_fetch.py) :
def read_current_from_device(self):
uri = “https://{0}:{1}/mgmt/tm/sys/ucs”.format(
self.client.provider[‘server’],
self.client.provider[‘server_port’],
)
resp = self.client.api.get(uri)
…
This is for V2Manager class, whatever this means (seems to be in use for TMOS >=12.1.0). V1Manager class (so <12.1.0) uses this one : https://{0}:{1}/mgmt/tm/util/bash.
A version check happen at the start of the script :
def is_version_v1(self):
“”"Checks to see if the TMOS version is less than 12.1.0
Versions prior to 12.1.0 have a bug which prevents the REST
API from properly listing any UCS files when you query the
/mgmt/tm/sys/ucs endpoint. Therefore you need to do everything
through tmsh over REST.
:return: bool
“”"
version = tmos_version(self.client)
if Version(version) < Version(‘12.1.0’):
return True
else:
return False
Which seems to indicate you could still query the REST API over tmsh. A quick Google search indicate that it uses /mgmt/tm/util/bash as path, therefore corresponding the V1Manager class referenced in ansible module.
I don’t know why you can’t seem to be able to reach the default ucs API endpoint path on your instances, if that’s something expected or if F5 changed something on more recent version, but you can perhaps still manually use the v1 path with ansible.builtin.uri module or shell / command to achieve what you’re trying to do. You could also manually change path in module file and maintain a fork module of your own.
I’d start to look into changelog for a potential API path change or something like that. Maybe try to query this endpoint manually through curl, and some others so you might have a better idea on what is happening. The error message you got might not be accurate, or it might mask an underlying one.
Hope this helps !