Perhaps you didn't reload apache? With the script alias you have below,
http://172.16.0.1/rt should not be working. Look at the apache error logs
- that will give some clues as to what's going on.
James Moseley
ty chan chanty...@yahoo.com wrote:
Hi Moseley,
Here is my configuration on
I see a database upgrade for postgres, but nothing for mysql going from
3.8.2 to 3.8.3. Perhaps I'm missing something...
James Moseley
Kevin Falcone falc...@bestpractical.com wrote:
Yes, there is a database upgrade involved.
___
FWIW, I do not see this behavior using RT 3.8.3 and FireFox 3.0.10.
James Moseley
Jo Rhett jrh...@netconsonance.com wrote:
It only happens when you click Show for outbound messages - ie, to
AdminCcs or Requestors. Normal ticket display is unaffected.
Everyone has their own way of doing things, but I find upgrading RT is
easiest by essentially installing RT from scratch upon each upgrade and
then soft-linking to the new directory. In this way, upgrades are
straightforward and not hard to follow. For example:
download new RT source
Did you apply the schema updates as indicated in steps 4-6 of
UPGRADING.mysql?
James Moseley
Dominic Lepiane dominic.lepi...@ptgrey.com wrote:
Dear RT users,
I am trying to get our RT installation moved from RT 3.6.6 to 3.8.3.
I've tried following the upgrade steps in the UPGRADING.mysql and
Hmmm, usually the cause of binary attachments not displaying properly is
because folks didn't apply the 'mysql 4.1' update...
James Moseley
Dominic Lepiane dominic.lepi...@ptgrey.com wrote:
I did run steps 4-6 from the UPGRADING.mysql , yes. I've included the
generated SQL below. I should
Kevin, I'm not seeing that behavior. I've upgraded to 3.8.3 and when using
the RichText editor, I still get /p generated when hitting 'enter'.
Basically, everything is double-spaced. Nor can I find a way to change
that behavior in preferences, in the settings when editing an individual
ticket,
Kevin, clearing the browser cache worked. I now see single spacing using
both IE8 and Firefox 3.0.10 when editing via the RT editor.
Thanks!
James Moseley
Kevin Falcone falc...@bestpractical.com wrote:
Sounds like your upgrade didn't take or you have cached JS
I see the behavior I
You need to uncomment the LoadModule line and add:
IfModule !mod_fastcgi.c
AddHandler fcgid-script fcg fcgi fpl
/IfModule
This assumes you have installed mod_fcgid from CentOS's yum respository and
are not trying to use mod_perl or mod_fastcgi.
James Moseley
What's the specific problem with putting a password in a crontab or a shell
script that cron calls? Only folks with root access could see the
password. Besides, you've got to put the RT database password in plain
text in your RT_SiteConfig file, unless you are running without a password,
which
Asif, in your email today, your syntax did not contain what I wrote below
yesterday. Furermore, with this syntax, as said before, you sould not have
SetLogo statements in your RT_SiteConfig.pm config. This is only one way
to do it - I am sure there are others. I would search the wiki and the
Perhaps folks were thinking you'd look in the Wiki first:
http://wiki.bestpractical.com/view/ChangeLogo
You will need to pay attention to the 3.5+ section regarding copying the
Logo file to RT_HOME/local/html/Elements/Logo
If you choose that method, you need not put all the SetLogo stuff in
We don't seem to have any problems using a 200x90 logo in our 3.8.2
install.
We have simply done the following:
Create a logo file: path to RT/local/htmlElements/Logo
With:
div id=logo
a href=insert your URL hereimg
src=%$RT::WebImagesURL%/your-image.jpg alt=%loc(Your company
name)%/a
%
Hi Kim, if I were you I would simply reinstall. Unless you had any
customizations, should be a piece of cake - the only thing you'll need is a
copy of your RT_SiteConfig.pm file.
If you were running an older version of RT, you'll need to follow the steps
to upgrade the RT database.
James
The scrips are all located in the database. Most other RT source
customizations would be located in the path-to-rt/local directory.
James Moseley
Kimberly McKinnis
Yes, I got both messages... Moving on.
Next step is setting up the mysql database.
3. If you already haven't done so, set a root password for your mysql
server - from a root prompt do (make sure mysqld is running):
mysqladmin -u root password root_password
(don't include the in the actual
Unless you are using a browser on the server itself, that isn't going to
work. More than likely you are using a remote browser.
In that case, use the IP address of the server, or better yet the FQ host
name:
http://10.10.10.10/rt/or http://server.domain.com/rt/
Obviously for the latter to
Grant, again, if you want help, do not post requests or replies directly to
individual members of the RT list without including the list...
The reason you are getting this error is because at some point in the email
chain, a linewrap inserted a carriage return into the 'configure' command
and you
First and foremost, please email your requests for assistance to the
mailing list. With that said, the main installation guide is Unix-generic,
but if you know enough about your particular OS, you shouldn't have too
many problems. Cent OS/RHEL (and other) guides do exist in the official RT
setup
When you moved the RT instance to another server, did you upgrade it? If
so, did you upgrade the database as well? If you moved the database, how
exactly did you do that? What version are you running now?
I'm running 3.8.2, and all table names are capitalized (with the exception
of one) , so
FYI, as my co-worker just pointed out, the FM* tables below are due to a
very old installation of RTFM... Those can be disregarded.
James Moseley
+-+
| Tables_in_rt3 |
+-+
| ACL |
| Attachments |
|
The first line of the error log tells you everything. You need version 18,
but you only have version installed. This problem is likely due to the
perl upgrade that happened as a result of the system upgrade.
You can use CPAN to update the perl module, or you can install it via yum:
yum upgrade
Did you update DNS? Did you update this part of the Apache config?:
Change this to Listen on specific IP addresses as shown below to
# prevent Apache from glomming onto all bound IP addresses (0.0.0.0)
# e.g. Listen 12.34.56.78:80
Listen ip address:port
James Moseley
Michael Mai
Asif, it's getting really hard to tell whose writing what with the way you
are quoting/replying to emails. It sounds like you are trying to not only
upgrade RT but also move your database from mysql 4.x.x to 5.x.x.
I would start by getting your current RT database dumped and successfully
moved
Our users have certainly noticed this bug, so it affects us to some degree.
In order to submit this as a bug, create an RT ticket here:
http://rt3.fsck.com/?user=guestpass=guest
James Moseley
Sean
Some mail environments are even more strict and require that the rDNS entry
of the IP address match with the corresponding forward DNS record.
With that said, I think it's reasonable that an owner of a mail server
sending legit email make sure that the IP address of their mail server have
an rDNS
For what it's worth, an end-user of mine noticed the same thing today...
He's also using IE 7.
James Moseley
Michael Finn
mf...@nbutexas.c
Yes, any time you make a change to RT_SiteConfig.pm, it's necessary to
restart apache and clear out the mason cache. Reference the readme docs.
James Moseley
Gary Casterline
Thanks. Good to know Jesse.
James Moseley
Jesse Vincent
je...@bestpracti
cal.com
Is mysql setup to listen to TCP connections via localhost and/or it's
'real' IP address?
James Moseley
Jerrad Pierce
jpie...@cambridg
Sorry, didn't realize RT had the capability to use local Unix sockets...
Does the owner of the httpd process have access to the mysql socket?
Assuming nothing changed within the RT directory, it might be helpful for
you to post the mysql you used to tune the database and the syntax used to
back
Report the results of a diff between the old version of my.cnf and the one
after your changes.
James Moseley
Jerrad Pierce
jpie...@cambridg
OK, and your certain my.cnf is setup so that this is the correct socket:
/var/lib/mysql/mysql.sock?
And the httpd process (assuming Apache here) has proper permissions to
mysql.sock?
James Moseley
The easiest thing for you to do is have the entire site in SSL. Is there a
particular reason you only want the login credentials passed using secure
sockets layer?
James Moseley
___
http://lists.bestpractical.com/cgi-bin/mailman/listinfo/rt-users
Why not just drop the database and then readd it using the
rt-setup-database script?
James Moseley
Joseph Spenner
joseph85...@yaho
Read the README/upgrade docs. The rt-setup-database does exactly that,
setup the database - it can also upgrade the database as well.
You can always look at the scripts themselves to see what they do.
James Moseley
What type of RAID system are you using and how fast are the disks?
James Moseley
Mathew
mathew.sny...@gm
On Fri, Jan 16, 2009 at 02:37:30PM -0500, Mathew wrote:
We presently have our RT installation running on the same hardware as
the database: an Intel 1550 box with 4 cores of about 2.5GHz each and 8GB
RAM. We've been plagued with speed issues even after upgrading to this.
We realized initial
Jesse Vincent je...@bestpractical.com wrote:
At 8 gigs of RAM on a well-tuned system, most of what RT is pulling out
of the database should always be cached in memory. If MySQL is going to
disk on every query, the game's over and you're better off sobbing
quietly into a stiff drink than
Fair enough.
James Moseley
Jesse Vincent
je...@bestpracti
cal.com
Andy Smith a.sm...@ukgrid.net wrote:
Hi All,
OK I think I have a solution to my problem which I hope results in
a correctly upgrade database. The last piece of the puzzle, is what
are the content files and what do you need to do with them? Well I
guessed that you are meant to do this:
Andy Smith a.sm...@ukgrid.net wrote:
Hi James,
ok, well thats nice I managed to get to the instructions for 3.8
using a bit of guesswork :P.
The only reason Im on 3.8.1 is that it is the current ports version in
FreeBSD. If its worth waiting for 3.8.2 I'm happy to wait, Im not in
any
In the next release, could this script be made host-aware? The script
currently assumes the db server is on the same host as the RT installation.
I know, it's more of a nuisance than anything, but perhaps it could be
built to use the login/host credentials in RT_SiteConfig.pm, just like the
Jesse Vincent je...@bestpractical.com wrote:
On Wed, Jan 07, 2009 at 06:19:20PM -0600, jmose...@corp.xanadoo.com wrote:
So, in the setup-database script, do you tell it to stop at 3.8.0, or
3.7.87? Meaning, if I choose 3.8.0, does the script stop after
installing
the 3.7.87 schema
To all,
I sent this before and got no response. Is it showing?
We are planning to move up to 3.8.2 soon (a month?). One of
the things
my users do not like is that there are many CF's that are sowing up in
their Modify Ticket screen that some use, some don't ,but needs to
Did you complete step four in UPGRADING.mysql?
James Moseley
Narendra Mehta
narendra.me...@s
Jesse Vincent
je...@bestpracti
cal.com
Why not just install 3.8.1 from source on a test server (it's really not
that hard and installing perl modules via yum or CPAN is not that complex),
make copy of your database, and then upgrade the schema on that test
database?
That's the only way you can adequately do a POC.
James Moseley
Jessie, I guess it just depends on the size of the organization, how many
queues they have, and how many open tickets are open at a given time. In
our environment, we have over 30,000 resolved tickets while we usually have
no more than a few dozen open in any given queue. It makes much more
You've installed RT from a Fedora package and are now trying to upgrade
from source. You are going to run into all sorts of problems...
Just install RT 3.8.1 from source as a new install in a DIFFERENT
directory. Run configure, make testdeps, etc.
Then run the upgrade database steps, etc.
First, make sure you cc the RT users group, not just respond directly to
the poster...
If this is a new install, uninstall the Fedora package then install RT from
source.
The first step is correctly run the 'configure' operation:
./configure --prefix=/path/to/rt --with-web-user=apache
Perhaps the script itself is poorly named, but the instructions state that
the 4.0 mysql upgrade script must be run if you are upgrading as it changes
the DB schema.
It's not a script only for those running 4.x.x versions of mysql, it's for
everyone that's upgrading.
James Moseley
I suspect our html-cleaning stuff is stripping out the color on
display. That could probably get relaxed if people think it's a good
idea.
Thanks, Jessie. To me personally, the value of editing in color doesn't
make much sense if others can't see those colors when looking at a ticket
via the GUI
1) Is email completion supposed to work out of the box?
Are you talking about when adding people via the 'People' section of
the ticket? If so, I've noticed that I used to be able to put
usernames in the Email field, but now that doesn't work.
You are correct. I'd love for this missing shortcut
Did you verify the contents of /opt/rt3/etc? The script output indicates
that the schema and acl files in that directory don't exist, which means
the preceding operation failed:
./install-sh -c -m 0755 -d /opt/rt3/etc
cd etc ./install-sh -c -m 0644 acl.Informix acl.Pg acl.Oracle
acl.mysql
So, if you are upgrading from 3.6.x to 3.8.0, you don't need to follow the
step below?
You may also need to update RT's database. To find out, type:
ls etc/upgrade
Also, which step that you refer to below fails? Fails for everyone?
James Moseley
Interesting. We're running 5.0.27 and the script worked just fine.
However, I ran all the schema/acl/insert upgrade commands using all the
5.7.xx directories in etc/upgrade.
Still waiting to hear from Ruslan to confirm 100% that you don't need to
run all these commands and you need only run the
Ruslan, thanks for sharing this.
James Moseley
Ruslan Zakirov
[EMAIL PROTECTED]
l.com
When replying or commenting on a ticket, I can see the text color, if color
is selected. When viewing the source, I can also see the span tags.
But after updating the ticket, the span tags disappear and thus the color
does too when displaying the ticket.
Am I missing something in the
Looking through the RT_Config.pm file, I've set this in my site config
file:
Set($PreferRichText, 1);
However, colored text still does not display in emails sent by RT nor in
the ticket display view - only while creating text while commenting,
replying, etc. Bold and italicized email displays,
Just installed RT 3.8.0 using a static copy of our RT (3.6.7) database.
Everything looks, well, different!
With that said, our personalized Logo is not displaying in the browser
window (works just fine in 3.6.7). Actually, the logo HTML code shows up
in the page source of the browser, but it
BTW, if I change the default theme to be 3.5 or 3.4, the logo displays.
I've actually added some code into the CSS files, and while the logo
displays, it looks terrible...
James Moseley
[EMAIL PROTECTED]
Guys, sorry for the quick alarm. We're going to have to mess with the CSS
layout to get the web2 layout to display our logo to our liking.
James Moseley
Scott Hebert
Changing:
Set($rtname, 'new.name');
Set($EmailSubjectTagRegex, qr/(?:old\.name|new\.name)/i);
Should be all you have to do.
James Moseley
Graeme Fowler
As far as I can tell, the 'sendmailpipe' reference in the RT config files
is simply an argument that tells the RT email functions how to call the
sendmail binary.
I think your problems will go away if you simply remove all your sendmail
configuration arguments from your RT_SiteConfig.pm file and
David, what kind of disks are you using on your DB server? Generally
speaking, mod_perl is slower than FastCGI - have you tried the latter?
Next, have you tried any manual queries against your DB server to see if
that's the source of the slowness?
For reference, I just displayed a ticket with
I would chalk it up to a temporary problem (bug) with Apache or the FastCGI
module...
James Moseley
[EMAIL PROTECTED]
.uk
http://wiki.bestpractical.com/view/ManualInstallation
James Moseley
Ruslan Zakirov
[EMAIL PROTECTED]
Have you read Ruslan's email yet? If you are running mod_perl 2.x, you
don't need: PerlModule Apache2::compat
I'd remove that first and then see if things work before worrying about
which version of mod_perl you have or switching to FastCGI.
But, to answer your questions:
More than likely, if
First, you have to add the user if they don't already exist. Next, you
have to make sure that user has appropriate permissions to the queue the
ticket is in.
James Moseley
John BORIS
It's really up to you how you want to handle how people are associated to a
ticket. With that said, typically the person 'opening' the ticket or
making the request is the requestor.
The owner of the ticket is the person that is working the ticket or that
the ticket is assigned to. You can very
This says it all:
The RT_* files are set as root:apache and the permission is set to 500
How is the Apache process (running as user 'apache', I assume) supposed to
access your RT configs if they have no permissions to do so? Chmod those
files to 640 or something similar - that should fix your
John, use the name of the new server (or whatever 172.31.6.213 resolves to)
in RT_SiteConfig. Also, any reason you are using a virtual host?
Lastly, I assume you are calling RT at the following URL:
http://172.31.6.213/rt/
James Moseley
You defintitely need to be calling it via http://172.31.6.213/rt/
As to your other errors, those are definitely problems. Did you install RT
via source or yum repository? Either way, have you done a 'make testdeps'
in the source directory to ensure all the Perl modules have been installed?
Have you checked the other config files in /etc/httpd/conf.d? Anyways,
that error is just informational and doesn't affect anything.
So the question is, does RT work now?
James Moseley
John BORIS
John - don't worry about the Apache warning about the module for now. You
also didn't make the necessary changes. You put some in, left others out,
then left the old mod_perl stuff in place. Also, you didn't adjust the
paths in the configs I sent you. Note that my path only had /opt/rt, not
Better yet, import the self-generated CA cert (that you used to sign each
of your SSL certs) into each browser as a trusted root certificate
authority and no more warnings...
James Moseley
You can generate your own SSL certs for free. The
only reason you'd want to buy them from Thawt
Danie, if you've got customers using RT and you want them to use SSL, then
pay up for a legitimate certificate.
James Moseley
There is a self-signed cert on the box. But we need to give Internet
access
to customers and do not want them to get a certificate warning. So they
must either
Compiling stuff on Solaris and making it work is always difficult at best.
James Moseley
mailing list
[EMAIL PROTECTED]
Reposting to see if some light can be shed. I am very curious...
Performing queries using 'Requestor = X OR Cc = X OR AdminCc =X' in 3.6.3
yielded very fast results - a few seconds. However, doing the same search
in 3.6.4 resulted in a timeout after four minutes.
Using 'Watcher = X' is the
OK - here you go:
mysql describe sessions;
+-+-+--+-+---+---+
| Field | Type| Null | Key | Default | Extra |
+-+-+--+-+---+---+
| id | varchar(32) | NO | PRI
You state starting apache still fails... I assume you did do the
following:
yum install mod_fcgid
Honestly, I would uninstall the RT3 RPM and install RT from source.
Documentation is fairly straightforward.
James Moseley
Choose the method you prefer, but I would at least try. Other than Apache
and either mod_perl or some flavor of a FastCGI module, your only
dependencies are Perl modules. Install the CPAN installer and use the
'make fixdeps' option within the RT source code to install what it can.
Then install
What's your complete apache RT config? Not the whole httpd.conf file, just
the stuff relative to RT.
James Moseley
mailing list
[EMAIL
A couple of things. I assume you have this somewhere in your apache
configs:
LoadModule fastcgi_module modules/mod_fastcgi.so
Apache probably wouldn't start up without loading this module in your
configuration, but...
Also, what are the following values in RT_SiteConfig.pm?
Set($WebPath)
In your previous post, you mentioned that sometimes you want to create a
particular user, but can't because this new users email address is already
taken by a user that got automatically created.
To fix that, you do exactly what you are doing - you manipulate/change the
address of the inactive
I am running RT 3.6.3 on two systems: An older Linux system using Apache 2
and Apache's mod_fastcgi module and an FC6 system with Apache 2 and
mod_fcgid (an alternative to the Apache mod_fastcgi module). Both
installations work just fine.
Prior to installing 4.6.4, I upgraded the few Perl
Try using a / at the end of your URL. For example:
http://rt.example.com/rt/
You might also want to post the output of the 'codes' you receive after
login.
James Moseley
Mendez, Felix
Those deferred messages should be in:
/var/spool/clientmqueue
Simply take a look at the files to make sure they are the emails in
question, then delete them. In other odd cases, they may have ended up in
one of the directories under /var/spool/postfix - but those directories are
used for
The documentation that comes with all RT releases outlines the steps
involved in modifying your database when RT is upgraded. Yes, you'll have
to run several scripts that will end up modifying your database in order
for it to be compatible with RT3.
James Moseley
OK. But obviously one or more of the patches you applied broke something -
I would take a look at each patch that was installed and compare them
against the dependency requirements for RT. Then only back out of those
patches. Backing out of Sun supplied patches (which I assume these are) is
Which Solaris patches were installed? Have you tried backing out of them?
James Moseley
Steve Holmes
[EMAIL PROTECTED]
I would look for perl or apache patches. Were these simply Solaris OS
patches for security, or was this a jumbo patch which includes updates for
other installed software. Problem with new versions of Solaris is that
they come with Apache, Sendmail, Perl (I believe) - I wouldn't install any
of
I am not sure Postfix knows how to deal with the RT mailgate code when used
inside virtual alias or virtual mailbox maps. I would simply create
entries in your alias file like this:
ck1: |/usr/sbin/rt-mailgate --queue general --action correspond --url
http://192.168.0.150/rt3/;
Also, what's
So, the mailgate code is in the aliases file? If so, I would remove all
references of the 'user' ck1 from your virtual maps, this will only confuse
things. Of course, this assumes that you are running Postfix on the same
box as RT...
Can you include your RT entries in the alias file in this
It appears that from behavior seen in the logs, email to that alias is
getting put through the virtual aliasing mechanism. Make sure you have the
following in main.cf:
alias_maps = hash:/etc/postfix/aliases
If adding the above fixes the problem, great. If not, add:
local_recipient_maps
Torben, without doing the research, I am not even sure if the RT code knows
how to parse html email. I think you'd be better off just sending plain
text email through the mailgate. I am sure this topic has been covered
before somewhere in the list archives.
I think perhaps your lack of
You need to make sure the proper scripts have been added either to the
global or queue-specific 'scrips' list. It's been a long time since I
setup RT from scratch, but I believe by default, CC's are not emailed when
making a comment. Either way, you need to make sure that you have 'On
Comment'
You need to make a list of the files you have changed so that you can
update those files with your changes in any new installation.
James Moseley
joey
I'd suggest searching the list archives as others have probably had similar
questions. With that said, you would need to upgrade the schema on the 2nd
database as well. Then do a database dump on the appropriate tables.
Since you'll probably have duplicate ticket numbers, you would need to
1 - 100 of 107 matches
Mail list logo
