nova cell v2

在批量创建75台虚拟机机器后，在随后的一段时间内，nova list 接口会卡住

debug方式： tailf -f /var/log/kolla/nova/nova-api.log 三台ccontrol

GET /v2.1/d33c6c311f9a4389aecc3ba52a0e5db1/servers/detail?all_tenants=1 HTTP/1.1" status: 200

可以看到请求已经接收到

但是随后看到了两个timeout

Cell 00000000-0000-0000-0000-000000000000 is not responding and hence skipped from the results.

另一个节点就会返回eventlet的reset 链接的traceback

ts=1 HTTP/1.1" status: 200 len: 119364 time: 60.6210210

2020-09-07 15:26:23.772 30 INFO nova.osapi_compute.wsgi.server [req-47904eb4-f008-4d1e-a1cf-e8be775b44d6 7b6a8a4f5d4e4815949607b865b38ba8 d33c6c311f9a4389aecc3ba52a0e5db1 - default default] Traceback (most recent call last):
File "/usr/lib/python2.7/dist-packages/eventlet/wsgi.py", line 515, in handle_one_response
...
File "/usr/lib/python2.7/dist-packages/eventlet/greenio/base.py", line 371, in _send_loop
return send_method(data, *args)
error: [Errno 104] Connection reset by peer

cellv2.png

https://xcodest.me/nova-cell-v2.html

2018 openstack峰会演示的场景，基于73个cell管理了9000+物理机器
https://object-storage-ca-ymq-1.vexxhost.net/swift/v1/6e4619c416ff4bd19e1c087f27a43eea/www-assets-prod/presentation-media/2018-OpenStack-Summit-Berlin-Scaling-Nova-with-CellsV2.pdf