Project

General

Profile

Actions

Bug #45147

closed

Module 'diskprediction_local' takes forever to load

Added by Kefu Chai about 4 years ago. Updated 5 months ago.

Status:
Resolved
Priority:
Normal
Assignee:
-
Category:
diskprediction_local
Target version:
-
% Done:

100%

Source:
Tags:
backport_processed
Backport:
octopus, pacific
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

debian sid, python3.8,2,

sha1: 5e540f1191975fac38733e765312295775a52ad4

ceph-mgr never finishes loading diskprediction_local:

2020-04-20T17:21:04.157+0800 7f2d9860ee40  1 mgr[py] Loading python module 'diskprediction_local'
2020-04-20T17:21:04.201+0800 7f2d9860ee40 10 mgr[py] Computed sys.path '/var/ssd/ceph/src/pybind:/var/ssd/ceph/build/lib/cython_modules/lib.3:/var/ssd/ceph/src/python-common::/usr/lib/python38.zip:/usr/l\
ib/python3.8:/usr/lib/python3.8/lib-dynload:/var/ssd/ceph/src/pybind/mgr:/usr/local/lib/python3.8/dist-packages:/usr/lib/python3/dist-packages:/usr/lib/python3.8/dist-packages'
2020-04-20T17:21:04.429+0800 7f2d7d7fa700 10 monclient: tick
2020-04-20T17:21:04.429+0800 7f2d7d7fa700 10 monclient: _check_auth_rotating renewing rotating keys (they expired before 2020-04-20T17:20:34.435106+0800)
2020-04-20T17:21:04.429+0800 7f2d7d7fa700 20 monclient: _un_backoff reopen_interval_multipler now 1
2020-04-20T17:21:05.433+0800 7f2d7d7fa700 10 monclient: tick
2020-04-20T17:21:05.433+0800 7f2d7d7fa700 10 monclient: _check_auth_rotating renewing rotating keys (they expired before 2020-04-20T17:20:35.435416+0800)
2020-04-20T17:21:06.433+0800 7f2d7d7fa700 10 monclient: tick
2020-04-20T17:21:06.433+0800 7f2d7d7fa700 10 monclient: _check_auth_rotating renewing rotating keys (they expired before 2020-04-20T17:20:36.435629+0800)
2020-04-20T17:21:07.433+0800 7f2d7d7fa700 10 monclient: tick
2020-04-20T17:21:07.433+0800 7f2d7d7fa700 10 monclient: _check_auth_rotating renewing rotating keys (they expired before 2020-04-20T17:20:37.435829+0800)
2020-04-20T17:21:08.429+0800 7f2d7d7fa700 10 monclient: tick
2020-04-20T17:21:08.429+0800 7f2d7d7fa700 10 monclient: _check_auth_rotating renewing rotating keys (they expired before 2020-04-20T17:20:38.435989+0800)
2020-04-20T17:21:09.429+0800 7f2d7d7fa700 10 monclient: tick

Related issues 6 (1 open5 closed)

Related to mgr - Bug #42764: Test failure: test_diskprediction_local (tasks.mgr.test_module_selftest.TestModuleSelftest)ResolvedSage Weil

Actions
Related to mgr - Bug #43447: mgr/diskprediction: diskprediction module fails to initialize with newer SciPy versionsNewSage Weil

Actions
Related to mgr - Bug #50196: TestModuleSelftest.test_diskprediction_localResolvedKefu Chai

Actions
Copied to mgr - Backport #50165: pacific: Module 'diskprediction_local' takes forever to loadRejectedActions
Copied to mgr - Backport #50166: nautilus: Module 'diskprediction_local' takes forever to loadRejectedActions
Copied to mgr - Backport #50167: octopus: Module 'diskprediction_local' takes forever to loadResolvedCory SnyderActions
Actions #1

Updated by Kefu Chai about 4 years ago

  • Related to Bug #42764: Test failure: test_diskprediction_local (tasks.mgr.test_module_selftest.TestModuleSelftest) added
Actions #2

Updated by Kefu Chai about 4 years ago

i think it's caused by numpy.

see the discussions at

- https://mail.python.org/pipermail/python-dev/2019-January/156095.html and
- https://bugs.python.org/issue10915

seems the short-term fix is to either

- disable diskprediction_local
- run diskprediction_local in a dedicated interpreter instead of in a sub-interpreter

Actions #3

Updated by Tomasz Torcz almost 4 years ago

FYI, the same happens on Fedora 33.

Actions #4

Updated by Kiefer Chang almost 4 years ago

  • Related to Bug #43447: mgr/diskprediction: diskprediction module fails to initialize with newer SciPy versions added
Actions #5

Updated by Kefu Chai almost 4 years ago

Jeff Layton also ran into this issue on fedora 32 with python3.8.

Actions #6

Updated by Josh Durgin almost 4 years ago

  • Priority changed from Normal to Urgent
Actions #7

Updated by Kefu Chai almost 4 years ago

  • Status changed from New to Fix Under Review
  • Assignee set to Kefu Chai
  • Pull request ID set to 34846

as this issue only impacts python3.8, and neither el8 nor ubuntu bionic ships python3.8. i am not adding any LTS branches to "backport" field of this ticket.

Actions #8

Updated by Kefu Chai almost 4 years ago

  • Status changed from Fix Under Review to New

the workaround fix was merged. let's keep this issue open until we have a proper fix.

Actions #9

Updated by Kefu Chai almost 4 years ago

  • Assignee deleted (Kefu Chai)
  • Priority changed from Urgent to High
Actions #10

Updated by Kefu Chai almost 4 years ago

  • Priority changed from High to Urgent

https://github.com/ceph/ceph/pull/34894 fails the tests on ubuntu focal.

we should address the test failures on ubuntu/focal

Actions #12

Updated by Kefu Chai about 3 years ago

  • Status changed from New to Pending Backport
  • Backport set to nautilus, octopus, pacific
Actions #13

Updated by Backport Bot about 3 years ago

  • Copied to Backport #50165: pacific: Module 'diskprediction_local' takes forever to load added
Actions #14

Updated by Backport Bot about 3 years ago

  • Copied to Backport #50166: nautilus: Module 'diskprediction_local' takes forever to load added
Actions #15

Updated by Backport Bot about 3 years ago

  • Copied to Backport #50167: octopus: Module 'diskprediction_local' takes forever to load added
Actions #16

Updated by Kefu Chai about 3 years ago

  • Related to Bug #50196: TestModuleSelftest.test_diskprediction_local added
Actions #17

Updated by Konstantin Shalygin about 2 years ago

  • Backport changed from nautilus, octopus, pacific to octopus, pacific
Actions #18

Updated by Backport Bot over 1 year ago

  • Tags set to backport_processed
Actions #19

Updated by Konstantin Shalygin 5 months ago

  • Status changed from Pending Backport to Resolved
  • Priority changed from Urgent to Normal
  • % Done changed from 0 to 100
Actions

Also available in: Atom PDF