Re: The same app name and symbol name
Michael Petlan <[email protected]>
| Newsgroups | gmane.linux.oprofile |
|---|---|
| Organization | Red Hat |
| Message-ID | <1447093510.15375.64.camel@Rudolf-RHEL-7> |
I have reproduced the behaviour with libgfortran: $ opreport --demangle=smart --symbols --long-filenames --merge all samples % app name symbol name [...] 1 0.0104 /usr/lib64/libgfortran.so.3.0.0 /usr/lib64/libgfortran.so.3.0.0 After installing debuginfo for the library, the same report showed symbol name as well: 1 0.0104 /usr/lib64/libgfortran.so.3.0.0 parse_format_list You even don't need to re-run operf, just opreport. That is also a bit safer, since new measurement may have different results. Try to check, whether the debuginfo for PostgreSQL is present. On RHEL/Fedora it should be in postgresql-debuginfo package. If not, install it and re-run the opreport command. Hope it helps, Michael On Mon, 2015-11-09 at 11:56 -0500, Tim Chou wrote: > Hi Michael, > > > Here is my code: > #!/bin/sh > > > echo "Delete previous files" > > rm -rf test* > > rm -rf /tmp/test* > > rm -rf op-re* > > > echo "Start OProfile" > > nohup sudo operf --system-wide --separate-thread --separate-cpu > > /dev/null & > > > echo "Start Tests" > > dbt2-run-workload -c 20 -a pgsql -d 100 -w 10 -n -o /tmp/test8 > test8 > > echo "Finish Tests" > > > echo "Start Generating Report" > > #sudo opreport --demangle=smart --symbols --long-filenames --merge > tgid $(which postgres) -o ./op-result-postgres > > sudo opreport --demangle=smart --symbols --long-filenames --merge all > -o ./op-result-postgres > > echo "Finish Generating Report" > > > sudo pkill operf > > echo "Finish OProfile" > > > Thanks, > Tim > > > > 2015-11-09 10:53 GMT-05:00 Michael Petlan <[email protected]>: > On Mon, 2015-11-09 at 05:36 -0500, Tim Chou wrote: > > Hi All, > > > > > Hi Tim, > > > Recently, I'm testing the performance of PostgreSQL with > oprofile > > v1.0.0. > > > > > > However, I found the app name and symbol name are the same > in my > > opreport output, like below: > > samples % app name symbol name > > 28 0.0857 /usr/bin/postgres /usr/bin/postgres > > 8 0.0245 /usr/bin/psql /usr/bin/psql > > > > which command(s) did you use for producing it? > > > What's wrong with my program? > > > > Looking forward to your reply. > > > > Thanks, > > Tim > > Michael > > > > > > > ------------------------------------------------------------------------------ > > Presto, an open source distributed SQL query engine for big > data, initially > > developed by Facebook, enables you to easily query your data > on Hadoop in a > > more interactive manner. Teradata is also now providing full > enterprise > > support for Presto. Download a free open source copy now. > > > http://pubads.g.doubleclick.net/gampad/clk?id=250295911&iu=/4140 > > _______________________________________________ > > oprofile-list mailing list > > [email protected] > > https://lists.sourceforge.net/lists/listinfo/oprofile-list > > > > ------------------------------------------------------------------------------ Presto, an open source distributed SQL query engine for big data, initially developed by Facebook, enables you to easily query your data on Hadoop in a more interactive manner. Teradata is also now providing full enterprise support for Presto. Download a free open source copy now. http://pubads.g.doubleclick.net/gampad/clk?id=250295911&iu=/4140