No check that healthmonitor delay should >= timeout
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
octavia |
Invalid
|
Low
|
Unassigned |
Bug Description
issue is in kilo branch
healthmonitor delay is 10, timeout is 12
it does not make sense
[root@opencos2 ~(keystone_admin)]# neutron lbaas-healthmon
+------
| Field | Value |
+------
| admin_state_up | True |
| delay | 10 |
| expected_codes | 305,205 |
| http_method | GET |
| id | 6d29f448-
| max_retries | 10 |
| pools | {"id": "591be59b-
| tenant_id | be58eaec789d44f
| timeout | 12 |
| type | PING |
| url_path | /api/ |
+------
Changed in neutron: | |
assignee: | nobody → dongjuan (dong-juan1) |
Changed in neutron: | |
importance: | Undecided → Low |
affects: | neutron → octavia |
A health check sent every 10 seconds with a 12 second delay is possible, although not all drivers may support overlapping health checks like this. I do agree there should be an administratively configureable lower bound on health check intervals (delay) to prevent user using the LBaaS to preform a DoS. This becomes more important with https:/ /bugs.launchpad .net/neutron/ +bug/1600326, since delay values are interpreted in milliseconds.