RT 3.8 on Ubuntu Hardy

Hi, I’ve installed RT 3.8 on an Ubuntu Hardy server from the Debian package
at

http://svn.debian.org/wsvn/pkg-request-tracker/packages/request-tracker3.8/trunk/debian/

It works totally fine aside from the fact that it’s got a significant visual
bug. Here’s how it’s produced:

  1. Click on queue in the Quick Search menu on the right hand side.
  2. Click on a ticket in that queue.
  3. Go back to the list of tickets.
  4. Refresh the page.

The error that occurs is that nothing gets rendered properly, so it’s just a
white background with a bunch of blue links for the tickets and text for the
different fields of each ticket. In /var/log/apache2/error.log I see stuff
like:

[Tue Mar 10 13:38:48 2009] [notice] child pid 24484 exit signal Segmentation
fault (11)
Out of memory!
Callback called exit.

and:

[Tue Mar 10 18:02:36 2009] [crit]: Apache2::RequestIO::print: (104)
Connection reset by peer at /usr/share/perl5/HTML/Mason/ApacheHandler.pm
line 1034 (/usr/share/request-tracker3.8/libexec/webmux.pl:150)

My boss suggested changing MaxRequestsPerChild in /etc/apache2/apache2.conf
to a low number other than 0, but that still did not fix the error. If
anyone has any suggestions I would greatly appreciate any advice you can
provide.

Shawn

Last time I saw that on an install (not debian) it needed a good
run-through of the instructions in UPGRADING.mysql

Shawn Smith wrote:

Hi, I’ve installed RT 3.8 on an Ubuntu Hardy server from the Debian
package at

http://svn.debian.org/wsvn/pkg-request-tracker/packages/request-tracker3.8/trunk/debian/

It works totally fine aside from the fact that it’s got a significant
visual bug. Here’s how it’s produced:

  1. Click on queue in the Quick Search menu on the right hand side.
  2. Click on a ticket in that queue.
  3. Go back to the list of tickets.
  4. Refresh the page.

The error that occurs is that nothing gets rendered properly, so it’s
just a white background with a bunch of blue links for the tickets and
text for the different fields of each ticket. In
/var/log/apache2/error.log I see stuff like:

[Tue Mar 10 13:38:48 2009] [notice] child pid 24484 exit signal
Segmentation fault (11)
Out of memory!
Callback called exit.

and:

[Tue Mar 10 18:02:36 2009] [crit]: Apache2::RequestIO::print: (104)
Connection reset by peer at
/usr/share/perl5/HTML/Mason/ApacheHandler.pm line 1034
(/usr/share/request-tracker3.8/libexec/webmux.pl:150)

My boss suggested changing MaxRequestsPerChild in
/etc/apache2/apache2.conf to a low number other than 0, but that still
did not fix the error. If anyone has any suggestions I would greatly
appreciate any advice you can provide.


Shawn


The rt-users Archives

Community help: http://wiki.bestpractical.com
Commercial support: sales@bestpractical.com

Discover RT’s hidden secrets with RT Essentials from O’Reilly Media.
Buy a copy at http://rtbook.bestpractical.com

Drew Barnes
Applications Analyst
Network Resources Department
Raymond Walters College
University of Cincinnati

Fix for: Out of memory! Callback called exit.

First login to your mysql database

Mysql –p –u root –D rtdb

This will log you in as root(-u) in the rtdb database(-D) with a prompt for a password(-p). Next, we want to change the session tables. Type in this exact command when you get to the

mysql> prompt (note capitalization areas and DO NOT include the ""s):

“ALTER TABLE sessions CHANGE a_session a_session LONGBLOB;”

Shawn Smith wrote:

Hi, I’ve installed RT 3.8 on an Ubuntu Hardy server from the Debian
package at

http://svn.debian.org/wsvn/pkg-request-tracker/packages/request-tracker3.8/trunk/debian/

It works totally fine aside from the fact that it’s got a significant
visual bug. Here’s how it’s produced:

  1. Click on queue in the Quick Search menu on the right hand side.
  2. Click on a ticket in that queue.
  3. Go back to the list of tickets.
  4. Refresh the page.

The error that occurs is that nothing gets rendered properly, so it’s
just a white background with a bunch of blue links for the tickets and
text for the different fields of each ticket. In
/var/log/apache2/error.log I see stuff like:

[Tue Mar 10 13:38:48 2009] [notice] child pid 24484 exit signal
Segmentation fault (11)
Out of memory!
Callback called exit.

and:

[Tue Mar 10 18:02:36 2009] [crit]: Apache2::RequestIO::print: (104)
Connection reset by peer at
/usr/share/perl5/HTML/Mason/ApacheHandler.pm line 1034
(/usr/share/request-tracker3.8/libexec/webmux.pl:150)

My boss suggested changing MaxRequestsPerChild in
/etc/apache2/apache2.conf to a low number other than 0, but that still
did not fix the error. If anyone has any suggestions I would greatly
appreciate any advice you can provide.


Shawn


The rt-users Archives

Community help: http://wiki.bestpractical.com
Commercial support: sales@bestpractical.com

Discover RT’s hidden secrets with RT Essentials from O’Reilly Media.
Buy a copy at http://rtbook.bestpractical.com

Steve Christou
UWM Information Security
Phone: 414-229-1100
Email: Osa-list@uwm.edu

Interesting, I can’t reproduce the error anymore after following your
instructions. Thanks a lot! Would you mind offering a quick explanation as
to why you think this fixes the problem? I’m a little confused about what
it’s doing and what it’s fixing. Thanks again.On Tue, Mar 10, 2009 at 2:28 PM, Steven Christou christ66@uwm.edu wrote:

Fix for: Out of memory! Callback called exit.

First login to your mysql database

Mysql –p –u root –D rtdb

This will log you in as root(-u) in the rtdb database(-D) with a prompt for
a password(-p). Next, we want to change the session tables. Type in this
exact command when you get to the

mysql> prompt (note capitalization areas and DO NOT include the ""s):

“ALTER TABLE sessions CHANGE a_session a_session LONGBLOB;”

Shawn Smith wrote:

Hi, I’ve installed RT 3.8 on an Ubuntu Hardy server from the Debian
package at

http://svn.debian.org/wsvn/pkg-request-tracker/packages/request-tracker3.8/trunk/debian/

It works totally fine aside from the fact that it’s got a significant
visual bug. Here’s how it’s produced:

  1. Click on queue in the Quick Search menu on the right hand side.
  2. Click on a ticket in that queue.
  3. Go back to the list of tickets.
  4. Refresh the page.

The error that occurs is that nothing gets rendered properly, so it’s
just a white background with a bunch of blue links for the tickets and
text for the different fields of each ticket. In
/var/log/apache2/error.log I see stuff like:

[Tue Mar 10 13:38:48 2009] [notice] child pid 24484 exit signal
Segmentation fault (11)
Out of memory!
Callback called exit.

and:

[Tue Mar 10 18:02:36 2009] [crit]: Apache2::RequestIO::print: (104)
Connection reset by peer at
/usr/share/perl5/HTML/Mason/ApacheHandler.pm line 1034
(/usr/share/request-tracker3.8/libexec/webmux.pl:150)

My boss suggested changing MaxRequestsPerChild in
/etc/apache2/apache2.conf to a low number other than 0, but that still
did not fix the error. If anyone has any suggestions I would greatly
appreciate any advice you can provide.


Shawn


The rt-users Archives

Community help: http://wiki.bestpractical.com
Commercial support: sales@bestpractical.com

Discover RT’s hidden secrets with RT Essentials from O’Reilly Media.
Buy a copy at http://rtbook.bestpractical.com


Steve Christou
UWM Information Security
Phone: 414-229-1100
Email: Osa-list@uwm.edu

Shawn

Shawn Smith wrote:

Interesting, I can’t reproduce the error anymore after following your
instructions. Thanks a lot! Would you mind offering a quick explanation as
to why you think this fixes the problem? I’m a little confused about what
it’s doing and what it’s fixing. Thanks again.

There are database schema changes in RT between versions. You really need
to go through UPGRADING.mysql and do what it says; that one thing isn’t
going to cover all the changes.

In particular, its quite likely that all your binary attachments are
non-functional. Try downloading one and opening it.

And please don’t top post :slight_smile:

– ============================
Tom Lahti
BIT Statement LLC

(425)251-0833 x 117
http://www.bitstatement.net/
– ============================

Go to this website and read the Continuous login part:
http://wiki.bestpractical.com/view/TroubleShooting
Probably can explain it 1000x better than I can, but basically its a
table issue converting from a longtext to longblob. The command below
does exactly that.

Do you still get the 2nd error?

Shawn Smith wrote:

Interesting, I can’t reproduce the error anymore after following your
instructions. Thanks a lot! Would you mind offering a quick
explanation as to why you think this fixes the problem? I’m a little
confused about what it’s doing and what it’s fixing. Thanks again.

Fix for: Out of memory! Callback called exit.

First login to your mysql database

Mysql �p �u root �D rtdb

This will log you in as root(-u) in the rtdb database(-D) with a
prompt for a password(-p). Next, we want to change the session
tables. Type in this exact command when you get to the

mysql> prompt (note capitalization areas and DO NOT include the ""s):

"ALTER TABLE sessions CHANGE a_session a_session LONGBLOB;"



Shawn Smith wrote:
> Hi, I've installed RT 3.8 on an Ubuntu Hardy server from the Debian
> package at
>
>
http://svn.debian.org/wsvn/pkg-request-tracker/packages/request-tracker3.8/trunk/debian/
>
>
> It works totally fine aside from the fact that it's got a
significant
> visual bug.  Here's how it's produced:
>
> 1. Click on queue in the Quick Search menu on the right hand side.
> 2. Click on a ticket in that queue.
> 3. Go back to the list of tickets.
> 4. Refresh the page.
>
> The error that occurs is that nothing gets rendered properly, so
it's
> just a white background with a bunch of blue links for the
tickets and
> text for the different fields of each ticket.  In
> /var/log/apache2/error.log I see stuff like:
>
> [Tue Mar 10 13:38:48 2009] [notice] child pid 24484 exit signal
> Segmentation fault (11)
> Out of memory!
> Callback called exit.
>
> and:
>
> [Tue Mar 10 18:02:36 2009] [crit]: Apache2::RequestIO::print: (104)
> Connection reset by peer at
> /usr/share/perl5/HTML/Mason/ApacheHandler.pm line 1034
> (/usr/share/request-tracker3.8/libexec/webmux.pl:150)
>
>
> My boss suggested changing MaxRequestsPerChild in
> /etc/apache2/apache2.conf to a low number other than 0, but that
still
> did not fix the error.  If anyone has any suggestions I would
greatly
> appreciate any advice you can provide.
>
> --
> Shawn
>
------------------------------------------------------------------------
>
> _______________________________________________
> http://lists.bestpractical.com/cgi-bin/mailman/listinfo/rt-users
>
> Community help: http://wiki.bestpractical.com
> Commercial support: sales@bestpractical.com
<mailto:sales@bestpractical.com>
>
>
> Discover RT's hidden secrets with RT Essentials from O'Reilly Media.
> Buy a copy at http://rtbook.bestpractical.com


--
Steve Christou
UWM Information Security
Phone: 414-229-1100
Email: Osa-list@uwm.edu <mailto:Osa-list@uwm.edu>


Shawn

Steve Christou
UWM Information Security
Phone: 414-229-1100
Email: Osa-list@uwm.edu

Yes I seem to be still getting the second error. There’s also a new error
that’s occurring:

[Tue Mar 10 18:36:25 2009] [crit]: Apache2::RequestIO::print: (104)
Connection reset by peer at /usr/share/perl5/HTML/Mason/ApacheHandler.pm
line 1034 (/usr/share/request-tracker3.8/libexec/webmux.pl:150)[Tue Mar 10
18:50:32 2009] [error]: Couldn’t get principal for not loaded object
(/usr/share/request-tracker3.8/lib/RT/User_Overlay.pm:1113)
[Tue Mar 10 18:50:32 2009] [error]: Group::HasMember was called with an
argument that isn’t an RT::Principal or id. It’s (undefined)
(/usr/share/request-tracker3.8/lib/RT/Group_Overlay.pm:1031)
[Tue Mar 10 18:52:20 2009] [error]: Couldn’t get principal for not loaded
object (/usr/share/request-tracker3.8/lib/RT/User_Overlay.pm:1113)

The first one in that list seems to be one of the originals that I posted.

Also, Tom: Yes, binary attachments seem to be messed up. I tried opening an
Excel spreadsheet on our current, older instance of RT and it worked, but on
my newer instance it did not. I thought I went through the UPGRADING.mysql
file when I originally migrated the database, but I will peruse it to see if
there’s anything that stands out.

Shawn Smith wrote:

Also, Tom: Yes, binary attachments seem to be messed up. I tried opening an
Excel spreadsheet on our current, older instance of RT and it worked, but on
my newer instance it did not. I thought I went through the UPGRADING.mysql
file when I originally migrated the database, but I will peruse it to see if
there’s anything that stands out.

Is it possible you did the: perl etc/upgrade/upgrade-mysql-schema.pl db user
pass > sql.queries

but then didn’t run the queries in sql.queries? All that command does is
generate a batch of SQL commands to run; it doesn’t run them for you. You
then need to do:

mysql -u root -p rt3 < sql.queries

To actually run the SQL queries to upgrade your database.

– ============================
Tom Lahti
BIT Statement LLC

(425)251-0833 x 117
http://www.bitstatement.net/
– ============================

I found from documentation that I wrote up on the process of migrating stuff
that this is what I did, is this correct, or is there more that needs to be
done?

for x in /etc/request-tracker3.8/upgrade/*; do
for y in schema acl insert; do
/usr/sbin/rt-setup-database-3.8 --action $y --datadir $x
–dba rt_user --dba-password [password];
done;
done;

From UPGRADING.mysql:
If you’re installing a new RT then you can skip this file.

If you’re migrating from MySQL 4.0 to MySQL 4.1 and newer then you MUST
follow
instructions at the bottom of this file.

If you’re upgrading RT from versions prior to 3.8.0 then you MUST follow
instructions below even if your old RT was installed on MySQL 4.1 or newer.

=== Upgrading RT from versions prior to 3.8.0 ===

  1. Backup RT’s database. Test that you can restore from this backup.

  2. Follow instructions in the README file to step 7.

  3. Apply changes described in step 7, but only up to version 3.8.0.

  4. Apply the MySQL 4.0->4.1 schema changes. Included in RT is the script
    etc/upgrade/upgrade-mysql-schema.pl that generates SQL queries to
    upgrade the database’s schema. Run it:

    perl etc/upgrade/upgrade-mysql-schema.pl db user pass > sql.queries

  5. Check sanity of the SQL queries yourself or consult with your DBA.

  6. Apply the queries. Note that this step can take a while. It may require
    additional space on your hard drive comparable with size of your tables.

    mysql -u root -p rt3 < sql.queries

NOTE that ‘rt3’ is the default name of the RT database, change it in the
command above if you’re using a different name.

This step should not produce any errors or warnings. If you see any, restore
your database from the backup you made at step 1) and send a report to the
rt-users@lists.bestpractical.com mailing list.

  1. Continue from step 7 in the README and apply other upgrade scripts
    and follow
    other steps.

Shawn Smith wrote:

I found from documentation that I wrote up on the process of migrating
stuff that this is what I did, is this correct, or is there more that
needs to be done?

for x in /etc/request-tracker3.8/upgrade/*; do \

 for y in schema acl insert; do \
    /usr/sbin/rt-setup-database-3.8 --action $y --datadir $x \
      --dba rt_user --dba-password [password]; \

 done;\

done;



The rt-users Archives

Community help: http://wiki.bestpractical.com
Commercial support: sales@bestpractical.com

Discover RT’s hidden secrets with RT Essentials from O’Reilly Media.
Buy a copy at http://rtbook.bestpractical.com

Drew Barnes
Applications Analyst
Network Resources Department
Raymond Walters College
University of Cincinnati

I found from documentation that I wrote up on the process of migrating stuff
that this is what I did, is this correct, or is there more that needs to be
done?

for x in /etc/request-tracker3.8/upgrade/*; do
for y in schema acl insert; do
/usr/sbin/rt-setup-database-3.8 --action $y --datadir $x
–dba rt_user --dba-password [password];
done;
done;

More. In particular, steps 4, 5 and 6 in UPGRADING.mysql.

– ============================
Tom Lahti
BIT Statement LLC

(425)251-0833 x 117
http://www.bitstatement.net/
– ============================

Applying those queries fixed the issue where the attachments wouldn’t load
properly, so thanks everyone for your help. Everything works fine it seems,
although I do get these errors:

[Tue Mar 10 20:23:07 2009] [crit]: Apache2::RequestIO::rflush: (104)
Connection reset by peer at /usr/share/perl5/HTML/Mason/ApacheHandler.pm
line 1035 (/usr/share/request-tracker3.8/libexec/webmux.pl:150)
[Tue Mar 10 20:24:04 2009] [error]: Couldn’t get principal for not loaded
object (/usr/share/request-tracker3.8/lib/RT/User_Overlay.pm:1113)
[Tue Mar 10 20:24:04 2009] [error]: Group::HasMember was called with an
argument that isn’t an RT::Principal or id. It’s (undefined)
(/usr/share/request-tracker3.8/lib/RT/Group_Overlay.pm:1031)

They are seemingly random in occurrence; originally I thought that they
happened after I opened a .jpg attachment, but then I opened it again and
could not reproduce the error.

Again, thanks everyone for your help with fixing the database, I really
appreciate it.