One of the initial mistakes I made was to try and run the installation from a network mount. DON’T. Solaris must block the execution of code on some types of removable drives, which in this case was a VMWare Fusion shared folder.

A second thing that is not clear in the README files: Clearcase 7.1.1 on Solaris x86 does NOT support the GUI installation. Therefore, for me as much as any readers of this blog, I’ll document the steps to kick off a fresh/silent installation.

The install will be/have:

• A local install
• Clearcase Atria Licencing, local
• IBM Installation Manager installed first
• ClearCase installed second

1. Download your entitled product for Solaris X86, for CC7.1.1 which is:
   • CZ9XKML
   • IBM Rational ClearCase V7.1.1 Solaris x86 Platform Edition Multilingual

   If you need an evaluation copy, they are also available.

2. Copy the archive to your server, and extract it. I’ll use /var/ccinstall for my install source, but adjust as required:
   

   
   mkdir -p /var/ccinstall
   cp -p CZ9XKML.zip /var/ccinstall
   cd /var/ccinstall
   unzip CZ9XKML.zip
   cd disk1
   

3. Follow the IBM documented steps for preparing a silent install. For my steps the dir /var/ccinstall/disk1 will be esd_image_root.
4. Kick off the install:
   

   
   /var/ccinstall/disk1/InstallerImage_*/install \
   --launcher.ini \
   /var/ccinstall/disk1/InstallerImage_*/silent-install.ini \
   -silent \
   -input clearcase_response_sol_x86.xml \
   -log silentinstall.log
   

   
   If you want to also show the progress, include the switch -showVerboseProgress

At this point kick back and make a few coffees.
Enjoy.

In my case I am Using Solaris 9, and have Oracle 10g databases and client files for use by BOE.

When running the cmsdbsetup.sh script you get the following error pertaining to clntsh:

           Business Objects 

Current CMS Data Source: DBNAME 

err: Error: Failed to get cluster name. 
err: Error description: Unable to load clntsh 

select (Select a Data Source) 
reinitialize (Recreate the current Data Source) 
copy (Copy data from another Data Source) 
changecluster (Change current cluster name) 
selectaudit (Select an Auditing Data Source) 

[select(6)/reinitialize(5)/copy(4)/changecluster(3)/selectaudit(2)/back(1)/quit(0)] 
----------------------------------------------------------

This error “Unable to load clntsh” refers to the libclntsh.so library used by the Oracle client. Since BOE runs as 32bit, the 32bit Oracle client libraries should be accessible by the user running BOE.

If you are running a 64 bit Unix and a 64bit Oracle install check that the environment for the user running BOE (user that will run the CMS) has the 32bit libraries in the path:
LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$ORACLE_HOME/lib:$ORACLE_HOME/lib32

Then check that either the user is a member of the Oracle dba Unix group or everyone has permissions to access the 32bit libraries under Oracle 10g:

su - oracle
chmod o+rx $ORACLE_HOME/lib32/*

Feel free to leave any comments if you need help with this.

First off install some require packages, including the zziplib package:

  export PKG_PATH=http://mirror.aarnet.edu.au/pub/OpenBSD/4.5/packages/i386/
  pkg_add -v zziplib
  pkg_add -v autoconf-2.62

  mkdir /usr/local/src/
  cd /usr/local/src/
  wget http://pecl.php.net/get/zip
  tar zxvf zip
  cd zip-1.10.2/ 

  export AUTOCONF_VERSION=2.62
  phpize
  ./configure
  make
  make install

  cat << EOF >> /var/www/conf/php.ini
  extension=zip.so
  EOF
  sudo apachectl stop
  sudo apachectl start

Of course, this will go stale over time as new releases and versions come out, so YMMV.

Have you ever found that a filesystem is filling up fast, and dont know what is causing it? This one liner (which can be placed in a cron job if you like) is best run as a super user.

It will:

1. Search for all files in the current dir and subdirs that are modified in the last 3 days;
2. list them, filtering just the filesize and name/path;
3. sort/order by the largest file; and
4. Email you a copy of the report.

find / -type f -mtime -3 -exec ls -l {} \; 2>/dev/null \
| awk '{print $5 " " $9}' \
| awk '{printf "%12d\t%s\n", $1, substr($0,index($0,$2),80)}' \
| sort -r >/tmp/largefiles.txt && \
( uuencode /tmp/largefiles.txt largefiles.txt ) \
| mailx -s 'Large Files Report' -r mail2@youremail.com `cat /tmp/mailrecipients` &

Nice. If you just want to see the response on stdout then try this:

find / -type f -mtime -3 -exec ls -l {} \; 2>/dev/null \
| awk '{print $5 " " $9}' \
| awk '{printf "%12d\t%s\n", $1, substr($0,index($0,$2),80)}' \
| sort -r

Have fun.

How does this all relate to readline under Leopard?

The problem is the Intel Core2 Duo is a 64bit CPU. Apple supplied apps like Apache2 are compiled as Universal binary AND for 32/64bits. This means there are 4 architectures in the binaries.

lantrix@lexx:~ $ file /usr/sbin/httpd
/usr/sbin/httpd: Mach-O fat file with 4 architectures

This is how and why I arrived at compiling readline on my Mac 64bit intel under Leopard.

On to the readline problems. I start the compile of readline as x86_64 architecture, and this occurs:

gcc -dynamic -arch x86_64 -bind_at_load -arch_only `/usr/bin/arch` -install_name /usr/local/lib/libreadline.5.2.dylib -current_version 5.2 -compatibility_version 5 -v -o libreadline.5.2.dylib readline.so vi_mode.so funmap.so keymaps.so parens.so search.so rltty.so complete.so bind.so isearch.so display.so signals.so util.so kill.so undo.so macro.so input.so callback.so terminal.so text.so nls.so misc.so xmalloc.so history.so histexpand.so histfile.so histsearch.so shell.so mbutil.so tilde.so compat.so -lncurses
Using built-in specs.
Target: i686-apple-darwin9
Configured with: /var/tmp/gcc/gcc-5465~16/src/configure --disable-checking -enable-werror --prefix=/usr --mandir=/share/man --enable-languages=c,objc,c++,obj-c++ --program-transform-name=/^[cg][^.-]*$/s/$/-4.0/ --with-gxx-include-dir=/include/c++/4.0.0 --with-slibdir=/usr/lib --build=i686-apple-darwin9 --with-arch=apple --with-tune=generic --host=i686-apple-darwin9 --target=i686-apple-darwin9
Thread model: posix
gcc version 4.0.1 (Apple Inc. build 5465)
i686-apple-darwin9-gcc-4.0.1: -compatibility_version only allowed with -dynamiclib
make[1]: *** [libreadline.5.2.dylib] Error 1
make: [shared] Error 2 (ignored)

Google is your friend, and MacOSXHints forum user dmacks has an answer why this occurs.

The readline compile scripting is ugly (doesn’t use standard build tools, then fakes it…poorly at times). You need -dynamiclib not -dynamic to compile a shared library on OS X, but this indicates a deeper problem: need to patch support/shobj-conf to know darwin9* as a Darwin/MacOS system if you’re on Leopard (I think readline5.2 only recognizes darwin8*, i.e., Tiger).

Still this is not a fix…. But some further searching and I find a Christopher Rasch on a Watir mailing list who shows how to get around this. Effectivly you need to change the makefile for readlines shlibs to use the -dynamiclib switch when compiling.

tar zxvf readline-5.2.tar.gz 
cd readline-5.2
export MACOSX_DEPLOYMENT_TARGET=10.5
export CFLAGS="-arch x86_64 -g -Os -pipe -no-cpp-precomp"
export CCFLAGS="-arch x86_64 -g -Os -pipe"
export CXXFLAGS="-arch x86_64 -g -Os -pipe"
export LDFLAGS="-arch x86_64 -bind_at_load"
./configure
cd shlib
sed -e 's/-dynamic/-dynamiclib/' Makefile > Makefile.good
mv Makefile.good Makefile
cd ..
make && sudo make install

This gave me a local x86_64 architecure readline library under Leopard. I was able to successfully compile PostgreSQL as -arch x86_64 against the readline library as well.

lantrix@lexx:/usr/local/src/readline-5.2 $ ls -l /usr/local/lib/libreadline.*
-r-xr-xr-x  1 root  lantrix  272480  3 Jan 02:06 /usr/local/lib/libreadline.5.2.dylib
lrwxr-xr-x  1 root  lantrix      21  3 Jan 02:06 /usr/local/lib/libreadline.5.dylib -> libreadline.5.2.dylib
-rw-r--r--  1 root  lantrix  934864  3 Jan 02:06 /usr/local/lib/libreadline.a
lrwxr-xr-x  1 root  lantrix      21  3 Jan 02:06 /usr/local/lib/libreadline.dylib -> libreadline.5.2.dylib

Do what you will with them

I was trying to get OpenBSD bootstrapped using the PXEBoot NIC in a server. I decided to use my Mac which has tftp and bootpd installed. This post is quite in depth and technical so if you are game then read on.

The quick and short of it was I turned it all on, and copied my pxeboot image, like this from the Terminal:

sudo apachectl start
sudo service tftp start
sudo cp /Library/WebServer/Documents/pub/OpenBSD/4.4/i386/pxeboot  /private/tftpboot/

Then I neeed to setup the bootp server which comes with Internet Sharing. The idea was to add the pxeboot filename needed by OpenBSD (the file copied to tftpboot above) to the bootp (AKA dhcp) server options. The important file here is /etc/bootpd.plist. If this file doesn’t exist when Internet Sharing starts then bootpd will create it, removing it when it stops. But courtesy of Jules.FM “if the file already exists when it starts, the Mac will leave it alone and not overwrite or remove it”. So to add new dhcp options you perform these steps:

1. Start Internet Sharing
2. Copy the file somewhere safe: “cp /etc/bootpd.plist /tmp/”
3. Stop Internet Sharing
4. Edit /tmp/bootpd.plist
5. Add your required options
6. Copy the file back in place: “sudo cp /tmp/bootpd.plist /etc/”
7. Start Internet Sharing

Since the Mac bootpd.plist file has no option for “filename” documented and since there appears to be a bug in the bootpd implementation with respect to supplying dhcp_options I added these data options to the bootpd.plist file:

<key>dhcp_option_66</key>
<data>
wKgCAQ==
</data>
<key>dhcp_option_67</key>
<data>
cHhlYm9vdA==
</data>
<key>dhcp_option_93</key>
<data>
AA==
</data>

In RFC2132 (Paragraphs 9.4 and 9.5) it specifies that options 66 and 67 are for the tftp server and boot filename. However the bootpd bug meant I had to encode the string “pxeboot” to hexadecimal “0x707865626F6F74” then Base64 encode it being “cHhlYm9vdA==” and that is way too time consuming to stuff around and try something else everytime the tftp file retrieval fails.

Don’t try using the dhcp_option_66 I had either as it is an encoded IP of my tftp server, not yours.

I got it all going, and the MacBookPro worked as planned, but the bootp server offered a filename of “pxeboot” and the silly Intel 10/100 card decided to fail due to missing files on the tftp server. Many hours later and I figured I should use a network sniffer instead of trying to turn on tftpd logging.

A packet dump showed \377 or 0xff being appended to the filename by the boot client (intel Nic on remote server) when using tftp to request the file. This made the filename “pxeboot” look like “pxeboot\377” in Wireshark’s view of the packet, as discussed here by other pioneers in netbooting.

If I get a full working implementation of dhcpd/tftp/pxe using the Mac – and actually working independent of the intel nic problem on the server, I’ll update this entry or write a complete run down.

Don’t hold your breath though. I rebuilt the OpenBSD server using bsd.rd (being a ramdisk installation) so my need is no longer a driver.

OpenBSD by default apache runs in a chroot jail, thereby making it more secure in case the www server is compromised. I have talked about this before.

However if you use wordpress or some other web application that needs MySQL (and I’m talking where apache and MySQL reside on the same host), then the chroot jail will not allow your webapp access to the mysql socket file which is by default located in /var/run/mysql/

There are various solutions available online to wait x number of seconds and then to create symlinks, etc – but this was not always working for me – with the result the server would sometimes be up without database availability.

Here is my solution (thanks to NoMoa for the idea) – very simple, no postfix symlink but there is a symlink to allow other applications to access MySQL locally via the default socket location.

First I create the run dirs in the apache jail:

mkdir -p /var/www/var/run/mysql
chown _mysql:_mysql /var/www/var/run/mysql
chmod 711 /var/www/var/run/mysql

Then I add this to the bottom of /etc/rc.local on the webhost.

#mySQL
if [ X"${mysql}" == X"YES" -a -x /usr/local/bin/mysqld_safe ]; then
        echo -n " mysqld"; /usr/local/bin/mysqld_safe --user=_mysql --log --open-files-limit=256 --socket=/var/www/var/run/mysql/mysql.sock &
        if [ ! -L /var/run/mysql/mysql.sock ]; then
                mkdir -p /var/run/mysql
                chown _mysql:_mysql /var/run/mysql
                chmod 711 /var/run/mysql
                ln -fs /var/www/var/run/mysql/mysql.sock /var/run/mysql/mysql.sock
        fi
fi

mysql=YES

Note: you may need to change the open-files-limit to suit, but the above paths are as per a standard OpenBSD 4.3 install using the supplied MySQL pkgs.

I tested this all out on OpenBSD 4.3 and it now works fine for me. Whenever the server comes up I never get the dreaded word press DB failure screen. Good luck

• If you use vertical splits, this will help move left and right across the split. Put in your ~/.vimrc
  

  " Map multi window keys
  set wmw=0
  " CTRL-H move to left window
  nmap <c -h> </c><c -w>h</c><c -w><bar>
  " CTRL-L move to right window
  nmap <c -l> </c><c -w>l</c><c -w><bar>
  </bar></c></bar></c>

• When scrolling up and down a window, you can use zz to jump the current line to the middle of the window. If you want this on always ala Scroll locking, then you can use this function. It is a toggle option. use \zz to toggle it. Put in your ~/.vimrc
  

  " Map \zz to lock scroll to middle of window
  map <leader>zz :let &scrolloff=999-&scrolloff<cr>
  </cr></leader>

• In a window split (of any sort) if you want to maximise to the current window, this will do it for you. When you press CTRL-W then o it will maximise the current view, then when pressed again will return your split arrangement! Put in your ~/.vimrc
  

  " Max/unmax splits
  nnoremap <c -W>O :call MaximizeToggle ()<cr>
  nnoremap <c -W>o :call MaximizeToggle ()<cr>
  nnoremap <c -W></c><c -O> :call MaximizeToggle ()<cr>
  
  function! MaximizeToggle()
    if exists("s:maximize_session")
      exec "source " . s:maximize_session
      call delete(s:maximize_session)
      unlet s:maximize_session
      let &hidden=s:maximize_hidden_save
      unlet s:maximize_hidden_save
    else
      let s:maximize_hidden_save = &hidden
      let s:maximize_session = tempname()
      set hidden
      exec "mksession! " . s:maximize_session
      only
    endif
  endfunction
  </cr></c></cr></c></cr></c>

In this case your private key is stored unprotected on your own computer, and anybody who gains access to that will be able to generate signatures (login to servers) as if they were you They will be able to log in to your server under your account.
I’ll reiterate that: This will allow ANYONE holding this file to access ANY server AS YOU where you have uploaded the public keys.

This means that in the case your laptop or computer is lost or stolen, your unix accounts are effectively compromised.

For this reason, your private key is recommended to be encrypted when it is stored on your local machine, using a pass phrase of your choice. To minimise this risk you should choose a strong pass phrase to be applied to the private key when generation occurs.

There are two ways to generate a key pair.

1. If you are using openssh then generate the keypair under your unix login as follows:
   

   $ ssh-keygen -C "My development key 05 May 2008" -t rsa
   Generating public/private rsa key pair.
   Enter file in which to save the key (/home/lantrix/.ssh/id_rsa): 
   Enter passphrase (empty for no passphrase): **type in a strong password here**
   Enter same passphrase again: **retype in your strong password here**
   Your identification has been saved in /home/lantrix/.ssh/id_rsa.
   Your public key has been saved in /home/lantrix/.ssh/id_rsa.pub.
   The key fingerprint is:
   1a:aa:bb:44:09:38:ec:1d:1c:2d:27:c8:cc:dd:ee:ff My development key 05 May 2008
   $ cat ~/.ssh/id_rsa.pub >> ~/.ssh/authorized_keys
   

   
   Then copy ~/.ssh/id_rsa the password protected and encrypted private key to a secure place on your client machine to use (placing it in your ~/.ssh/ folder – remember to set permissions to 600).

   If you want to use this openssh keypair with putty on a windows client, you will need to follow an extra step. Use PuttyGen menu to load your generated “id_rsa” file you transferred to your windows client.

2. If you are only going to use putty to connect to UNIX servers it is better to generate the keypair in putty
   See the putty documentation for instructions on generating your key pair.
   Ensure you export the public keyfile to ~/.ssh/authorized_keys on each UNIX server you want to login (and chmod 600 on the file).

A Helpful tip

You can use putty to “cache” your key (to prevent constant retyping of your password when logging into servers) in a secure fashion using this component of putty.

Dont think of SSH keypairs as a means of easier logins. When used correctly it will in fact provide a more secure login; as your password is never passed over the network.

During the initialisation of the new DB, I got the following error:

creating directory /var/postgresql/data/pg_tblspc ... ok
selecting default max_connections ... 10
selecting default shared_buffers ... 50
creating configuration files ... ok
creating template1 database in /var/postgresql/data/base/1 ... FATAL:  could not create semaphores: No space left on device
DETAIL:  Failed system call was semget(1, 17, 03600).

After some searching, I found the solution. It looks like the problem is that the semaphores from the old installation are still hanging around, and never got cleaned up. You can use ipcs and ipcrm to clean then up and run your initdb again.

First list the existing semaphores:

server:~# ipcs                                                                                                  
Message Queues:
T       ID     KEY        MODE       OWNER    GROUP

Shared Memory:
T       ID     KEY        MODE       OWNER    GROUP
m    65536    5432001 --rw------- _postgresql _postgresql

Semaphores:
T       ID     KEY        MODE       OWNER    GROUP
s  1376256          0 --rw-------      www      www
s  1245185          0 --rw-------      www      www
s    65538    5432001 --rw------- _postgresql _postgresql
s    65539    5432002 --rw------- _postgresql _postgresql
s    65540    5432003 --rw------- _postgresql _postgresql
s    65541       5419 --rw-r--r--      www      www

server:~# ipcrm -s 65538   
server:~# ipcrm -s 65539 
server:~# ipcrm -s 65540 

Semaphores:
T       ID     KEY        MODE       OWNER    GROUP
s  1376256          0 --rw-------      www      www
s  1245185          0 --rw-------      www      www
s    65541       5419 --rw-r--r--      www      www

server:/usr/local/share/postgresql# su - _postgresql 
#initdb                                                                                                         
The files belonging to this database system will be owned by user "_postgresql".
This user must also own the server process.

The database cluster will be initialized with locale C.

creating directory /var/postgresql/data ... ok
creating directory /var/postgresql/data/global ... ok
creating directory /var/postgresql/data/pg_xlog ... ok
creating directory /var/postgresql/data/pg_xlog/archive_status ... ok
creating directory /var/postgresql/data/pg_clog ... ok
creating directory /var/postgresql/data/pg_subtrans ... ok
creating directory /var/postgresql/data/pg_twophase ... ok
creating directory /var/postgresql/data/pg_multixact/members ... ok
creating directory /var/postgresql/data/pg_multixact/offsets ... ok
creating directory /var/postgresql/data/base ... ok
creating directory /var/postgresql/data/base/1 ... ok
creating directory /var/postgresql/data/pg_tblspc ... ok
selecting default max_connections ... 40
selecting default shared_buffers ... 1000
creating configuration files ... ok
creating template1 database in /var/postgresql/data/base/1 ... ok

<ifmodule mod_rewrite.c>
RewriteEngine On
#prevent slashdot effect
RewriteCond %{HTTP_USER_AGENT} !^CoralWebPrx
RewriteCond %{HTTP_REFERER} blogspot\.com [NC,OR]
RewriteCond %{HTTP_REFERER} reddit\.com [NC,OR]
RewriteCond %{HTTP_REFERER} digg\.com [NC,OR]
RewriteCond %{HTTP_REFERER} news\.slashdot\.org [NC,OR]
RewriteCond %{HTTP_REFERER} slashdot\.org
RewriteRule ^(.*)$ http://techdebug.com.nyud.net/$1 [R,L]
#Rewrite images to allow CDN to serve them
RewriteCond %{HTTP_USER_AGENT} !^CoralWebPrx
RewriteRule ^(.*)/(.*\.(gif|png|jpe?g))$ http://techdebug.com.nyud.net/$1/$2 [R,L]
#Wordpress rewites
RewriteBase /
RewriteCond %{REQUEST_FILENAME} !-f
RewriteCond %{REQUEST_FILENAME} !-d
RewriteRule . /index.php [L]
</ifmodule>

First write your message:

cat << EOF > /tmp/mailmsg
Hi this is a message
And this is the second line
EOF

Then populate your recipient list, comma delimited as per the mailx(1) man page:

cat << EOF > /tmp/mailrecipients
john.doe@nodomain.com.it,jack.black@someplace.co.za
EOF

. You need to uuencode your binary attachments, and you can send as many as you need.
(cat /tmp/mailmsg ; uuencode /path/to/file.txt file.txt ; uuencode /location/of/otherfile otherfile) | mailx -s 'Subject' -r myemail@some.place.mx `cat /tmp/mailrecipients`

You need to specify each file name twice, once for source file to encode, and once for the encoded file name; as per the uuencode(1C) man page. If you are sending from some local account on the server, the -r switch allows you to specify an alternate return address for the recipients (in other words your normal email address).

PS: watch for the quotes and backticks. Dont mix them up!

Hope this helps you out someday.

In this post, I will show you how to use a Perl script to enforce Clearcase labeling conventions. This example is directed toward Clearcase on UNIX (i.e. Solaris or similar) and assumes you have Perl installed, working and have a basic knowledge of how to program in Perl. It is a reworked version of the windows script supplied by IBM on Developerworks.

This is a long post, but a good one if you are a new clearcase admin who needs to enforce label names.

Lets get into it.

For this script we would like to allow only project defined labels. This list shows examples which we will aim to enforce:

• A user label: USER_ASMITH_NEW
• A defect label: DEFECT_12345
• A Project label: PROJECTA_PRERELEASE
• Another Project label: PROJECTB_FINAL
• A release label: REL_1.0
• A baseline label: BASELINE_STRATA

In all examples you actually want the users to define some part of the label, so for our script we will enforce only the first part up to and including the underscore; i.e. the part in italics ENFORCE_USERINFO. You can change this as needed for your project.

Initially you will need to write a perl script that can be used in a trigger event. The perl script will take the requested label (lbtype) from the clearcase label creation event, and then process that label for you. You can start the script be choosing some preferred label prefixes for your projects (as per our example list of labels above):

#
# Configuration Variables
#
my $USERPREFIX = 'USER';
my $DEFECTPREFIX = 'DEFECT';
my $PROJ1_PREFIX = 'PROJECTA';
my $PROJ2_PREFIX = 'PROJECTB';
my $REL_PREFIX = 'REL';
my $BASEPREFIX = 'BASELINE';

After you have your variables set, put in place a subroutine to gather the current Unix userID. This will prevent a user from creating user labels with other users account names in it. This subroutine will get called later when you break apart the label with regex.

sub checkUserName {
  my $user = shift(@_);

  my $currentUser = uc($ENV{CLEARCASE_USER});

 if ($user !~ m/^($currentUser)/) {
    `clearprompt proceed -mask proceed -type error -prompt \"Can't create label for this user\"`;
    exit 1;
  }
  exit 0;
}

Next you need to gather the actual lbtype requested from the operation. This script will be run as a pre-operation script, unlike tip number 02 which was a post operation script. The idea is to check the label BEFORE it is actually created.

{
  my $label = $ENV{CLEARCASE_LBTYPE};

  # check, if label is in uppercase format
  if ($label ne uc($label)) {
    `clearprompt proceed -mask proceed -type error -prompt \"CC Error: Label not in uppercase\"`;
    exit 1;
  }

It is good practice to have your labels all uppercase to prevent them from being confused with branch type names (which normally you would have all in lowercase).

Next you can breakdown the label using regex and use some conditional checks to compare the label against all the allowed prefixes. As long as the prefix is valid, we don’t care what comes after it. The one exception is the chosen user label convention. It will require the user prefix, followed by a username that must match the current unix login name.

  # decompose label
  if ($label =~ m/^($USERPREFIX)_(.*)_(.*)$/) {
    &checkUserName($2);
  } elsif ($label =~ m/^($DEFECTPREFIX)_(.*)$/) {
    exit 0;
  } elsif ($label =~ m/^($PROJ1_PREFIX)_(.*)$/) {
    exit 0;
  } elsif ($label =~ m/^($PROJ2_PREFIX)_(.*)$/) {
    exit 0;
  } elsif ($label =~ m/^($REL_PREFIX)_(.*)$/) {
    exit 0;
  } elsif ($label =~ m/^($BASEPREFIX)_(.*)$/) {
    exit 0;

Lastly should the new label not match any of your regex checks then the script will exit with a return code of one. This will abort the creation of the invalid label type.

  } else {
    # no valid prefix found
    `clearprompt proceed -mask proceed -type error -prompt \"CC Error: No valid Label specified - please read project guidelines\"`;
    exit 1;
  }
}

You can download the whole script if you are in a rush.

In this script you have seen a number of variables starting with $CLEARCASE. These variables will be populated as environment variables when the label creation trigger event occurs and launches the perl script.

The script should be saved somewhere it can be run on the UNIX server hosting the VOB, so when Clearcase triggers it can run the script. An appropriate location would be in the home directory of the VOB owner, for example:
/home/ccadmin/ccscripts/test_label_name.pl

One you have saved the script, you need to actually make the trigger in Clearcase. Consider the following on creating this trigger type:

• We need the trigger to run before the event completes, this is what is known as a “Pre Operation”
• We want the trigger to be enacted before a new type is made, of type labeltype.
• We want the trigger to execute our perl script on the event occurring

You can research these options by viewing the man page of the command used to make a new trigger type:
cleartool man mktrtype
Once you have read the man page, you should see some examples and options that can be used. Have a go at running the command yourself.

This is how I created the trigger to cover all our considerations:

cleartool mktrtype -c "Trigger to validate new labels on lbtype creation" -type -preop mktype \
-lbtype -all -exec /home/ccadmin/ccscripts/test_label_name.pl new_label_trigger

Once that has been run, you can see your new trigger listed in the trigger type list:

cleartool lstype -kind trtype
21-Jan.09:21   ccadmin     trigger type "new_label_trigger"
  "Trigger to validate new labels on lbtype creation"

Next time a user creates a label type, they will either have success if they follow the guidlines you have communicated to them, or they will get an error message explaining why the label type failed.

See if you can tie this in with the previous tip, so that you get an email when a new lbtype is created. Over and out.

Lets have a go at it.

Write a shell script that uses mailx to send a message with some Clearcase trigger variables embedded in it, something like this:

# Comma delimited email list of mail recipients
adminmail="youraddress@your.domain.com"
#This gets the host name
host=`uname -a |awk '{print $2}'`

#Begin send of message
echo "\
Creation of brtype \"$CLEARCASE_BRTYPE\"
by: $CLEARCASE_USER
comment: $CLEARCASE_COMMENT" \
| /usr/bin/mailx -s "[$host:CLEARCASE] New branchtype $CLEARCASE_BRTYPE created"

In this script you will see a number of variables starting with $CLEARCASE. These variables will be populated as environment variables when a trigger event occurs and launches the shell script.

The script should be saved somewhere it can be run on the UNIX server hosting the VOB, so when Clearcase triggers it can run the script. An appropriate location would be in the home directory of the VOB owner, for example:
/home/ccadmin/ccscripts/mail_new_brtype.sh

One you have saved the script, you need to actually make the trigger in Clearcase. Consider the following on creating this trigger type

• We need the trigger to run after the event, this is what is known as a “Post Operation”
• We want the trigger to be enacted after a new type is made, of type branchtype.
• We want the trigger to execute our shell script on the event occurring

You can research these options by viewing the man page of the command used to make a new trigger type:
cleartool man mktrtype

Once you have had a read of the man page, you should see some examples and options that can be used. Have a go at running the command yourself. This is what I came up with and should cover all our considerations:
cleartool mktrtype -c "Trigger to email cfgmgr on new brtype creation" -type -postop mktype -brtype -all -exec /home/ccadmin/ccscripts/mail_new_brtype.sh new_branch_trigger
My example shown here is relevant to Clearcase V6, so YMMV.

Once that has been run, you can see your new trigger listed in the trigger type list:

cleartool lstype -kind trtype
18-Dec.10:27   ccadmin     trigger type "new_branch_trigger"
  "Trigger to email cfgmgr on new brtype creation"

Next time a user creates a branch type, you will receive an email with the branch type name, branch type comment and whom created it!

Look for more upcoming tips; including how to make a trigger use a perl script to validate label names.

first I read that DBI was a memory hog so I wrote it out and put in a native postgres call, but the pgsql for perl was borked.
(comment from coding friend “DBI is fine man, works for the massess”). Next I changed the while loop over the the file to a Tail::File method, but the module was badly documenting and had issues.

I actually ended up using IO::File to tail the log and looped over the lines (in a mad loop), but it was still broken.
So I went back to DBI, which as my friend pointed out was never a problem. It worked!

It turned out the loop was the CPU hog and all I had to do was put in a sleep 1; if there was no new line in the tail.
In the process I rewrote the whole script and it barely resembles the original except for the idea. Now it has 0.5% CPU load at best. A win for coding. I’ll post the script soon and also send it to dansguardian.

Comment from perl coding friend: of course – i guessed that after u left.

I had some trouble with getting it working in the default apache chroot jail, and found very little information out there on this app in a chroot jail. Here are my notes I recorded and and steps I took. Hope it helps someone else, but YMMV:

Download chpasswd to /tmp

  cd /tmp<br />
  tar -zxvf chpasswd-2.2.3.tar.gz<br />
  cd chpasswd-2.2.3<br />
  ./configure --enable-cgidir=/var/www/cgi-bin/ --prefix=/etc<br />
  make &&  make install<br />
  make clean<br />

  mkdir /var/www/etc<br />
  mkdir /var/www/tmp<br />
  mkdir -p /var/www/var/tmp<br />
  chmod 777 /var/www/tmp<br />
  chmod 777 /var/www/var/tmp<br />
  mv /etc/chpasswd* /var/www/etc/<br />

  ldd /var/www/cgi-bin/chpasswd.cgi<br />

/var/www/cgi-bin/chpasswd.cgi:
  Start    End      Type Ref Name
  00000000 00000000 exe   1  /var/www/cgi-bin/chpasswd.cgi
  03d1e000 23d4f000 rlib  1  /usr/lib/libc.so.38.2
  09e55000 09e55000 rtld  1  /usr/libexec/ld.so

  mkdir -p /var/www/usr/lib<br />
  mkdir -p /var/www/usr/libexec<br />
  cp /usr/lib/libc.so.38.2 /var/www/usr/lib/<br />
  cp /usr/libexec/ld.so /var/www/usr/libexec/<br />

  cd /var/www/etc<br />
  vi chpasswd.conf<br />
  vi ipauth<br />

  mkdir /var/www/etc/squid<br />
  touch /var/www/etc/squid/passwd<br />
  chown root:www /var/www/etc/squid/passwd<br />
  chmod 660 /var/www/etc/squid/passwd<br />

  vi /etc/squid/squid.conf<br />
  <samp>
    auth_param basic program /usr/local/squid/libexec/ncsa_auth /var/www/etc/squid/passwd<br />
    auth_param basic children 5<br />
    auth_param basic realm Home Localnet to Internet access<br />
    auth_param basic credentialsttl 2 hours<br />
  </samp>

  squid -k reconfigure<br />

How do I list all files and file versions going into a specific build that is labelled?
Assume your label is TR1_PRE_RELEASE

  cleartool find . -version 'lbtype (TR1_PRE_RELEASE)' -print

  cleartool find . -version 'version (.../tst/LATEST)' -print

  cleartool find . -version 'version (/main/LATEST) && ! lbtype(TR1_PRE_RELEASE)' -print

  cd /<clearcase mount path>/
  clearfsimport -preview -recurse /<your app dir>/* .|grep -v unchan |grep -v identi

read more | digg story

You may be asked by someone what command level your system supports. In the case of MQ V6.0 on Solaris, the following is true:

MQCMDL_LEVEL_600
Level 600 of system control commands.
This value is returned by the following:

Websphere MQ for AIX V6.0
Websphere MQ for HP-UX V6.0
Websphere MQ for i/Series V6.0
WebSphere MQ for Linux V6.0
Websphere MQ for Solaris V6.0
Websphere MQ for Windows V6.0
Websphere MQ for z/OS V6.0

See IBM’s doco for further versions or systems.