script exits prematurely with no stderr output, but with system errors

[Larry Martell]

On Sun, Mar 20, 2016 at 4:55 PM, Larry Martell <larry.martell at gmail.com> wrote:
> On Sun, Mar 20, 2016 at 4:47 PM, Joel Goldstick
> <joel.goldstick at gmail.com> wrote:
>> On Sun, Mar 20, 2016 at 4:32 PM, Larry Martell <larry.martell at gmail.com>
>> wrote:
>>
>>> I have a script that I run a lot - at least 10 time every day. Usually
>>> it works fine. But sometime it just stops running with nothing output
>>> to stdout or stderr. I've been trying to debug this for a while, and
>>> today I looked in the system logs and saw this:
>>>
>>> abrt: detected unhandled Python exception in
>>> '/home/prod_user/python/make_workitem_list.py'
>>> abrtd: Directory 'pyhook-2016-03-19-22:20:43-26461' creation detected
>>> abrt-server[3688]: Saved Python crash dump of pid 26461 to
>>> /var/spool/abrt/pyhook-2016-03-19-22:20:43-26461
>>> abrtd: Executable '/home/prod_user/python/make_workitem_list.py'
>>> doesn't belong to any package and ProcessUnpackaged is set to 'no'
>>> abrtd: 'post-create' on
>>> '/var/spool/abrt/pyhook-2016-03-19-22:20:43-26461' exited with 1
>>> abrtd: Deleting problem directory
>>> '/var/spool/abrt/pyhook-2016-03-19-22:20:43-26461'
>>> abrtd: make_workitem_list: page allocation failure. order:1, mode:0x20
>>> abrtd: Pid: 31870, comm: make_workitem_list Not tainted
>>> 2.6.32-573.12.1.el6.x86_64 #1
>>>
>>> I have never seen anything like this before. Usually, if there is an
>>> unhandled exception something is dumped to stderr. Anyone have any
>>> idea what is going on? How can I get it to not delete this crash dump
>>> it mentioned? I guess I can put a big exception handler around the
>>> enter script with a traceback.
>>>
>>> This is on Red Hat Enterprise Linux Server release 6.7 (Santiago).
>>
>> Googling I found this:
>> http://stackoverflow.com/questions/2628901/interpreting-kernel-message-page-allocation-failure-order1
>>
>> It seems that the kernel can't allocate memory is a likely cause.

I modified the program to use less memory and I am not getting the
page allocation failure any more, but I am still getting the unhanded
exceptions messages in the system logs.

> Yes, I was thinking that as well about the "page allocation failure"
> message, but it's almost like there were 2 errors, the first being the
> unhandled exception. But why would it not output something to stderr?

I was able to configure the abrt daemon to capture these unhanded
exceptions and they are just typical ValueErrors and TypeErrors (which
I will of course deal with in my program). But I wonder why these
exceptions were not just printed to stderr as other unhanded
exceptions are. What is special about these that made the OS get
involved?