Re: BUG in 10.1 - dsa_area could not attach to a segment that hasbeen freed

On 11/29/2017 05:23 PM, Alexander Voytsekhovskyy wrote:
> thanks for detailed instructions!
> 
> Here it is:
> 
> Note: breakpoint 1 also set at pc 0x5560252c21f5.
> Breakpoint 2 at 0x5560252c21f5: file
> /build/postgresql-10-qAeTPy/postgresql-10-10.1/build/../src/backend/utils/mmgr/dsa.c,
> line 1736.
> (gdb) bt
> #0  0x00007f53e538e9b3 in __epoll_wait_nocancel () at
> ../sysdeps/unix/syscall-template.S:84
> #1  0x000055602516a791 in WaitEventSetWaitBlock (nevents=1,
> occurred_events=0x7ffc2849a460, cur_timeout=-1, set=0x556026359498) at
> /build/postgresql-10-qAeTPy/postgresql-10-10.1/build/../src/backend/storage/ipc/latch.c:1048

Unfortunately, that's still not it.  This shows the backtrace at the
moment when gdb attached to the process (because it interrupts the
exectution). You need to do "c" first, to continue the execution.

So it should be:

1) gdb -p $PID
2) (gdb) handle SIGUSR1 noprint nostop
3) (gdb) break dsa.c:1736
4) (gdb) c
5) run the query, gdb should interrupt at the breakpoint
6) (gdb) bt


regards

-- 
Tomas Vondra                  http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services