Re: BUG #14531: server process (PID 12714) was terminated by signal 11: Segmentation fault

Thanks for your reply firstly.

I did not mention anything about logs thinking the back trace would be
enough to solve the issue. Sorry for that.
I have enabled the *log_statement=all* already and each time it gives
different queries. But until now i have seen postmaster crashed during
SELECT query.

Eg: Crashed yesterday while executing query [PID] 11221 DETAIL: Failed
process was running: select n.nspname, c.relname, a.attname, a.atttypid,
t.typname, a.attnum, a.attlen, a.atttypmod, a.attnotnull, c.relhasrules,
c.relkind, c.oid, pg_get_expr(d.adbin, d.adrelid), case t.typtype when 'd'
then t.typbasetype else 0 end, t.typtypmod, c.relhasoids from
(((pg_catalog.pg_class c inner join pg_catalog.pg_namespace n on n.oid =
c.relnamespace and c.oid = 19175) inner join pg_catalog.pg_attribute a on
(not a.attisdropped) and a.attnum > 0 and a.attrelid = c.oid) inner join
pg_catalog.pg_type t on t.oid = a.atttypid) left outer join pg_attrdef d on
a.atthasdef and d.adrelid = a.attrelid and d.adnum = a.attnum order by
n.nspname, c.relname, attnum

For your Information the previous query in the log is "select COLUMN_VALUE
from TABLE_NAME where COLUMN_VALUE=3"

The crash is reproducible with different queries.

Previous crashed query(just format) log:- select column1,column2, column3
from table_name where column4>-1

Regards,
Naveen Ch.

On Tue, Feb 7, 2017 at 8:09 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> naveenallin(at)gmail(dot)com writes:
> > We are facing frequent crash issue in postgresql due to segmentation
> fault.
> > We have tried various ways to debug the issue such as enabling debug
> level
> > log etc., but we could not get any idea what is the reason behind the
> crash.
>
> The postmaster should at least have logged what statement a crashed
> backend was running (assuming you've not disabled the statistics
> subsystem). Is there any consistency to that? It might be worth
> turning on full statement logging (log_statement=all) so that you
> can see everything the crashed sessions did.
>
> The backtrace you show doesn't look much like any known 9.6 bug,
> so we'd be very interested if you can put together a reproducible
> test case.
>
> regards, tom lane
>