CSS/Segfault problem

Just checking in to see if anyone had come up with a reproducible solution
to the CSS/Segfault problem that has been discussed repeatedly. I am
seeing it on a RH 7.1 system.

Dave

David C. Troy [dave@toad.net] 410-544-6193 Sales
ToadNet - Want to go fast? 410-544-1329 FAX
570 Ritchie Highway, Severna Park, MD 21146-2925 www.toad.net

Come to think of it, I haven’t seen that one since my last
install (2.0.8). I checked my error_log and nothing for at
least 30 days. I built every thing by hand. Here’s some snips from
my install notes:

I’m building postgres, perl, apache, and mod-perl from sources:
perl-5.6.1
postgresql-7.1.3
apache_1.3.22
mod_perl-1.26
rt-2-0-8

Mail server is postfix from SuSe7.2 RPMs

fisrt I built a new kernel.
develrt:/usr/include # uname -r
2.4.9-ac9-2

Then I built perl 5.6.1 with libperl as shared object:
develrt:/usr/local/build/perl-5.6.1 # sh Configure -Duseshrplib
develrt:/usr/local/build/perl-5.6.1 # make
develrt:/usr/local/build/perl-5.6.1 # make test
develrt:/usr/local/build/perl-5.6.1 # make install

Building postgres

develrt:/usr/local/build/postgresql-7.1.3 # cat startbuild
#!/bin/bash

This is the flags I’m using to configure building postgres

./configure --enable-locale
–with-perl
–enable-multibyte
–enable-unicode-conversion
–enable-syslog
–with-python
–with-openssl

develrt:/usr/local/build/postgresql-7.1.3 # ./startbuild
develrt:/usr/local/build/postgresql-7.1.3 # make
develrt:/usr/local/build/postgresql-7.1.3 # make install

develrt:/usr/local/build/postgresql-7.1.3 # grep postgres /etc/passwd
postgres:x:26:2:Postgres Database Admin:/usr/local/pgsql:/bin/bash
develrt:/usr/local # chown -R postgres.daemon pgsql/
develrt:/usr/local # su - postgres
postgres@develrt:~ > /usr/local/pgsql/bin/initdb -D
/usr/local/pgsql/data
postgres@develrt:~ > /usr/local/pgsql/bin/pg_ctl -D
/usr/local/pgsql/data -l log

The “-i” flag is needed so that rt can connect over the loopback
interface
I set up this start command line in /etc/init.d/boot.local:

start postgres

/usr/bin/sudo -u postgres /usr/bin/pg_ctl -o “-i -p 5432” -D
/var/lib/pgsql/data

Building apache with mod_perl non-DSO

develrt:/usr/local/build/mod_perl-1.26 # perl Makefile.PL
APACHE_SRC=…/apache_1
develrt:/usr/local/build/mod_perl-1.26 # make && make test && make
install
develrt:/usr/local/build/mod_perl-1.26 # cd …/apache_1.3.22/
develrt:/usr/local/build/apache_1.3.22 # make install

  1. Satisfy RT’s myriad dependencies

ran make fixdeps, configured cpan. Looked good up to DBD-Pg:

CPAN.pm: Going to build M/ME/MERGL/DBD-Pg-1.01.tar.gz

Configuring Pg
Remember to actually read the README file !
please set environment variables POSTGRES_INCLUDE and POSTGRES_LIB !
Running make test
Make had some problems, maybe interrupted? Won’t test
Running make install
Make had some problems, maybe interrupted? Won’t install

I’m buidling it by hand:
develrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # export
POSTGRES_HOME=/usr/loc
develrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # perl Makefile.PL
velrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # make
develrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # make test
develrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # su postgres
develrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # make test
develrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # exit
develrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # make install

Looks good. Trying make fixdeps again:

got everything but Apache::Session. Fails on install because failed
tests were looking for Storable.pmi and Digest/MD5.pm. So I installed
that with cpan:
develrt:/usr/local/build/rt-2-0-8 # perl -MCPAN -e shell
cpan> install Storable
cpan> install Digest::MD5

develrt:/usr/local/build/.cpan/build/Apache-Session-1.54 # make test
develrt:/usr/local/build/.cpan/build/Apache-Session-1.54 # make install

My kit is all good. But wait! there was a post on the RT users list:

have you installed DBI::SearchBuilder 0.46-test1, as recommended in
last week’s conversation about performance tuning?

I think I had better do it. Oh, well. All I could find was
DBIx::SearchBuiler-0.46. I hope that is recent enough.

  1. 5a FOR A NEW INSTALLATION:

First I couldn’t connect to the database:
I then restarted postgres with the -i flag:
sudo -u postgres postmaster -i -D /usr/local/pgsql/data/ &

  -i     Allows  clients  to  connect  via  TCP/IP (Internet
          domain)  connections.  Without  this  option,  only
          local  Unix domain socket connections are accepted.

This way the RT install could connect over the loopback interface.On Thu, Nov 29, 2001 at 08:14:45PM -0500, David C. Troy wrote:

Just checking in to see if anyone had come up with a reproducible solution
to the CSS/Segfault problem that has been discussed repeatedly. I am
seeing it on a RH 7.1 system.

Dave

=====================================================================
David C. Troy [dave@toad.net] 410-544-6193 Sales
ToadNet - Want to go fast? 410-544-1329 FAX
570 Ritchie Highway, Severna Park, MD 21146-2925 www.toad.net


rt-users mailing list
rt-users@lists.fsck.com
http://lists.fsck.com/mailman/listinfo/rt-users

I beleive the primary difference for me was building perl 5.6.1
with libperl as shared object.

I have built RT no less than 6 times. This was the primary difference
in my last build.

ashleyOn Fri, Nov 30, 2001 at 09:41:42AM -0600, Jay Kramer wrote:

I want to take a poll about this, because I have “NEVER” seen this problem,
but, the only thing I can think I have done differently is that I built
everything myself, and did not use the RPM’s or DEB’s or whatnot :wink: How
many people that have seen this problem built from source? How many
installed their distributions packages? What distribution? What were
versions of ‘the important stuff’ did you have installed? I keep leaning
towards this being a packaging problem with one (or more) of the
distributions…

Jay

Come to think of it, I haven’t seen that one since my last
install (2.0.8). I checked my error_log and nothing for at
least 30 days. I built every thing by hand. Here’s some snips from
my install notes:

I’m building postgres, perl, apache, and mod-perl from sources:
perl-5.6.1
postgresql-7.1.3
apache_1.3.22
mod_perl-1.26
rt-2-0-8

Mail server is postfix from SuSe7.2 RPMs


fisrt I built a new kernel.
develrt:/usr/include # uname -r
2.4.9-ac9-2


Then I built perl 5.6.1 with libperl as shared object:
develrt:/usr/local/build/perl-5.6.1 # sh Configure -Duseshrplib
develrt:/usr/local/build/perl-5.6.1 # make
develrt:/usr/local/build/perl-5.6.1 # make test
develrt:/usr/local/build/perl-5.6.1 # make install


Building postgres

develrt:/usr/local/build/postgresql-7.1.3 # cat startbuild
#!/bin/bash

This is the flags I’m using to configure building postgres

./configure --enable-locale
–with-perl
–enable-multibyte
–enable-unicode-conversion
–enable-syslog
–with-python
–with-openssl

develrt:/usr/local/build/postgresql-7.1.3 # ./startbuild
develrt:/usr/local/build/postgresql-7.1.3 # make
develrt:/usr/local/build/postgresql-7.1.3 # make install

develrt:/usr/local/build/postgresql-7.1.3 # grep postgres /etc/passwd
postgres:x:26:2:Postgres Database Admin:/usr/local/pgsql:/bin/bash
develrt:/usr/local # chown -R postgres.daemon pgsql/
develrt:/usr/local # su - postgres
postgres@develrt:~ > /usr/local/pgsql/bin/initdb -D
/usr/local/pgsql/data
postgres@develrt:~ > /usr/local/pgsql/bin/pg_ctl -D
/usr/local/pgsql/data -l log

The “-i” flag is needed so that rt can connect over the loopback
interface
I set up this start command line in /etc/init.d/boot.local:

start postgres

/usr/bin/sudo -u postgres /usr/bin/pg_ctl -o “-i -p 5432” -D
/var/lib/pgsql/data


Building apache with mod_perl non-DSO

develrt:/usr/local/build/mod_perl-1.26 # perl Makefile.PL
APACHE_SRC=…/apache_1
develrt:/usr/local/build/mod_perl-1.26 # make && make test && make
install
develrt:/usr/local/build/mod_perl-1.26 # cd …/apache_1.3.22/
develrt:/usr/local/build/apache_1.3.22 # make install


  1. Satisfy RT’s myriad dependencies

ran make fixdeps, configured cpan. Looked good up to DBD-Pg:

CPAN.pm: Going to build M/ME/MERGL/DBD-Pg-1.01.tar.gz

Configuring Pg
Remember to actually read the README file !
please set environment variables POSTGRES_INCLUDE and POSTGRES_LIB !
Running make test
Make had some problems, maybe interrupted? Won’t test
Running make install
Make had some problems, maybe interrupted? Won’t install

I’m buidling it by hand:
develrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # export
POSTGRES_HOME=/usr/loc
develrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # perl Makefile.PL
velrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # make
develrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # make test
develrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # su postgres
develrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # make test
develrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # exit
develrt:/usr/local/build/.cpan/build/DBD-Pg-1.01 # make install

Looks good. Trying make fixdeps again:

got everything but Apache::Session. Fails on install because failed
tests were looking for Storable.pmi and Digest/MD5.pm. So I installed
that with cpan:
develrt:/usr/local/build/rt-2-0-8 # perl -MCPAN -e shell
cpan> install Storable
cpan> install Digest::MD5

develrt:/usr/local/build/.cpan/build/Apache-Session-1.54 # make test
develrt:/usr/local/build/.cpan/build/Apache-Session-1.54 # make install

My kit is all good. But wait! there was a post on the RT users list:

have you installed DBI::SearchBuilder 0.46-test1, as recommended in
last week’s conversation about performance tuning?

I think I had better do it. Oh, well. All I could find was
DBIx::SearchBuiler-0.46. I hope that is recent enough.


  1. 5a FOR A NEW INSTALLATION:

First I couldn’t connect to the database:
I then restarted postgres with the -i flag:
sudo -u postgres postmaster -i -D /usr/local/pgsql/data/ &

  -i     Allows  clients  to  connect  via  TCP/IP (Internet
          domain)  connections.  Without  this  option,  only
          local  Unix domain socket connections are accepted.

This way the RT install could connect over the loopback interface.

On Thu, Nov 29, 2001 at 08:14:45PM -0500, David C. Troy wrote:

Just checking in to see if anyone had come up with a reproducible
solution

to the CSS/Segfault problem that has been discussed repeatedly. I am
seeing it on a RH 7.1 system.

Dave

=====================================================================
David C. Troy [dave@toad.net] 410-544-6193 Sales
ToadNet - Want to go fast? 410-544-1329 FAX
570 Ritchie Highway, Severna Park, MD 21146-2925 www.toad.net


rt-users mailing list
rt-users@lists.fsck.com
http://lists.fsck.com/mailman/listinfo/rt-users


rt-users mailing list
rt-users@lists.fsck.com
http://lists.fsck.com/mailman/listinfo/rt-users