Avoid poll_interval to be the same as HTTP Keep-Alive timeout

When OpenStack API services are behind Apache2, HTTP Keep-Alive timeout
is 5 seconds out of the box. Polling operations with the same interval
as HTTP Keep-Alive timeout can cause intermittent failures like:

Unable to establish connection to
https://nova.example.com:8774/v2.1/servers/UUID:
('Connection aborted.', RemoteDisconnected('Remote end closed connection
without response',))

Let's avoid the same value by default as Apache2's HTTP Keep-Alive
timeout.

Closes-Bug: #1946912
Change-Id: Ibda414b129a44d38ed3c3a4b5a43fd45e63ec122
This commit is contained in:
Nobuto Murata 2021-10-13 14:30:29 +09:00 committed by Andrey Kurilin
parent 20590f3071
commit 0ec255aad2
1 changed files with 1 additions and 1 deletions

View File

@ -238,7 +238,7 @@ OPTS = {"openstack": [
deprecated_group="benchmark",
help="Server resize timeout"),
cfg.FloatOpt("nova_server_resize_poll_interval",
default=5.0,
default=4.0,
deprecated_group="benchmark",
help="Server resize poll interval"),
# "resize_confirm": (0, 200, 2)