)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":11628,"name":"Michael Johnson","email":"johnsomor@gmail.com","username":"johnsom"},"change_message_id":"366d1714ce9040b0d8608fbdf45cba84b837f957","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"41137c93_d7953292","updated":"2025-08-26 16:13:34.000000000","message":"Can you explain more of why you need this?\n\nIf the member server is reachable via routing from the VIP subnet, you do not need to specify the subnet_id option when creating a member object on the pool. If subnet_id is not specified, no subnet will be plugged into the amphora instances.","commit_id":"f109d341a9ddc1a6f216fe5c88da3a4592c3881b"},{"author":{"_account_id":31671,"name":"manu","email":"manumoh@cisco.com","username":"manuvakery"},"change_message_id":"2867f90542733153cf8c522041b7acfe0915d81a","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"4711e989_93a232e5","in_reply_to":"41137c93_d7953292","updated":"2025-08-26 17:08:46.000000000","message":"This is not something a client without openstack expertise will know or will follow . Also when you create an LB from horizon it automatically populate the subnet-id.  We have hundreds of such LBs in production( RHOSP 16.1) where backend members falls into full subnets. During the  OSP17 migration existing lb failover will not work as there is no ip left in those subnets. Even a new LB created  with a member from full subnet will fail.","commit_id":"f109d341a9ddc1a6f216fe5c88da3a4592c3881b"},{"author":{"_account_id":11628,"name":"Michael Johnson","email":"johnsomor@gmail.com","username":"johnsom"},"change_message_id":"7b9a3049115becd7376da3d31ecffd45ae8816ad","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"88406938_be55f008","updated":"2025-08-26 18:20:25.000000000","message":"I don\u0027t think this is a good idea. It is perfectly valid to plug a member into provider networks, so this change becomes confusing.\nAlso, the listed use case is working as expected/designed. If you specify a subnet-id on a member, it will need to be able to allocate an IP on that subnet to function correctly.","commit_id":"62ea1a0e7c446f9250418bd57a245c88c5f587dc"},{"author":{"_account_id":31671,"name":"manu","email":"manumoh@cisco.com","username":"manuvakery"},"change_message_id":"89c967b46091939dbbe4a6799440b2f0d88ead18","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"d155234c_51249082","in_reply_to":"88406938_be55f008","updated":"2025-09-01 10:12:32.000000000","message":"I agree to your point. But its a problem for an operator where majority of provider networks consists of multiple subnets (say 5-6). One day customer can create lb and add members without any issues and another day the same LB cannot be failover due to ip not available. The main problem is we can\u0027t enforce customer not to provide the subnet-id  while adding the member object. Its kind of wasting a lot of ips too.","commit_id":"62ea1a0e7c446f9250418bd57a245c88c5f587dc"}]}