Fix lb healthmonitor delay/timeout unit

This patch changes description of lb healthmonitor
delay/timeout unit from second to millisecond. The
actual unit is millisecond in code implementation
of Neutron-lbaas but existing API reference erroneously
describe it as second. Neutron team has decided to
correct the document rather then fix code implementation
to make them consistent. More discussion can be found
here:

https://review.openstack.org/#/c/339260/1

Change-Id: I43bbdf45a7f79f4a52dca0009fa195e9cc03d667
This commit is contained in:
yanyanhu 2016-07-11 22:47:19 -04:00 committed by Yanyan Hu
parent 0f0c104bce
commit 920d41a603
2 changed files with 8 additions and 8 deletions

View File

@ -45,12 +45,12 @@ properties:
# can be PING, TCP, HTTP, or HTTPS.
type: 'PING'
# The amount of time, in seconds, between sending probes to members.
delay: 10
# The amount of time, in milliseconds, between sending probes to members.
delay: 10000
# The maximum time in seconds that a monitor waits to connect before it
# times out. This value must be less than the delay value.
timeout: 5
# The maximum time in milliseconds that a monitor waits to connect before
# it times out. This value must be less than the delay value.
timeout: 5000
# The number of allowed connection failures before changing the status
# of the member to INACTIVE. A valid value is from 1 to 10.

View File

@ -222,13 +222,13 @@ class LoadBalancingPolicy(base.Policy):
default=PING,
),
HM_DELAY: schema.Integer(
_('The amount of time in seconds between sending '
_('The amount of time in milliseconds between sending '
'probes to members.'),
default=10,
),
HM_TIMEOUT: schema.Integer(
_('The maximum time in seconds that a monitor waits to '
'connect before it times out.'),
_('The maximum time in milliseconds that a monitor waits '
'to connect before it times out.'),
default=5,
),
HM_MAX_RETRIES: schema.Integer(