Project

General

Profile

Tasks #57709

Handle complex role names in teuthology tests

Added by Laura Flores 2 months ago. Updated 2 months ago.

Status:
New
Priority:
Normal
Assignee:
-
Category:
-
Target version:
-
% Done:

0%

Tags:
Reviewed:
Affected Versions:

Description

Currently, most real-world clusters have daemons with complex names, which teuthology and vstart environments do not reflect.

When we define roles in teuthology tests, like this:

roles:
- - mon.a
  - mon.c
  - mgr.y
  - mds.a
  - osd.0
  - osd.1
  - osd.2
  - osd.3
  - client.0
  - node-exporter.a
  - alertmanager.a
- - mon.b
  - mds.b
  - mgr.x
  - osd.4
  - osd.5
  - osd.6
  - osd.7
  - client.1
  - prometheus.a
  - grafana.a
  - node-exporter.b

It is currently not possible to make the daemon ids more complex, i.e.

roles:
- - mon.a.complex
  - mon.c.complex
  - mgr.y.complex
  - mds.a.complex
  - osd.0
  - osd.1
  - osd.2
  - osd.3
  - client.0
  - node-exporter.a
  - alertmanager.a
- - mon.b.complex
  - mds.b.complex
  - mgr.x.complex
  - osd.4
  - osd.5
  - osd.6
  - osd.7
  - client.1
  - prometheus.a
  - grafana.a
  - node-exporter.b

Having the ability to name complex roles could help catch bugs like https://tracker.ceph.com/issues/57700 before they surface on real clusters.

orig.config.yaml View (5.6 KB) Laura Flores, 09/29/2022 05:15 PM

teuthology.log View (402 KB) Laura Flores, 09/29/2022 05:15 PM


Related issues

Related to mgr - Bug #57700: mgr/telemetry: ValueError: too many values to unpack (expected 2) in get_mempool Resolved

History

#1 Updated by Laura Flores 2 months ago

  • Related to Bug #57700: mgr/telemetry: ValueError: too many values to unpack (expected 2) in get_mempool added

#2 Updated by Laura Flores 2 months ago

  • Project changed from Ceph to teuthology

#3 Updated by Laura Flores 2 months ago

Example of a run I tried when making the daemon names more complex:
http://pulpito.front.sepia.ceph.com/lflores-2022-09-29_00:18:10-upgrade:telemetry-upgrade:quincy-x-wip-telemetry-perf-fix-distro-default-smithi/

The test failed here:

2022-09-29T00:36:01.150 INFO:tasks.cephadm:Config: {'cephadm_branch': 'quincy', 'cephadm_git_url': 'https://github.com/ceph/ceph', 'conf': {'osd': {'osd_class_default_list': '*', 'osd_class_load_list': '*', 'debug ms': 1, 'debug osd': 20, 'osd shutdown pgref assert': True}, 'mgr': {'debug mgr': 20, 'debug ms': 1}, 'mon': {'debug mon': 20, 'debug ms': 1, 'debug paxos': 20}}, 'image': 'quay.io/ceph/daemon:latest-quincy-devel', 'create_rbd_pool': True, 'flavor': 'default', 'log-ignorelist': ['\\(MDS_ALL_DOWN\\)', '\\(MDS_UP_LESS_THAN_MAX\\)'], 'sha1': '97833a6a81fed7f868e1d544816cfbdf254fdb43'}
2022-09-29T00:36:01.151 INFO:tasks.cephadm:Cluster image is quay.io/ceph/daemon:latest-quincy-devel
2022-09-29T00:36:01.152 INFO:tasks.cephadm:Cluster fsid is b156fd74-3f8e-11ed-8431-001a4aab830c
2022-09-29T00:36:01.153 INFO:tasks.cephadm:Choosing monitor IPs and ports...
2022-09-29T00:36:01.153 ERROR:teuthology.contextutil:Saw exception from nested tasks
Traceback (most recent call last):
  File "/home/teuthworker/src/git.ceph.com_git_teuthology_35ea38a9840006713a3d42472a2c536a25e88c15/teuthology/contextutil.py", line 31, in nested
    vars.append(enter())
  File "/usr/lib/python3.6/contextlib.py", line 81, in __enter__
    return next(self.gen)
  File "/home/teuthworker/src/github.com_ljflores_ceph_6796dedf5c3e8cfdbdcbaf17ffca235d71639d73/qa/tasks/cephadm.py", line 1520, in initialize_config
    mon_bind_addrvec=config.get('mon_bind_addrvec', True),
  File "/home/teuthworker/src/github.com_ljflores_ceph_6796dedf5c3e8cfdbdcbaf17ffca235d71639d73/qa/tasks/ceph.py", line 508, in get_mons
    assert mons
AssertionError

#4 Updated by Laura Flores 2 months ago

Also available in: Atom PDF