|
|
|
Oct 18 11:25:26 c0558 kernel: get[4386]: segfault at ffffffff ip
00000000ffffffff sp 00000000ffa6497c error 14
Oct 18 11:25:29 c0558 abrtd:
Directory 'ccpp-2012-10-18-11:25:26-4386' creation detected
Oct 18 11:25:29
c0558 abrt[10384]: Saved core dump of pid 4386 (/usr/genesis/e99/get/get) to
/var/spool/abrt/ccpp-2012-10-18-11:25:26-4386 (120422400 bytes)
Oct 18
11:25:30 c0558 abrtd: Executable '/usr/genesis/e99/get/get' doesn't belong to
any package
Oct 18 11:25:30 c0558 abrtd: 'post-create' on
'/var/spool/abrt/ccpp-2012-10-18-11:25:26-4386' exited with 1
Oct 18 11:25:30
c0558 abrtd: Corrupted or bad directory
/var/spool/abrt/ccpp-2012-10-18-11:25:26-4386, deleting
Having been underwhelmed by help and advice, I can now, 7 months on, answer my own question. I post the answer here just in case it is useful to anyone else.
Edit the file /etc/abrt/abrt-action-save-package-data.conf to change the line
ProcessUnpackaged = no
to
ProcessUnpackaged = yes
This now allows applications not installed via the package manager to create core dumps. This change is necessary, but I am not entirely sure whether it is sufficient: you may also need to issue a command like the following in the shell that will receive the core dump:
limit coredumpsize unlimited
When a core dump is created it appears in the current directory of the process that crashed. A copy of the core dump and quite a few supporting files are placed in /var/spool/abrt/ccpp-<yyyy>-<mm>-<dd>-<HH>:<MM>:<SS>-<pid>. These supporting files appear to be very useful, since they provide lists of things like environment variables, libraries loaded, open file descriptors, and even an extract from /var/log/messages.
[/etc/abrt/abrt-action-save-package-data.conf] 의 항목을 아래와 같이 수정후
/etc/init.d/abrtd restart
OpenGPGCheck = no
ProcessUnpackaged =
yes
