Bug 468971 - autofs segfaults when using LDAP mount maps
Summary: autofs segfaults when using LDAP mount maps
Status: RESOLVED FIXED
: 551722 (view as bug list)
Alias: None
Product: openSUSE 11.1
Classification: openSUSE
Component: Basesystem (show other bugs)
Version: Final
Hardware: x86-64 openSUSE 11.1
: P3 - Medium : Critical with 3 votes (vote)
Target Milestone: ---
Assignee: Leonardo Chiquitto
QA Contact: E-mail List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2009-01-23 19:11 UTC by Ian Grant
Modified: 2010-05-24 14:38 UTC (History)
5 users (show)

See Also:
Found By: ---
Services Priority:
Business Priority:
Blocker: ---
Marketing QA Status: ---
IT Deployment: ---


Attachments
Core dumped by automount (1.50 MB, application/octet-stream)
2010-05-21 15:05 UTC, Sven Bilke
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Ian Grant 2009-01-23 19:11:13 UTC
User-Agent:       Mozilla/5.0 (Macintosh; U; PPC Mac OS X 10.5; en-US; rv:1.9.0.5) Gecko/2008120121 Firefox/3.0.5

We have automount maps served by LDAP. They have worked fine with all previous opensuse versions including 11.0. Now opensuse 11.1 automount process segfaults.

Reproducible: Always

Steps to Reproduce:
1. Configure automount maps (sorry, I can't describe this, it's too complicated.)
2. Start autofs with 'rcautofs start'
3. cd to an automounted directory
Actual Results:  
Jan 23 17:45:15 tywi kernel: automount[3917]: segfault at 7f76873f62e0 ip 00007f76873f62e0 sp 00007f76839ef0f8 error 14<6>automount[3916]:
 segfault at 7f76873f62e0 ip 00007f76873f62e0 sp 00007f7683af00f8 error 14 in libc-2.9.so[7f7687f39000+14f000] in libc-2.9.so[7f7687f39000
+14f000]
Jan 23 17:45:15 tywi kernel: 
Jan 23 17:45:15 tywi kernel: automount[3914]: segfault at 7f76873f62e0 ip 00007f76873f62e0 sp 00007f768656c0f8 error 14 in libc-2.9.so[7f7
687f39000+14f000]


Expected Results:  
I expected it to work!
Comment 1 Andreas Jaeger 2009-01-29 08:18:29 UTC
So, it crashes in glibc.  Petr, any ideas?
Comment 2 Petr Baudis 2009-01-29 09:37:22 UTC
This can be bug in glibc or some memory corruption, impossible to say. Ian, can you please try to gather a coredump?
Comment 3 Matthias Koenig 2009-01-29 10:02:23 UTC
Which version are you running?
Have you installed the latest autofs update? I am asking because this update fixes a segfault problem (Bug 458336).
Comment 4 Ian Grant 2009-01-29 17:54:28 UTC
The RPM is autofs-5.0.3-82.28.1

I have a core dump but it's over 10MB in size. You can get it from http://www.cl.cam.ac.uk/~ig206/automount.core

This is an x86_64 machine and the associated log message with this core dump was

Jan 29 15:34:24 vignemale kernel: automount[4182]: segfault at 7fd5962f62e0 ip 00007fd5962f62e0 sp 00007fd592af10f8 error 14 in libc-2.9.so[7fd596e39000+14f000]
Comment 5 Ian Grant 2009-01-30 12:17:32 UTC
Note that the NEEDINFO is satisfied by my previous comment
Comment 6 Matthias Koenig 2009-02-04 13:13:26 UTC
I have an updated autofs package in my buildservice test project.
Can you please check if the problem still exists there?
You can find the package here:
http://download.opensuse.org/repositories/home://makoenig://test/openSUSE_11.1/
Comment 7 Ian Grant 2009-02-11 14:45:58 UTC
This RPM works. It also resolves the issues I reported in another bug about the syntax error in the mount maps.
Comment 9 Jiri Kosina 2009-12-04 14:11:39 UTC
*** Bug 551722 has been marked as a duplicate of this bug. ***
Comment 10 Forgotten User sLJ7K2dvxj 2010-01-04 14:04:15 UTC
OpenSUSE-11.2 has autofs-5.0.4-6.1, which should have this bug fixed.
Comment 11 Sven Bilke 2010-05-21 00:53:29 UTC
The bug still exists in OpenSuse 11.2, the error message is

automount[17820]: segfault at 7f532df3e7c8 ip 00007f532edcb93d sp 00007fff0142de40 error 4 in libcom_err.so.2.1[7f532edca000+3000]
Comment 12 Leonardo Chiquitto 2010-05-21 01:16:40 UTC
Sven, what's the autofs package version?
Comment 13 Sven Bilke 2010-05-21 13:56:41 UTC
Leonardo, the autofs version is 5.0.4-6.1 

If that helps: it appears autofs reads from a closed socket (I guess, talking to DNS), here are a few lines from 

strace automount -f -v


connect(5, {sa_family=AF_INET, sin_port=htons(53), sin_addr=inet_addr("192.168.8.2")}, 28) = 0

a lot of lines later

[pid 10170] shutdown(5, 2 /* send and receive */) = 0
[pid 10170] close(5)                    = 0
[pid 10170] read(4294967295, 0x7f3f1e249940, 5) = -1 EBADF (Bad file descriptor)
[pid 10170] shutdown(4294967295, 2 /* send and receive */) = -1 EBADF (Bad file descriptor)
[pid 10170] close(4294967295)           = -1 EBADF (Bad file descriptor)

and then a few lines later

[pid 10170] munmap(0x7f3f1b0fc000, 2118136) = 0
[pid 10170] --- SIGSEGV (Segmentation fault) @ 0 (0) ---

Sven
Comment 14 Leonardo Chiquitto 2010-05-21 14:09:42 UTC
Sven, thanks for the quick reply.

We released a maintenance update for AutoFS on 11.2 to fix bug #556479. This could be the same problem as the one you're seeing here. The packages are available in the Updates repository:

http://download.opensuse.org/update/11.2/rpm/x86_64/autofs-5.0.4-6.2.1.x86_64.rpm
http://download.opensuse.org/update/11.2/rpm/i586/autofs-5.0.4-6.2.1.i586.rpm

Can you please test and report if the problem is still reproducible? If yes, we need a core dump to be able to investigate further. To generate the core you'll probably need to set HARDCORELIMIT and SOFTCORELIMIT to "unlimited" in /etc/sysconfig/ulimit, restart the machine and reproduce the problem. Thanks!
Comment 15 Sven Bilke 2010-05-21 14:36:58 UTC
The problem still exists. Where should I send the core?
Comment 16 Leonardo Chiquitto 2010-05-21 14:52:24 UTC
You can attach the compressed core here in the bug (see "Add an attachment" below).
Comment 17 Sven Bilke 2010-05-21 15:05:05 UTC
Created attachment 363819 [details]
Core dumped by automount
Comment 18 Leonardo Chiquitto 2010-05-21 21:15:45 UTC
Thanks for attaching the core dump, it confirms that this is not the same problem.

The problem you're seeing was also reported in:
  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=527545

At least on Debian it was caused by a Kerberos update. We should open a new bug report to handle this but, before we do that, can you please apply all the updates released for 11.2 so far? Perhaps the problem is already resolved. Thanks.
Comment 19 Sven Bilke 2010-05-24 14:02:53 UTC
Leonardo, I made sure I have updated everything on the workstation, as well as the server for LDAP/KERBEROS (running Debian Lenny). The issue persists.
Comment 20 Leonardo Chiquitto 2010-05-24 14:09:28 UTC
Sven, thanks again. As this is a different problem, can you please open a new bug report against 11.2?
Comment 21 Sven Bilke 2010-05-24 14:34:15 UTC
Done:

https://bugzilla.novell.com/show_bug.cgi?id=608295
Comment 22 Leonardo Chiquitto 2010-05-24 14:38:33 UTC
Thanks! Back to RESOLVED/FIXED.