إيه اللي بيحصل فعلا جوه OpenStack لما تعمل server create؟ بالدليل | Deep Dive 0

الأمر ده كلنا كتبناه مية مرة: openstack server create. بتدوس Enter، وفي ثواني الـ VM بتبقى ACTIVE. بس إيه اللي حصل في الثواني دي بالظبط؟ في الحلقة دي بنمشي ورا request واحد جوه OpenStack خطوة بخطوة: Keystone و nova-api و nova-conductor و nova-scheduler و Placement و Neutron و Glance و Ceph — وكل خطوة بأمر وسطر log من السيستم نفسه. وفي الآخر بنكسر الـ cloud تلات مرات ونشوف كل مرة وقع فين. 🇬🇧 English subtitles available — turn on CC. We follow one openstack server create request through every OpenStack service — Keystone, nova-api, conductor, scheduler, Placement, Neutron, Glance and Ceph — with a command and a log line as proof at every step. Then we break the cloud 3 times and trace each failure. ⏱️ Chapters 00:00 What happens in these seconds? 00:55 Intro 01:26 The lab & one command 02:18 The players + 3 ways services talk 04:06 1 · Keystone: the token 04:55 2 · nova-api: 202 Accepted 08:18 3 · Scheduling: conductor → scheduler → Placement 15:24 4 · On the compute: queue, port, Ceph clone, vif-plugged 23:48 One XML file holds all of OpenStack 25:52 Where did the seconds go? 27:32 The full map 29:19 Breaking the cloud 3 times 31:32 Break 2: scheduler down 34:53 Break 3: No valid host 36:21 The 3-command method 37:12 Next: Keystone 🧪 The lab Kolla-Ansible 2025.1 · 3 controllers (HA) · 3 computes · 3 Ceph nodes · ML2/OVS + VXLAN 💻 Commands used in the episode openstack server create --flavor m1.tiny --image cirros-raw --network demo-net --wait sa3y-vm openstack token issue openstack --debug server create --flavor m1.tiny --image cirros-raw --network demo-net sa3y-vm-2 openstack server event list sa3y-vm-2 grep $REQUEST_ID /var/log/kolla/nova/nova-scheduler.log openstack resource provider allocation show $SERVER_ID openstack resource provider list docker exec rabbitmq rabbitmqctl list_queues name consumers messages openstack port list --server sa3y-vm-2 openstack port show $PORT_ID -c fixed_ips -c binding_host_id -c binding_vif_type -c status rbd ls vms rbd info vms/${SERVER_ID}_disk | grep parent grep $SERVER_ID /var/log/kolla/nova/nova-compute.log | grep network-vif-plugged docker exec nova_libvirt virsh dumpxml $INSTANCE_NAME openstack server event show sa3y-vm-2 $REQUEST_ID grep $SERVER_ID /var/log/kolla/nova/nova-compute.log | grep "Took" 💥 The 3 breaks 1) Wrong password → HTTP 401 (Keystone) 2) nova-scheduler stopped on all controllers → message stuck in the queue → rpc_response_timeout (60 s) → ERROR 3) Flavor with a 1 PB disk → No valid host was found (Placement) 💼 LinkedIn:   / ahmed-saey-eldin-224a28130   #OpenStack #CloudComputing #TelcoCloud #Ceph #DevOps