This document describes key steps to set up Coda. For detailed step by step instructions on setting up Coda refer to topic "The Coda Administration and User Manual"  at Coda Web site. This Web page also contains references to introductory documentation about Coda. Also one can install complete documentation on his local host by downloading latest zipped and tarred archive file from any of the Coda download sites. This documentation also contains man pages on all Coda administrations tools, e.g., cfs. pdbtool etc. For example, this documentation is installed at this  location.

Note that the steps described below apply only to Coda installation on Linux. Further, after following the instructions described below, you are still not able to install Coda or make it work properly, you need to revisit the detailed documentation provided on the Coda Web site to find any clues. If reading documentation does not help send an  e-mail with appropriate symptom messages to Coda server mailing list. One can find instructions to subscribe to Coda server mailing list on the Coda Web site.

Coda has two major configurable software components; Coda server and Coda client. Most preferrably, for Coda server there should be a dedicated machine. There can be more than one Coda clients for a given Coda server, most preferably one Coda client per machine.

A. Coda Server Installation

1. First install following rpms on Coda server host.

# rpm -Uvh lwp-release.i386.rpm
# rpm -Uvh rvm-release.i386.rpm
# rpm -Uvh rpc2-release.i386.rpm
# rpm -Uvh coda-debug-server-release.i386.rpm

Where "release" stands for your release of choice. These rpms can be downloaded from the various sites referred at http://www.coda.cs.cmu.edu/. Most preferably one should choose release as the latest or few versions lower than the latest.

2. Then install the Coda server using vice-setup.

Notes for Coda server installation:

1. If you are installing Coda server by remotely logging onto the chosen host, make sure that the root user environment is exported appropriately on the shell where you doing  the installation. This is because Coda server installation uses many other binaries (or shell scripts) which are located in standard locations like "/usr/bin" etc. If these are not included in your executable search path, the installation will fail during some intermediate step.

2. For creating Coda admin during server installation, do not specify Coda admin user id as 1, even though instructions say you can use any Coda admin user id other than 0. On specifying Coda admin user id of 1, it is seen using pdbtool that Coda admin was not created because user id of 1 was already assigned to System user. pdbtool is a tool run on Coda server host and is used to create/remove Coda users/groups.  Under this circumstance any login attempt to Coda server (using clog) from a client under Coda admin will fail, perhaps because user id 1 is already used by System user. Further note that Coda user should not be confused with a UNIX user.

3. When creating a Coda volume on server as described in the INSTALL.linux help file, available at the Coda download site one should use the following command:

createvol_rep coda:root E0000100 /vicepa

where, "coda:root" is root volume.

But, the online documentation available at the Coda Web site says, specify the Coda root volume as "coda.root". So, one should use correct root volume name to createvol_rep command, the same as that specified during server installation, to avoid any unwanted problems caused due to this root volume name discrepancy.

B. Coda Client Installation

1. First install following rpms on Coda client host.

# rpm -Uvh lwp-release.i386.rpm
# rpm -Uvh rvm-release.i386.rpm
# rpm -Uvh rpc2-release.i386.rpm
# rpm -Uvh coda-debug-client-release.i386.rpm

Where "release" in all the above packages, most preferably, should  be same as what you chose on corresponding packages in the Coda server installation.

2. Activate your Coda client using command syntax:

venus-setup <comma separated Coda server hosts> <cachesizeinkb>

For example, venus-setup  wait.uccs.edu  40000

Off course, Coda server should be up and running as before you activate the client.

Notes on Coda usage on client machine:

1. After your Coda client is successfully activated go the "/coda" directory on the client host. Run the "ls" command in that directory. If you see something like "NOT_REALLY_CODA", means your Coda client/server setup is not correct and your client is not able to connect to the Coda server.

2. Once you confirm that your Coda client is connected successfully, the next step is to authenticate yourself with the Coda server using "clog" command, so that you can create/edit/delete files on the Coda file system. As a first step, authenticate yourself with the user id of Coda admin user that you specified in the Coda server setup, using clog tool. Type "clog" on the command prompt with no arguments or refer to its man page to see its usage. Note that the initial (defaulted) password for Coda admin user is "changeme", which is prompted from the "clog" command.

3.  Once you are successfully authenticated to the Coda server, you can now play with "/coda" file system and create/edit/delete your own files. You can create additional user and groups with appropriate privileges and working file systems (called Coda volumes) using "pdbtool" and "cfs" tools. Refer to chapter on "User Administration" in The Coda Administration and User Manual.

C. Shutting down Coda server

On the server type the command as below:

#volutil shutdown

This will stop the server.  If somehow the server, which is codasrv process is stil not down, you might need to kill it.

D. Shutting down the Coda client

To halt the Coda client type on client host:

# vutil shutdown
# umount /coda

If somehow the client, which is venus process is still not down, you might need to kill it.