Postgres Pro
Downloads
Home   >   mailing lists

Thread: RE: 3867653, EMM1 cluster issue on 23rd May--core generated --design feedback

RE: 3867653, EMM1 cluster issue on 23rd May--core generated --design feedback

From
M Tarkeshwar Rao
Date:

Hi All,

 

We are getting following core with following use case:

 

Case: We made max_connections as 15 and 10 different processes opening 15 connections with server. It is running fine. But after few hours it giving following

          core. When we increase the max_connections then core come again but it take more hours to come.

 

Program terminated with signal 11, Segmentation fault.

#0  0x00007f712e7a3bf0 in pqRowProcessor () from /tmp/3pp/postgres//lib/libpq.so.5

Missing separate debuginfos, use: debuginfo-install CXC1741717-R2N-EM20_ICP2001.x86_64

(gdb)

(gdb) bt

#0  0x00007f712e7a3bf0 in pqRowProcessor () from /3pp/postgres//lib/libpq.so.5

#1  0x00007f712e7adef2 in pqParseInput3 () from /3pp/postgres//lib/libpq.so.5

#2  0x00007f712e7a4e78 in PQgetResult () from /3pp/postgres//lib/libpq.so.5

#3  0x00007f712e7a50af in PQexecStart () from /3pp/postgres//lib/libpq.so.5

#4  0x00007f712e7a5261 in PQexec () from /3pp/postgres//lib/libpq.so.5

#5  0x00007f712ea08add in PostGresSqlExecutor::executeSql (this=0x14c50090, aSqlStatement=0x1341fee8 "SELECT * from tab_1",

    thePrefetchCount=<optimized out>) at PostGresSqlExecutor.cc:107

 

Regards

Tarkeshwar

Re: 3867653, EMM1 cluster issue on 23rd May--core generated --design feedback

From
Simon Riggs
Date:
On Mon, 12 Jul 2021 at 16:07, M Tarkeshwar Rao
<m.tarkeshwar.rao@ericsson.com> wrote:
>
> Hi All,
>
> We are getting following core with following use case:
>
> Case: We made max_connections as 15 and 10 different processes opening 15 connections with server. It is running
fine.But after few hours it giving following
 
>           core. When we increase the max_connections then core come again but it take more hours to come.
>
> Program terminated with signal 11, Segmentation fault.

Sounds like a memory leak. In libpq programs this is often caused by
forgetting to run PQclear().

-- 
Simon Riggs                http://www.EnterpriseDB.com/