2009-12-03 03:24:43 |
Robert Collins |
bug |
|
|
added bug |
2009-12-03 03:30:09 |
Robert Collins |
security vulnerability |
yes |
no |
|
2009-12-03 23:47:51 |
Francis J. Lacoste |
affects |
canonical-identity-provider |
launchpad-foundations |
|
2009-12-03 23:48:05 |
Francis J. Lacoste |
visibility |
private |
public |
|
2009-12-03 23:48:37 |
Francis J. Lacoste |
description |
We're seeing operational issues in the sso servers when they are brought up in r/o mode.
I'll attach more backtraces to the bug, but briefly:
encodings.utf_8 has its global variable codecs set to None, which is something the VM does when it thinks a module is being finalised, to make things cleanup safely and racefree.
This patch http://bugs.python.org/file15318/modcleanup.patch in http://bugs.python.org/issue6551 shows a similar issue that Python itself had in its test suite with near-identical symptoms.
My theory is:
14:11 < lifeless> I'm thinking a thread that starts up and does some naughty 'probe'
14:12 < lifeless> and thus sometimes it wins, and the system breaks
14:12 < lifeless> and other times it loses and the system doesn't break
Which sounded plausible to spiv, so I think bears investigating. |
We're seeing operational issues in the app servers when they are brought up in r/o mode.
I'll attach more backtraces to the bug, but briefly:
encodings.utf_8 has its global variable codecs set to None, which is something the VM does when it thinks a module is being finalised, to make things cleanup safely and racefree.
This patch http://bugs.python.org/file15318/modcleanup.patch in http://bugs.python.org/issue6551 shows a similar issue that Python itself had in its test suite with near-identical symptoms.
My theory is:
14:11 < lifeless> I'm thinking a thread that starts up and does some naughty 'probe'
14:12 < lifeless> and thus sometimes it wins, and the system breaks
14:12 < lifeless> and other times it loses and the system doesn't break
Which sounded plausible to spiv, so I think bears investigating.
|
|
2009-12-03 23:49:50 |
Francis J. Lacoste |
description |
We're seeing operational issues in the app servers when they are brought up in r/o mode.
I'll attach more backtraces to the bug, but briefly:
encodings.utf_8 has its global variable codecs set to None, which is something the VM does when it thinks a module is being finalised, to make things cleanup safely and racefree.
This patch http://bugs.python.org/file15318/modcleanup.patch in http://bugs.python.org/issue6551 shows a similar issue that Python itself had in its test suite with near-identical symptoms.
My theory is:
14:11 < lifeless> I'm thinking a thread that starts up and does some naughty 'probe'
14:12 < lifeless> and thus sometimes it wins, and the system breaks
14:12 < lifeless> and other times it loses and the system doesn't break
Which sounded plausible to spiv, so I think bears investigating.
|
We're seeing operational issues in the app servers when they are brought up in r/o mode.
I'll attach more backtraces to the bug, but briefly:
encodings.utf_8 has its global variable codecs set to None, which is something the VM does when it thinks a module is being finalised, to make things cleanup safely and racefree.
This patch http://bugs.python.org/file15318/modcleanup.patch in http://bugs.python.org/issue6551 shows a similar issue that Python itself had in its test suite with near-identical symptoms.
My theory is:
14:11 < lifeless> I'm thinking a thread that starts up and does some naughty 'probe'
14:12 < lifeless> and thus sometimes it wins, and the system breaks
14:12 < lifeless> and other times it loses and the system doesn't break
Which sounded plausible to spiv, so I think bears investigating.
It happened with several (but not all app servers) and restarting them manually usually work.
|
|
2009-12-04 00:01:43 |
Robert Collins |
launchpad-foundations: status |
New |
Triaged |
|
2009-12-04 00:01:52 |
Robert Collins |
launchpad-foundations: importance |
Undecided |
Critical |
|
2009-12-04 15:00:31 |
Francis J. Lacoste |
launchpad-foundations: importance |
Critical |
High |
|
2009-12-04 18:49:10 |
Gary Poster |
launchpad-foundations: assignee |
|
Gary Poster (gary) |
|
2009-12-15 16:09:23 |
Gary Poster |
launchpad-foundations: status |
Triaged |
In Progress |
|
2010-02-25 17:01:17 |
Gary Poster |
branch linked |
|
lp:~gary/launchpad/bug491705 |
|
2010-02-25 18:03:44 |
Gary Poster |
launchpad-foundations: milestone |
|
10.03 |
|
2010-03-30 18:04:56 |
Gary Poster |
launchpad-foundations: milestone |
10.03 |
10.04 |
|
2010-04-07 15:00:39 |
Gary Poster |
launchpad-foundations: status |
In Progress |
Fix Committed |
|
2010-04-09 11:20:39 |
Ursula Junque |
tags |
|
qa-needstesting |
|
2010-05-04 11:03:08 |
Gary Poster |
launchpad-foundations: status |
Fix Committed |
Fix Released |
|
2010-05-04 11:46:16 |
Gary Poster |
tags |
qa-needstesting |
qa-ok |
|