Apport is the Python script provided by Canonical to send the core dump to their server. And it seems it crashed too :)
The best way to debug this at this stage is the following:
1. Disable apport with /etc/init.d/apport stop
2. Stop gnocchi-metricd
3. Make sure you have no limit on core dump size: ulimit -c unlimited
4. Run gnocchi-metricd in foreground: gnocchi-metricd --debug
5. Wait for it to crash :)
6. When it segfaults, you should have a "core" file in the current directory
7. With that core file, run gdb as follow: gdb /usr/bin/python2.7 core
8. At the (gdb) prompt type: bt full
That should give you the full backtrace of the segmentation fault and what caused it. It's likely caused by a bug in librados, so it's pretty likely we won't be able to fix, but it can give us a hint.
Apport is the Python script provided by Canonical to send the core dump to their server. And it seems it crashed too :)
The best way to debug this at this stage is the following:
1. Disable apport with /etc/init.d/apport stop
2. Stop gnocchi-metricd
3. Make sure you have no limit on core dump size: ulimit -c unlimited
4. Run gnocchi-metricd in foreground: gnocchi-metricd --debug
5. Wait for it to crash :)
6. When it segfaults, you should have a "core" file in the current directory
7. With that core file, run gdb as follow: gdb /usr/bin/python2.7 core
8. At the (gdb) prompt type: bt full
That should give you the full backtrace of the segmentation fault and what caused it. It's likely caused by a bug in librados, so it's pretty likely we won't be able to fix, but it can give us a hint.