Segfaults

Thu Jun 23 15:57:58 UTC 2016

Hi Tyler,
that is not a file path for the dump but the pipe to the apport script and
it seems that fails.
Unfortunately that leaves my area of experience - I've never seen this
either.

I hate to not be able to help more, and that it might be a two stage issue.
But I think with that I'd redirect you to move the thread to ubuntu-devel
list an/or IRC.
There are the owners and experts on apport and such to finally get a proper
crash to analyze your original issue.

Kind Regards,
Christian

Christian Ehrhardt
Software Engineer, Ubuntu Server
Canonical Ltd

On Thu, Jun 23, 2016 at 4:21 PM, Tyler Petersen <tpetersen at xes-inc.com>
wrote:

> Hi Christian,
>
>
> I looked and there is no log for apport in /var/log/. Also I check dmesg
> and I
> now see this error.
>
>
> [2920471.605928] mdadm[26249]: segfault at 8 ip 00007f6cb8f0ed16 sp
> 00007fffd5bb20e8 error 4 in libc-2.23.so[7f6cb8e84000+1c0000]
> [2920471.606068] Core dump to |/usr/share/apport/apport 15730 11 0 26249
> pipe
> failed
>
>
> Which is different than before. I have never seen the core dump error
> before.
> Any ideas on what that means? Sounds like apport isn't working correctly.
>
>
> Thanks!
>
> Tyler
>
> ------------------------------
> *From: *"Christian Ehrhardt" <christian.ehrhardt at canonical.com>
> *To: *"Tyler Petersen" <tpetersen at xes-inc.com>
> *Cc: *"ubuntu-server" <ubuntu-server at lists.ubuntu.com>
> *Sent: *Thursday, June 23, 2016 1:08:09 AM
>
> *Subject: *Re: Segfaults
>
> Hi Tyler,
> well in that case we have to find why your apport isn't collecting the
> crash, because that is the start to look deeper into it.
> There is /var/log/apport* which might help you to understand what is going
> on.
>
> I quickly ran this:
> #include <signal.h>
> int main()
> {
>     raise(SIGSEGV);
> }
>
> And got:
> cat /var/log/apport.log
> ERROR: apport (pid 26269) Thu Jun 23 07:50:47 2016: called for pid 26268,
> signal 11, core limit 0
> ERROR: apport (pid 26269) Thu Jun 23 07:50:47 2016: executable:
> /tmp/segfaultme (command line "./segfaultme")
> ERROR: apport (pid 26269) Thu Jun 23 07:50:47 2016: executable does not
> belong to a package, ignoring
>
> Well that is fair, but at least it picked up the fault happening for sure.
>
> Next I let something segfault that is in a package
> md5sum /dev/urandom &
> kill --signal SIGSEGV $!
>
> Here I already got my "there was a crash popup by apport" since I'm on a
> desktop.
> In the log I found:
> ERROR: apport (pid 28384) Thu Jun 23 08:01:01 2016: called for pid 28159,
> signal 11, core limit 0
> ERROR: apport (pid 28384) Thu Jun 23 08:01:01 2016: executable:
> /usr/bin/md5sum (command line "md5sum /dev/urandom")
> ERROR: apport (pid 28384) Thu Jun 23 08:01:01 2016: gdbus call error:
> Error: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name
> org.gnome.SessionManager was not provided by any .service files
> ERROR: apport (pid 28384) Thu Jun 23 08:01:01 2016: debug: session gdbus
> call:
> ERROR: apport (pid 28384) Thu Jun 23 08:01:02 2016: wrote report
> /var/crash/_usr_bin_md5sum.1000.crash
>
>
> Debugging starts like:
> sudo apport-retrace --stdout --rebuild-package-info
>  /var/crash/_usr_bin_md5sum.1000.crash
> sudo apport-retrace --rebuild-package-info --gdb
> /var/crash/_usr_bin_md5sum.1000.crash
>
> I hope that helps to find why crashes aren't collected for you and then
> also about the first few steps into debugging if you like.
>
>
> Christian Ehrhardt
> Software Engineer, Ubuntu Server
> Canonical Ltd
>
> On Wed, Jun 22, 2016 at 9:45 PM, Tyler Petersen <tpetersen at xes-inc.com>
> wrote:
>
>> Hi everyone,
>>
>> I am still having problems with this. I made sure apport was running and
>> the /var/crash directory exists. But when
>> a segfault happens nothing is written to the directory. The segfaults
>> also seem to come and go.
>>
>> Thanks,
>>
>> Tyler
>>
>> ------------------------------
>> *From: *"Christian Ehrhardt" <christian.ehrhardt at canonical.com>
>> *To: *"Tyler Petersen" <tpetersen at xes-inc.com>
>> *Cc: *"ubuntu-server" <ubuntu-server at lists.ubuntu.com>
>> *Sent: *Wednesday, June 15, 2016 10:47:37 AM
>> *Subject: *Re: Segfaults
>>
>> Hi Tyler,
>> when an appliaction is crashing it should leave a .crash file in
>> /var/crash which stores debugging information.
>> Base on that you can automatically report a bug that has most of the info
>> to start with.
>> See https://help.ubuntu.com/community/ReportingBugs#Crash_reports for
>> more.
>>
>> If you want to look into it more yourself I'd recommend starting with
>> https://wiki.ubuntu.com/DebuggingProgramCrash
>> https://www.piware.de/2011/08/apport-retrace-made-useful/
>>
>> Kind regards,
>> Christian
>>
>> Christian Ehrhardt
>> Software Engineer, Ubuntu Server
>> Canonical Ltd
>>
>> On Wed, Jun 15, 2016 at 3:55 PM, Tyler Petersen <tpetersen at xes-inc.com>
>> wrote:
>>
>>> Hi everyone,
>>> We are currently experiencing some odd issues on our ubuntu servers
>>> (14.04). We
>>> keep seeing segfault messages for mdadm but we are not sure if they are
>>> a bug or
>>> a real problem. The servers are two completely different physical boxes.
>>> Both
>>> are running the 3.13.0-88-generic kernal and mdadm version
>>> 3.2.5-5ubuntu4.3.
>>> Here are the two most common errors we see: "[2201076.898901]
>>> mdadm[17298]:
>>> segfault at 8 ip 00007f3445891d16 sp 00007ffc2cd425f8 error 4 in
>>> libc-2.23.so[7f3445807000+1c0000]" and "[39304.894195] mdadm[17613]:
>>> segfault at
>>> 0 ip 00007fb29477cd16 sp 00007ffe8ce211d8 error 4 in
>>> libc-2.23.so[7fb2946f2000+1c0000]". Any help or advice would be greatly
>>> apperciated.
>>> Thanks,
>>> Tyler
>>>
>>> --
>>> ubuntu-server mailing list
>>> ubuntu-server at lists.ubuntu.com
>>> https://lists.ubuntu.com/mailman/listinfo/ubuntu-server
>>> More info: https://wiki.ubuntu.com/ServerTeam
>>>
>>
>>
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/ubuntu-server/attachments/20160623/4774afa4/attachment.html>