Issue 38 in dataparksearch: double free or corruption backtrace

4 views
Skip to first unread message

datapar...@googlecode.com

unread,
Feb 13, 2012, 2:39:06 PM2/13/12
to datapar...@googlegroups.com
Status: New
Owner: ----
Labels: Type-Defect Priority-Medium

New issue 38 by codebl...@elrod.me: double free or corruption backtrace
http://code.google.com/p/dataparksearch/issues/detail?id=38

We are exploring search options for use within the Fedora Project, and in
our tests of running the crawler (indexer), we are running into an
occasional backtrace. It seems to be almost at random, I don't notice
anything in common about the URL it hits before the backtrace each time.

I've attached the backtrace, and happy to provide any more information we
can to help get this solved.

What version of the product are you using? On what operating system?

indexer from dpsearch-4.53-pqsql

[root@junk09 dpsearch]# cat /etc/redhat-release
Red Hat Enterprise Linux Server release 6.2 (Santiago)

[root@junk09 dpsearch]# uname -a
Linux junk09.qa.fedoraproject.org 2.6.32-220.4.1.el6.x86_64 #1 SMP Thu Jan
19 14:50:54 EST 2012 x86_64 x86_64 x86_64 GNU/Linux

Please provide any additional information below.

Again, if there's anything we can do, we're happy to work with you try to
and get this fixed. :)

Attachments:
bt 7.1 KB

datapar...@googlecode.com

unread,
Feb 14, 2012, 5:46:54 AM2/14/12
to datapar...@googlegroups.com
Updates:
Owner: dp.max...@gmail.com

Comment #1 on issue 38 by dp.max...@gmail.com: double free or corruption
backtrace
http://code.google.com/p/dataparksearch/issues/detail?id=38

Perhaps you've faced a bug when libcrypto had been used in multithread
indexer without safety locks. This issue has been fixed in snapshot of 4.54
version. Please try it:
http://dataparksearch.googlecode.com/files/dpsearch-4.54-2012-02-14.tar.bz2

datapar...@googlecode.com

unread,
Feb 14, 2012, 7:57:15 PM2/14/12
to datapar...@googlegroups.com

Comment #2 on issue 38 by codebl...@elrod.me: double free or corruption
backtrace
http://code.google.com/p/dataparksearch/issues/detail?id=38

Hi, I tried the snapshot you linked to and the indexer is still randomly
failing in the same way. Is there any other information I can provide that
would be helpful in getting this fixed? We really like the looks of
Datapark so far, aside from this issue. :-).

Thanks!


datapar...@googlecode.com

unread,
Feb 15, 2012, 4:41:20 AM2/15/12
to datapar...@googlegroups.com

Comment #3 on issue 38 by dp.max...@gmail.com: double free or corruption
backtrace
http://code.google.com/p/dataparksearch/issues/detail?id=38

How many indexing threads do you run simultaneously (the value for -N
switch)?

datapar...@googlecode.com

unread,
Feb 15, 2012, 7:12:11 AM2/15/12
to datapar...@googlegroups.com
Updates:
Status: Fixed

Comment #4 on issue 38 by dp.max...@gmail.com: double free or corruption
backtrace
http://code.google.com/p/dataparksearch/issues/detail?id=38

Please try fresh snapshot:
http://dataparksearch.googlecode.com/files/dpsearch-4.54-2012-02-15.tar.bz2

I've moved SSL initialisation to the right place.

Reply all
Reply to author
Forward
0 new messages