This document describes how to install or upgrade Checkpoint or Checkpoint Commander onto a Linux/x86 host. It's not difficult, but please read carefully. NB. the rest of this document refers to "Checkpoint"; if you're installing Checkpoint Commander, substitute accordingly.
This document should be sufficient to install Checkpoint, and (hopefully) to allow you to troubleshoot configuration problems which are severe enough to prevent Checkpoint from running. Once you're up and running, you should be able to access the Checkpoint user guide. This will give you further information about running Checkpoint and administering your Checkpoint installation.
This document is part of the Checkpoint user guide. It isn't linked into the navigation for that guide (and doesn't follow its document naming convention) so that it can be available at install time, which is when it's primarily needed. The rest of the Checkpoint documentation is available after you've installed Checkpoint.
This depends on your Linux distribution. Some modern ones may come with a suitable version of java as part of the distribution itself; in other cases, you'll have to download it separately. You can get the Sun JRE from java.com. [Don't be put off by all the adverts which make it look like a domain-squatter; it is official. We checked.] The JRE distribution should tell you how to install it.
It doesn't matter where the JRE is installed, as long as the java executable is on the system PATH. Generally this will be taken care of by whatever mechanism installed Java. If this doesn't happen for some reason, the Checkpoint installer will abort (cleanly). If this happens you'll need to provide a symlink to the java executable, in a suitable folder on the system PATH.
You can install Checkpoint in one of two configurations.
Run the install script, as yourself, from an xterm
or similar,
from the folder where you unzipped the distribution. (X must be running.) If
you unzipped the distribution zip into ~yourself/tmp
, then the
code (where '>' represents the shell prompt) would be something like:
> cd ~yourself/tmp > source installSingleUser.sh
If you're upgrading, make sure that all Checkpoint applications and the Checkpoint Daemon have been shut down before continuing.
Run the install script, as root
, from the folder where you
unzipped the distribution. If you unzipped the distribution zip into
~yourself/tmp
, then the code (where '>' represents your shell
prompt, and '#' represents the root prompt) would be something like:
> sudo -i
(or)
> su root
# cd ~yourself/tmp # source installMultiUser.sh
[Tip: for fully-automatic install using all the defaults, pass the argument
--force
to the script, thus
source installMultiUser.sh --force
]
[NB. You don't need X running for multi-user install.]
The installer will ask you to set the browser to use with Checkpoint applications. It's important to set this correctly, otherwise Checkpoint applications won't work properly and will be missing important features.
Checkpoint will try to auto-detect your browser.
[Tip: the first place Checkpoint looks is /etc/alternatives/x-www-browser
.
For best results, make sure this is a symlink to your browser of choice.]
Mainstream browsers shouldn't now require any special treatment to work properly with Checkpoint (some earlier versions did). Check the release notes for recommended browser versions.
All Checkpoint applications except for Checkpoint Commander use a shared resource called the Checkpoint Daemon, and won't work properly unless it's running.
daemonctl
script as root:
# /opt/cpoint/bin/daemonctl start
[Once you're comfortable that everything's working properly, and IF your comfortable with doing so, we recommend that you start the daemon automatically from your system boot scripts. If this paragraph is gibberish, don't worry about it.]
In general you can simply leave the Daemon running in the background; it doesn't eat much. When you do need to shut it down:
daemonctl
script from a command line:
> /opt/cpoint/bin/daemonctl stop
daemonctl
script from a command line, as root:
# /opt/cpoint/bin/daemonctl stop
[NB. Don't stop the Daemon if there are any Checkpoint applications running. They'll stop working properly.]
[If you're upgrading an existing Checkpoint installation, skip this section.]
After Checkpoint has been installed for the first time, it needs to be configured. Checkpoint Command Centre will automatically start this process as soon as you run it.
root
or some other user) run the following
command (where '>' stands for the shell prompt)
Your chosen browser will be launched, and a series of web pages will guide you through the configuration process.> /opt/cpoint/bin/cpcc
All Checkpoint apps are run as executable (JStart) shell scripts, located in
/opt/cpoint/bin
. For example, Checkpoint Commander is
/opt/cpoint/bin/cpc
. Pick one! You may find it convenient to put
symlinks to commonly-used apps in /usr/local/bin
. Alternatively,
your desktop environment may allow you to add links to its launcher menu or
equivalent.
Many Checkpoint apps are installed, controlled and / or launched using
the Checkpoint Command Centre, /opt/cpoint/bin/cpcc
. Consult the
documentation for your application.
[This is a troubleshooting technique: bad lock files can (under unusual circumstances) prevent Checkpoint from running, which is why we've put it in here rather than in the main user guide.]
Every Checkpoint Persistent Object Store (aka. "CDB") has a main data file and a separate lock file. If the main data file is only used by the system user that owns it, the lock file's default permission set should be OK. However, others - ie. group and world - may not be able to access the data file even if the permissions on the data file allow them, because the permissions on the lock file don't.
To fix this, use the command-line utility /opt/cpoint/bin/lockfix
,
supplying the path to the affected main data file.
Applying lockfix
is always safe. If you get strange problems which
look as though they might be down to permissions, try it.
Don't delete lock files. If some local vandal does, then the lock file will be
re-created the next time any application touches the corresponding data file
(possibly with incorrect permissions). After root
has accidentally
locked the vandal in the tape safe for a day or so (or however long it takes for
the noise to die away; this varies from site to site, please ask your BOFH for
the exact figure, but only if you're sure this is a good idea) use
lockfix
again.
Some Checkpoint applications are services, which are accessed remotely. To support these, Checkpoint can be installed on a headless server (ie. one which does not have a GUI available). However, this is a more advanced technique which we will not cover here. (We generally do not recommend this: having a local GUI of some kind tends to make life easier.) See the Checkpoint User Guide for details.
[You only need to read this if you want strict conformance with the FHS. If you don't know what this is, you can skip this section.]
The root, shadow and temporary-log databases,
/opt/cpoint/lib/checkpoint*.cdb
,
/opt/cpoint/lib/shadow.cdb
, and
/opt/cpoint/lib/tmplog.cdb
respectively, must be mounted read-write
and may change even when system administration is not being performed.
Therefore, according to the FHS, they're supposed to live in
/var/opt/cpoint
instead of /opt/cpoint/lib
.
This violation arises because Checkpoint is a cross-platform application, and other platforms on which it must run (notably Windows) do not partition their filesystems along these axes.
To correct this, move the files referred to into the correct location, and provide symlinks thereto from the standard locations.