snmp can't be polled

[Alexander Chadin]

I have Ceilometer (stable/Pike) with the following polling.yaml

---
sources:
    - name: some_pollsters
      interval: 60
      meters:
        - cpu
        - memory.usage
        - network.incoming.bytes
        - network.incoming.packets
        - network.outgoing.bytes
        - network.outgoing.packets
        - disk.read.bytes
        - disk.read.requests
        - disk.write.bytes
        - disk.write.requests
    - name: meter_snmp
      interval: 60
      resources:
          - snmp://10.5.250.21
      meters:
       - "hardware.cpu*"
       - "hardware.memory*"
       - "hardware.disk*"
       - "hardware.network*"

Telemetry node can get stream of data using "snmpwalk -v 2c -c public 10.5.250.21"

But I get the following log on central.log:

2017-11-14 13:11:51.718 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.memory.total, no resources found this cycle
2017-11-14 13:11:51.718 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.disk.size.used, no resources found this cycle
2017-11-14 13:11:51.719 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.memory.cached, no resources found this cycle
2017-11-14 13:11:51.719 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.cpu.load.5min, no resources found this cycle
2017-11-14 13:11:51.719 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.memory.swap.total, no resources found this cycle
2017-11-14 13:11:51.719 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.cpu.load.1min, no resources found this cycle
2017-11-14 13:11:51.720 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.cpu.load.15min, no resources found this cycle
2017-11-14 13:11:51.720 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.network.ip.incoming.datagrams, no resources found this cycle
2017-11-14 13:11:51.720 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.memory.used, no resources found this cycle
2017-11-14 13:11:51.721 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.memory.swap.avail, no resources found this cycle
2017-11-14 13:11:51.721 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.memory.buffer, no resources found this cycle
2017-11-14 13:11:51.721 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.network.incoming.bytes, no resources found this cycle
2017-11-14 13:11:51.721 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.disk.size.total, no resources found this cycle
2017-11-14 13:11:51.722 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.network.outgoing.bytes, no resources found this cycle
2017-11-14 13:11:51.722 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.cpu.util, no resources found this cycle
2017-11-14 13:11:51.722 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.network.ip.outgoing.datagrams, no resources found this cycle
2017-11-14 13:11:51.722 15544 INFO ceilometer.agent.manager [-] Skip pollster hardware.network.outgoing.errors, no resources found this cycle

I've made some investigation and found out that candidate_res variable (https://github.com/openstack/ceilometer/blob/stable/pike/ceilometer/agent/manager.py#L160) has [] value.

[gordon chung (chungg)]

i don't have time currently to debug but as a shot in the dark. what happens if you align snmp://10.5.250.21 the same as the others. i know yaml is very specific about it's formatting.

---
sources:
    - name: some_pollsters
      interval: 60
      meters:
        - cpu
        - memory.usage
        - network.incoming.bytes
        - network.incoming.packets
        - network.outgoing.bytes
        - network.outgoing.packets
        - disk.read.bytes
        - disk.read.requests
        - disk.write.bytes
        - disk.write.requests
    - name: meter_snmp
      interval: 60
      resources:
        - snmp://10.5.250.21
      meters:
        - "hardware.cpu*"
        - "hardware.memory*"
        - "hardware.disk*"
        - "hardware.network*"

[Alexander Chadin (joker946)]

Unfortunately, it didn't fix the issue

[Ivan Denezhkin (pikwick)]

Setting coordination.backend_url in ceilometer.conf fixes the issue. Reason is that method get in Resources class add static resources after check < if self.agent_manager.partition_coordinator: > but partition_coordinator without setting coordination.backend_url is None

Is setting of coordination.backend_url necessery by design?

[gordon chung (chungg)]

it should not be. if this is the case, it is a bug.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/527108

[gordon chung (chungg)]

hopefully, the proposed patch fixes the issue.

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/527108
Committed: https://git.openstack.org/cgit/openstack/ceilometer/commit/?id=46c84ef7f7db3500e8864501bb517aca436003cb
Submitter: Zuul
Branch: master

commit 46c84ef7f7db3500e8864501bb517aca436003cb
Author: gord chung <email address hidden>
Date: Mon Dec 11 14:42:45 2017 +0000

    static resources not picked up

    polling is still broken if backend_url is not set. switch to
    tooz hashrings caused it to be ignored.

    Change-Id: Iace637dd5ed8a852db79b23e99918ca3e5c6605e
    Closes-Bug: #1732147

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/pike
Review: https://review.openstack.org/527521

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/527521
Committed: https://git.openstack.org/cgit/openstack/ceilometer/commit/?id=50bcd23528d0492b167f2a034cdb13312f309b55
Submitter: Zuul
Branch: stable/pike

commit 50bcd23528d0492b167f2a034cdb13312f309b55
Author: gord chung <email address hidden>
Date: Mon Dec 11 14:42:45 2017 +0000

    static resources not picked up

    polling is still broken if backend_url is not set. switch to
    tooz hashrings caused it to be ignored.

    Change-Id: Iace637dd5ed8a852db79b23e99918ca3e5c6605e
    Closes-Bug: #1732147
    (cherry picked from commit 46c84ef7f7db3500e8864501bb517aca436003cb)

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/ceilometer 9.0.4 release.

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/ceilometer 10.0.0 release.