Closed dlarson04 closed 3 years ago
I think this happens for the same reason when the model is in use by many agents it is deleted
root@plasmas1:~/hzn-work/mms/policy# hzn -v mms object delete --type application-mmsmodel --id purdue-hello-mms-config.json
[verbose] Reading configuration file: /etc/horizon/hzn.json
[verbose] Reading configuration file: /etc/default/horizon
[verbose] Config file does not exist: /root/.hzn/hzn.json.
[verbose] No project level configuration file found.
[verbose] The model management service url: https://cp-console.ieam-roks-scale-70ea81cdef68a2eb78ece6d890b7dad3-0000.us-south.containers.appdomain.cloud/edge-css
[verbose] DELETE https://cp-console.ieam-roks-scale-70ea81cdef68a2eb78ece6d890b7dad3-0000.us-south.containers.appdomain.cloud/edge-css/api/v1/objects/ieam-roks-scale/application-mmsmodel/purdue-hello-mms-config.json
[verbose] HTTP request timeout set to 30 seconds
[verbose] Encountered HTTP error: Delete "https://cp-console.ieam-roks-scale-70ea81cdef68a2eb78ece6d890b7dad3-0000.us-south.containers.appdomain.cloud/edge-css/api/v1/objects/ieam-roks-scale/application-mmsmodel/purdue-hello-mms-config.json": net/http: timeout awaiting response headers calling Model Management Service REST API DELETE https://cp-console.ieam-roks-scale-70ea81cdef68a2eb78ece6d890b7dad3-0000.us-south.containers.appdomain.cloud/edge-css/api/v1/objects/ieam-roks-scale/application-mmsmodel/purdue-hello-mms-config.json. HTTP status: . Will retry.
[verbose] HTTP code: 204
Object purdue-hello-mms-config.json deleted from org ieam-roks-scale in the Model Management Service
The reason seems to be that CSS receives the model update in plenty of time but then on the same thread handling the HTTP request, it does some type of notification on all 2994 agents which takes 4 minutes causing the timeout on the HTTP connection.