Opened 7 years ago

Closed 7 years ago

#16745 closed defect (fixed)

Random OperationalError: disk I/O error

Reported by: vbraun Owned by:
Priority: major Milestone: sage-6.4
Component: packages: standard Keywords: random_fail ipython
Cc: roed Merged in:
Authors: Volker Braun Reviewers: William Stein
Report Upstream: N/A Work issues:
Branch: f9cbe65 (Commits, GitHub, GitLab) Commit: f9cbe655c1efb81c898a8f613cd16bf3bf7e3047
Dependencies: #16921 Stopgaps:

Status badges

Description (last modified by jdemeyer)

I see this occasionally on different buildbots, so it is not a hardware problem. Looks like a race with the IPython sqlite history database. 

sage: sig_on_count() ## line 923 ##
0
sage: ascii_art(tuple(DyckWords(3))) ## line 933 ##
(                                   /\   )
(            /\    /\      /\/\    /  \  )
( /\/\/\, /\/  \, /  \/\, /    \, /    \ )
sage: sig_on_count() ## line 937 ##
0

**********************************************************************
Traceback (most recent call last):
  File "/Users/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/sage/doctest/forker.py", line 2117, in __call__
    atexit._run_exitfuncs()
  File "/Users/buildslave-sage/slave/sage_git/build/local/lib/python/atexit.py", line 24, in _run_exitfuncs
    func(*targs, **kargs)
  File "/Users/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/IPython/core/interactiveshell.py", line 3189, in atexit_operations
    self.history_manager.end_session()
  File "/Users/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/IPython/core/history.py", line 525, in end_session
    len(self.input_hist_parsed)-1, self.session_number))
OperationalError: disk I/O error

See also #16372.

Change History (12)

comment:1 Changed 7 years ago by vbraun_spam

  • Milestone changed from sage-6.3 to sage-6.4

comment:2 Changed 7 years ago by jdemeyer

  • Component changed from doctest framework to packages: standard
  • Keywords ipython added

Not sure why you think this is related to the doctest framework... It looks like a bug in IPython.

comment:3 Changed 7 years ago by jdemeyer

  • Description modified (diff)

comment:4 Changed 7 years ago by vbraun

The doctest framework should (and does) disable the saving of commands to the history. Maybe we do something wrong / not enough to turn it off though. Could be an IPython issue as well, I agree.

comment:6 Changed 7 years ago by vbraun

  • Branch set to u/vbraun/close_ipython_history

comment:7 Changed 7 years ago by vbraun

  • Commit set to f9cbe655c1efb81c898a8f613cd16bf3bf7e3047
  • Dependencies set to 16921

New commits:

​10132bcUpdate to IPython 2.2.0
​f9cbe65Explicitly close IPython history database

comment:8 Changed 7 years ago by vbraun

  • Authors set to Volker Braun

comment:9 Changed 7 years ago by jdemeyer

  • Dependencies changed from 16921 to #16921

comment:10 Changed 7 years ago by vbraun

  • Status changed from new to needs_review

At least the patch doesn't make it worse ;-) only time will tell whether it fixes this. I'll reopen if it does not...

Last edited 7 years ago by vbraun (previous) (diff)

comment:11 Changed 7 years ago by was

  • Reviewers set to William Stein
  • Status changed from needs_review to positive_review

I'm rubber stamping this.

comment:12 Changed 7 years ago by vbraun

  • Branch changed from u/vbraun/close_ipython_history to f9cbe655c1efb81c898a8f613cd16bf3bf7e3047
  • Resolution set to fixed
  • Status changed from positive_review to closed
Note: See TracTickets for help on using tickets.