{"id":124,"date":"2026-10-12T09:51:38","date_gmt":"2026-10-12T09:51:38","guid":{"rendered":"http:\/\/www.scmgalaxy.com\/tutorials\/2009\/04\/06\/cvs-configuration-install-cvs\/"},"modified":"2026-10-12T09:51:39","modified_gmt":"2026-10-12T09:51:39","slug":"cvs-configuration-install-cvs","status":"publish","type":"post","link":"https:\/\/www.devopsschool.com\/blog\/cvs-configuration-install-cvs\/","title":{"rendered":"CVS Installation and Configuration on Linux (Legacy Server and Client Setup)"},"content":{"rendered":"\n

Concurrent Versions System (CVS) is a legacy source control tool that some teams still operate for older applications. This guide turns the original quick notes into a clearer runbook for installing CVS, configuring a Linux server, and connecting clients with fewer setup mistakes.<\/p>\n\n\n\n

Originally published on 2009-04-06; fully revised and updated.<\/em><\/p>\n\n\n

\n
\"CVS<\/figure>\n<\/div>\n\n\n

Install CVS packages by Linux distribution<\/h2>\n\n\n\n

The original post referenced rpm<\/code>, up2date<\/code>, yum<\/code>, and apt-get<\/code>. The table below keeps that context and adds practical command choices used today on many systems.<\/p>\n\n\n\n

Platform<\/th>Preferred command<\/th>Legacy command from original post<\/th>Notes<\/th><\/tr><\/thead>
RHEL\/CentOS (yum-era)<\/td>sudo yum install cvs<\/code><\/td>up2date cvs<\/code><\/td>up2date<\/code> is obsolete on most current systems.<\/td><\/tr>
Fedora \/ newer RHEL family<\/td>sudo dnf install cvs<\/code><\/td>rpm -ivh cvs<\/code><\/td>Direct RPM install usually requires a full package filename\/path.<\/td><\/tr>
Debian\/Ubuntu<\/td>sudo apt-get install cvs<\/code><\/td>apt-get install cvs<\/code><\/td>Run apt-get update<\/code> first if package indexes are stale.<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n
# Example installs<\/span>\nsudo yum install cvs\nsudo dnf install cvs\nsudo apt-get update\nsudo apt-get install cvs<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
CVS server setup workflow<\/h2>\n\n\n\n
\nflowchart TD\n    A[Install CVS package] --> B[Create cvs service account]\n    B --> C[Initialize repository root]\n    C --> D[Configure xinetd pserver]\n    D --> E[Restart xinetd service]\n    E --> F[Client sets CVSROOT]\n    F --> G[cvs login and checkout]\n<\/pre>\n\n\n\n
Create a CVS service account and initialize the repository<\/h2>\n\n\n\n
The original article creates a dedicated cvs<\/code> user. Keep one repository root path consistent across all commands and configuration files.<\/p>\n\n\n
# Create service account (example)<\/span>\nsudo useradd -m -d \/home\/cvsroot -s \/sbin\/nologin cvs\nsudo passwd cvs\n\n# Initialize repository<\/span>\nsudo -u cvs cvs -d \/home\/cvsroot init<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Configure CVS pserver with xinetd (legacy method)<\/h2>\n\n\n\n
If your environment still uses xinetd and pserver, create or verify \/etc\/xinetd.d\/cvs<\/code> with values aligned to your host and repository path.<\/p>\n\n\n
service cvspserver\n{\n    disable         = no\n    socket_type     = stream\n    wait            = no\n    user            = cvs\n    group           = cvs\n    protocol        = tcp\n    server          = \/usr\/<\/span>bin\/cvs\n    server_args     = -f --allow-root=\/home\/<\/span>cvsroot pserver\n    bind            = 192.168<\/span>.1<\/span>.100<\/span>\n    port            = 2401<\/span>\n    log_on_failure  += USERID\n}<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
After saving the file, restart xinetd:<\/p>\n\n\n
sudo service xinetd restart\n# or, on systemd-based hosts<\/span>\nsudo systemctl restart xinetd<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Client configuration and first login<\/h2>\n\n\n\n
Use the server IP\/hostname and repository path in CVSROOT<\/code>, then authenticate and checkout a module.<\/p>\n\n\n
export<\/span> CVSROOT=:pserver:USERNAME@192.168<\/span>.1<\/span>.100<\/span>:\/home\/<\/span>cvsroot\ncvs login\ncvs checkout MODULE_NAME<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Access method comparison: pserver vs SSH<\/h2>\n\n\n\n
Method<\/th>Example CVSROOT<\/th>Security profile<\/th>Operational notes<\/th><\/tr><\/thead>
pserver<\/td>:pserver:USERNAME@HOST:\/home\/cvsroot<\/code><\/td>Legacy, weak password protection<\/td>Simple to enable with xinetd; avoid exposing broadly to untrusted networks.<\/td><\/tr>
SSH ext<\/td>:ext:USERNAME@HOST:\/home\/cvsroot<\/code><\/td>Encrypted transport via SSH<\/td>Usually preferred when available; reuses SSH account and key controls.<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
About the original Procedure #2<\/h2>\n\n\n\n
The original post includes manually creating CVS administrative files like Entries<\/code>, Root<\/code>, and Repository<\/code>. In normal usage, CVS manages these files automatically. A cleaner workflow is to initialize the repository once, then use checkout\/add\/commit from a client working copy.<\/p>\n\n\n
# Server-side one-time init<\/span>\nsudo -u cvs cvs -d \/home\/cvsroot init\n\n# Client workflow<\/span>\nexport CVSROOT=:pserver:USERNAME@192.168<\/span>.1<\/span>.100<\/span>:\/home\/cvsroot\ncvs checkout MODULE_NAME\ncd MODULE_NAME\ncvs add new_file.txt\ncvs commit -m \"Add new file\"<\/span><\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Legacy reference links from the original post<\/h2>\n\n\n\n