The Launchpad crash is the symptom, but it is the exact same bug. The
right thing to do is update the summary and description to reflect this;
I'll do that now.
** Summary changed:
- Updating FTI on one of our bugs causes PostgreSQL to SEGV
+ Able to trigger backend segmentation faults via tsearch2
** Description changed:
- Not only do our bugs cause syntax errors in FTI triggers (see bug
- 141627) they also cause postgresql to SEGV. When issuing an update on
- all bugs (which triggers FTI triggers to run) as we are doing for
- today's rollout, at least one bug with an ID between 50,000-100,000 (yay
- for binary search) causes a cruel crash in the backend. It is
- reproducible -- James T. triggered it three times today.
-
- NOTICE: A word you are indexing is too long. It will be ignored.
- CONTEXT: SQL statement "SELECT setweight(to_tsvector('default', coalesce($1,
'')), $2)||setweight(to_tsvector('default', coalesce($3, '')),
$4)||setweight(to_tsvector('default', coalesce($5, '')), $6) AS fti"
-
- We need a script which runs through the bugs and tries to update them;
- running this on carbon will tell us the who the culprits are. Once that
- is done we can finish issuing the SQL required to add the data that is
- required to finally fix bug 50663; we have had to roll out without it.
+ An update issued to the Launchpad database uncovered a bug in PostgreSQL
+ related to FTI; we were able to crash the backend when issuing a mass-
+ update. We have established a simple testcase that reproduces the
+ problem.
--
Able to trigger backend segmentation faults via tsearch2
https://bugs.launchpad.net/bugs/141637
You received this bug notification because you are a member of Ubuntu
Bugs, which is the bug contact for Ubuntu.
--
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
