I had similar needs and used the script by Jettro (above) as a starting point to manage our artifacts by marking their state as a property (e.g. tested, releasable, production) and then deletes old artifacts based on the number of artifacts in each state.
The script file itself and useful adjunct files can be obtained from:
https://github.com/brianpcarr/ArtifactoryCurator
The readme is:
Invoke with:
groovy ArtifactoryProcess.groovy [--dry-run] [--full-log] --function <func> --value <val> --web-server http://YourWebServer --repository yourRepoName --domain <com/YourOrg> Version1 ...
where:
--domain domain : Name of the domain to scan.
--dry-run : Don't change anything; just list what would be done
--full-log : Log miscellaneous steps for processing artifacts
--function function : function to perform on artifacts
--maxInState maxInState : name of csv file with states and max counts, optional
--must-have mustHave : property required before applying delete, mark,
download or clear, optional
--password password : Password to use to access Artifactory server.
--repository repoName : Name of the repository to scan.
--targetDir targetDir : target directory for downloaded artifacts
--userName userName : userName to use to access Artifactory server
--value value : value to use with function above, often required
--web-server webServer : URL to use to access Artifactory server
Example: groovy ArtifactoryCleanup.groovy --domain domain --dry-run --full-log --function function --maxInState maxInState.csv --must-have mustHave --password password --repository repoName --targetDir targetDir --userName userName --value value --web-server webServer 1.0.1 1.0.2
Supported functions include [clear, delete, config, download, mark, repoPrint]
Columns in config csv files can be [repoName, targetDir, maxInState, domain, value, userName, mustHave, webServer, password, function]
The ArtifactoryProcess script can be used in a couple of main modes as well as
a sort of hyper mode.
The first mode is to mark sets of artifacts as being in a particular state, e.g.
groovy.bat ArtifactoryProcess.groovy --function mark --value production --web-server http://YourWebServer/ --repository yourRepoName --domain <com.YourOrg> --userName fill-in-userID --password fill-in-password 1.0.45-zdf
would mark all artifacts in yourRepoName with a version of 1.0.45-zdf as
being in production.
The other mode is to cleanup old artifacts. This could be done in two stages
where previously marked artifacts are deleted and then additional artifacts
would be marked for deletion on the next run.
Whether the cleanup is done in two stages or one, the different states in which
an artifact can be is defined in a comma separated value (csv) file which has
the name of the state and the number of artifacts to retain in that state. The
last entry in the MaxInState.csv file is an unnamed state and is the maximum
number of otherwise unmarked artifacts should be retained.
There is also a hyper-mode (function config) where each step of a clean up is
read from a comma separated value (csv) file. In this case the first line will
name the parameters which are to be specified and the values for each step will
be in the following lines. It is recommended that parameters like user ID and
password be passed on the command line, not in the config file.
To run the script, it can be run from the git root as suggested above. However,
this requires that groovy 2.3 or higher be installed (parameters to closure
support was added then and is required for the closure implementation used).
This requires that your JVM be at least 1.7. If you do not wish to install
groovy on your server, you can comment out the @Grapes sections at the top (oh
for the conditional compilations of c days gone) and build the required jar file
with 'gradlew build'. You can then run the utility with something like:
java -jar build/libs/artifactoryProcess-run.jar --dry-run --full-log --function mark --value tested --web-server http://YourWebServer/ --repository yourRepoName --domain <com.YourOrg> --userName fill-in-userID --password fill-in-password 1.0.45-zdf
The main script is:
package artifactoryProcess
import com.xlson.groovycsv.CsvIterator
import groovy.util.logging.Log
import org.jfrog.artifactory.client.Artifactory
import org.jfrog.artifactory.client.Repositories
import org.jfrog.artifactory.client.model.impl.RepositoryTypeImpl
import org.jfrog.artifactory.client.DownloadableArtifact;
import org.jfrog.artifactory.client.ItemHandle;
import org.jfrog.artifactory.client.PropertiesHandler;
import org.jfrog.artifactory.client.model.Folder
import org.kohsuke.args4j.Argument;
import org.kohsuke.args4j.CmdLineParser;
import org.kohsuke.args4j.CmdLineException;
import org.kohsuke.args4j.ExampleMode;
import org.kohsuke.args4j.Option;
import groovy.transform.stc.ClosureParams;
import groovy.transform.stc.SimpleType;
import org.jfrog.artifactory.client.ArtifactoryClient;
import org.jfrog.artifactory.client.RepositoryHandle;
import com.xlson.groovycsv.CsvParser;
@Grapes([
@GrabResolver(name='jcenter', root='http://jcenter.bintray.com/', m2Compatible=true),
@Grab(group='net.sf.json-lib', module='json-lib', version='2.4', classifier='jdk15' ),
@Grab(group='org.codehaus.groovy.modules.http-builder', module='http-builder', version='0.7'),
@Grab( group='com.xlson.groovycsv', module='groovycsv', version='1.0' ),
@GrabExclude(group='org.codehaus.groovy', module='groovy-xml')
])
@Grapes( [
@Grab(group='org.kohsuke.args4j', module='args4j-maven-plugin', version='2.0.22'),
@GrabExclude(group='org.codehaus.groovy', module='groovy-xml')
])
@Grapes([
@Grab(group='org.jfrog.artifactory.client', module='artifactory-java-client-services', version='0.13'),
@GrabExclude(group='org.codehaus.groovy', module='groovy-xml')
])
/**
* This groovy class is meant to mark artifacts for release and clean up old versions.
*
* The first mode is to mark artifacts with properties such as tested, releasable and production.
*
* The second mode could be to mark artifacts for removal based on FIFO counts of artifacts in the
* states defined in the maxInState.csv file provided. If you say want at most 5 versions of an
* artifact in the production state and there are more than that, then the oldest versions could be
* marked for removal.
*
* The third mode could be the actual deletion of any artifacts which were marked for removal.
* The delay is to allow human intervention before wholesale deletion.
*
* The versionsToUse is an array of strings that are the start of builds that should be processed.
*
* There are two additional options. The first is the dryRun option. This way you can get
* an overview of what will be processed. If specified, no artifacts will be altered.
*
* Usage example
* groovy ArtifactoryProcess.groovy --dry-run --function mark --value production --must-have releasable --web-server http://yourWebServer/artifactory/ --domain <com.YourOrg> --repository libs-release-prod 1.0.1 1.0.2
*
* @author Brian Carr (snippets from Jettro Coenradie, David Carr and others)
*/
class ArtifactoryProcess {
public static final Set< String > validFunctions = [ 'mark', 'delete', 'clear', 'download', 'config', 'repoPrint' ];
public static final Set< String > validParameters = [ 'function', 'value', 'mustHave', 'targetDir', 'maxInState', 'webServer', 'repoName', 'domain', 'userName', 'password' ];
@Option(name='--dry-run', usage='Don\'t change anything; just list what would be done')
boolean dryRun;
@Option(name='--full-log', usage='Log miscellaneous steps for processing artifacts')
boolean fullLog;
// eg --function mark
@Option(name='--function', metaVar='function', usage="function to perform on artifacts")
String function;
// eg --value production
@Option(name='--value', metaVar='value', usage="value to use with function above, often required")
String value;
// eg --must-have releasable
@Option(name='--must-have', metaVar='mustHave', usage="property required before applying delete, mark, download or clear, optional")
String mustHave;
// eg --targetDir d:/temp/bin
@Option(name='--targetDir', metaVar='targetDir', usage="target directory for downloaded artifacts")
String targetDir;
// eg --maxInState MaxInState.csv
@Option(name='--maxInState', metaVar='maxInState', usage="name of csv file with states and max counts, optional")
String maxInState;
// eg --web-server 'http://artifactory01/artifactory/'
@Option(name='--web-server', metaVar='webServer', usage='URL to use to access Artifactory server')
String webServer;
// eg --repository 'libs-release-prod'
@Option(name='--repository', metaVar='repoName', usage='Name of the repository to scan.')
String repoName;
// eg --domain 'org/apache'
@Option(name='--domain', metaVar='domain', usage='Name of the domain to scan.')
String domain;
// eg --userName cleaner
@Option(name='--userName', metaVar='userName', usage='userName to use to access Artifactory server')
String userName;
// eg --password SomePswd
@Option(name='--password', metaVar='password', usage='Password to use to access Artifactory server.')
String password;
@Argument
ArrayList<String> versionsToUse = new ArrayList<String>();
class PathAndDate{
String path;
Date dtCreated;
}
class StateRecord {
String state;
int cnt;
List< PathAndDate > pathAndDate;
}
@SuppressWarnings(["SystemExit", "CatchThrowable"])
static void main( String[] args ) {
try {
new ArtifactoryProcess().doMain( args );
} catch (Throwable throwable) {
// Java returns exit code 0 if it terminates because of an uncaught Throwable.
// That's bad if we have a process like Bamboo depending on errors being non-zero.
// Thus, we catch all Throwables and explicitly set the exit code.
println( "Unexpected error: ${throwable}" )
System.exit(1)
}
System.exit(0);
}
List< StateRecord > stateSet = [];
Artifactory srvr;
RepositoryHandle repo;
private int numProcessed = 0;
String firstFunction;
String lastConfig;
void doMain( String[] args ) {
CmdLineParser parser = new CmdLineParser( this );
try {
parser.parseArgument(args);
if( function == 'config' && value == null ) {
throw new CmdLineException("You must provide a config.csv file as the value if you specify the config function.");
}
firstFunction = function; // Flag in case we recurse into config files, where did we start
if( function == 'config' ) {
processConfig();
return;
} else {
checkParms();
}
} catch(CmdLineException ex) {
System.err.println(ex.getMessage());
System.err.println();
System.err.println("groovy ArtifactoryProcess.groovy [--dry-run] [--full-log] --function <func> --value <val> --web-server http://YourWebServer --repository libs-release-prod --domain <com/YourOrg> Version1 ...");
parser.printUsage(System.err);
System.err.println();
System.err.println(" Example: groovy ArtifactoryProcess.groovy"+parser.printExample(ExampleMode.ALL)+" 1.0.1 1.0.2");
System.err.println();
System.err.println(" Supported functions include ${validFunctions}" );
System.err.println();
System.err.println(" Columns in config csv files can be ${validParameters}" );
return;
}
String stateLims;
if( maxInState != null && maxInState.size() > 0 ) stateLims = "(using stateLims)" else stateLims = "(no stateLims)"
println( "Started processing of $function with ${(value==null)?mustHave:value} $stateLims on $webServer in $repoName/$domain with $versionsToUse." );
withClient { newClient ->
srvr = newClient;
if( function == 'repoPrint' ) printRepositories();
else {
processRepo();
}
}
}
def processRepo() {
numProcessed = 0; // Reset count from last repo.
repo = srvr.repository( repoName );
processArtifactsRecursive( domain );
if( dryRun ) {
println "$numProcessed folders would have been $function[ed] with $value.";
} else {
println "$numProcessed folders were $function[ed] with $value.";
}
}
def processConfig() {
File configCSV = new File( value );
lastConfig = value; // Record which csv file we have last dived into.
Artifactory mySrvr = srvr; // Each line of config could have a different web server, preserve connection in case recursing
configCSV.withReader {
CsvIterator csvIt = CsvParser.parseCsv( it );
for( csvRec in csvIt ) {
if (fullLog) println("Step is ${csvRec}");
Map cols = csvRec.properties.columns;
String func = csvRec.function;
def hasFunc = cols.containsKey( 'function' );
def has = cols.containsKey( 'targetDir' );
if( cols.containsKey( 'function' ) && !noValue( csvRec.function ) ) function = csvRec.function ;
if( cols.containsKey( 'value' ) && !noValue( csvRec.value ) ) value = csvRec.value ;
if( cols.containsKey( 'targetDir' ) && !noValue( csvRec.targetDir ) ) targetDir = csvRec.targetDir ;
if( cols.containsKey( 'maxInState' ) && !noValue( csvRec.maxInState ) ) maxInState = csvRec.maxInState;
if( cols.containsKey( 'webServer' ) && !noValue( csvRec.webServer ) ) webServer = csvRec.webServer ;
if( cols.containsKey( 'repoName' ) && !noValue( csvRec.repoName ) ) repoName = csvRec.repoName ;
if( cols.containsKey( 'domain' ) && !noValue( csvRec.domain ) ) repoName = csvRec.domain ;
if( cols.containsKey( 'userName' ) && !noValue( csvRec.userName ) ) userName = csvRec.userName ;
if( cols.containsKey( 'password' ) && !noValue( csvRec.password ) ) password = csvRec.password ;
if( cols.containsKey( 'mustHave' ) ) mustHave = csvRec.mustHave; // Can clear out mustHave value
checkParms();
withClient { newClient ->
srvr = newClient;
processRepo();
}
srvr = mySrvr; // Restore previous web server connection
}
}
}
def checkParms() {
if( !noValue( maxInState ) ) {
stateSet.clear(); // Throw away any previous states from last step
File stateFile = new File( maxInState );
def RC = stateFile.withReader {
CsvIterator csvFile = CsvParser.parseCsv( it );
for( csvRec in csvFile ) {
String state = csvRec.properties.values[ 0 ];
String strCnt = csvRec.properties.values[ 1 ];
if( fullLog ) println( "State ${state} allowed ${strCnt}" );
int count = 0;
if( strCnt.integer ) count = strCnt.toInteger();
if( count < 0 ) count = 0;
// Iterator lies and claims there is a next when there isn't. Force break on empty state.
stateSet.add( new StateRecord( state: state, cnt: count, pathAndDate: [] ) );
}
}
}
String prefix;
if( firstFunction == 'config' && function != 'config' ) {
prefix = "While processing ${lastConfig} encountered, ";
} else prefix = '';
if( !validFunctions.contains( function ) ) {
throw new CmdLineException( "${prefix}Unrecognized function ${function}, function is required and must be one of ${validFunctions}." );
}
if( function == 'mark' && noValue( value ) ) {
throw new CmdLineException( "${prefix}You must provide a value to mark with if you specify the mark function." );
}
if( function == 'clear' && noValue( value ) ) {
throw new CmdLineException( "${prefix}You must provide a value to clear with if you specify the clear function." );
}
if( function != 'repoPrint' && noValue( domain ) ) {
throw new CmdLineException( "${prefix}You must provide a domain to use with the ${function} function." );
}
if( function == 'download' ) {
if( noValue( targetDir ) ) targetDir = '.';
}
if( noValue( webServer ) || noValue( userName ) || noValue( password ) || noValue( repoName ) ) {
throw new CmdLineException( "${prefix}You must provide the webServer, userName, password and repository name values to use." );
}
if( versionsToUse.size() == 0 && stateSet.size() == 0 && function != 'repoPrint' ) {
throw new CmdLineException( "${prefix}You must provide maxInState or a list of artifacts / versions to act upon." );
}
}
Boolean noValue( var ) {
return var == null || var == '';
}
/**
* Print information about all the available repositories in the configured Artifactory
*/
def printRepositories() {
Repositories repos = srvr.repositories();
List repoList = repos.list( RepositoryTypeImpl.LOCAL );
for( it in repoList ) {
println "key :" + it.key
println "type : " + it.type
println "description : " + it.description
println "url : " + it.url
println ""
};
}
/**
* Recursively removes all folders containing builds that start with the configured paths.
*
* @param path String containing the folder to check and use the childs to recursively check as well.
* @return Number with the amount of folders that were processed.
*/
private int processArtifactsRecursive( String path ) {
ItemHandle item = repo.folder( path );
// def RC = item.isFolder(); This lies, always returns true even for a file, go figure!
// def RC = path.endsWith('.xml'); // item.info() fails for simple files, go figure!
if( !path.endsWith( '.xml' ) &&
!path.endsWith( '.jar' ) &&
item.isFolder() ) {
Folder fldr;
try{
fldr = item.info()
} catch( Exception e ) {
println( "Error accessing $webServer/$repoName/$path" );
throw( e );
};
for( kid in fldr.children ) {
boolean processed = false;
if( stateSet.size() > 0 ) {
if( isEndNode( kid.uri )) {
processed = groupFolders( path + kid.uri );
}
} else {
versionsToUse.find { version ->
if( kid.uri.startsWith( '/' + version ) ) {
numProcessed += processItem( path + kid.uri );
return true; // Once we find a match, no others are interesting, we are outta here
} else return false; // Just formalize the on to next iterator
}
}
if( !processed ) {
processArtifactsRecursive( path + kid.uri );
}
}
}
/* If we are counting number in each state, our lists should be all set now */
if( stateSet.size() > 0 ) {
processSet();
}
return numProcessed;
}
private boolean processedThis( String vrsn, kid ) {
if( kid.uri.startsWith('/' + vrsn )) {
numProcessed += processItem( vrsn + kid.uri );
return true; // Once we find a match, no others are interesting, we are outta here
} else return false; // Just formalize the on to next iterator
}
// True if nodeName is of form int.int.other, could be one line, but how would you debug it.
private boolean isEndNode( String nodeName ){
int firstDot = nodeName.indexOf( '.' );
if( firstDot <= 1 ) return false; // nodeName starts with '/' which is ignored
int secondDot = nodeName.indexOf( '.', firstDot + 1 );
if( secondDot <= 0 ) return false;
String firstInt = nodeName.substring( 1, firstDot ); // nodeName starts with '/' which is ignored
if( !firstInt.isInteger() ) return false;
String secondInt = nodeName.substring( firstDot + 1, secondDot );
if( secondInt.isInteger() ) return true;
return false;
}
private boolean groupFolders( String path ) {
Map<String, List<String>> props;
stateSet.find { rec ->
ItemHandle folder = repo.folder( path );
if( rec.state.size() > 0 ) {
props = folder.getProperties( rec.state );
}
if( rec.state.size() <= 0 || props.size() > 0 ) {
PathAndDate nodePathDate = new PathAndDate();
nodePathDate.path = path;
nodePathDate.dtCreated = folder.info().lastModified;
rec.pathAndDate.add( nodePathDate ); // process this one
return true; // No others are interest, break out of iterator
} else return false; // On to next iterator
}
return true; // We always process all nodes which are end nodes
}
private boolean processSet() {
for( set in stateSet ) {
int del = set.cnt;
if( set.pathAndDate.size() < del ) {
del = set.pathAndDate.size() }
else {
set.pathAndDate.sort() { a,b -> b.dtCreated <=> a.dtCreated }; // Sort newest first to preserve newest
}
while( del > 0 ) {
set.pathAndDate.remove( 0 );
del--;
}
while( set.pathAndDate.size() > 0 ) {
numProcessed += processItem( set.pathAndDate[ 0 ].path );
set.pathAndDate.remove( 0 );
}
}
return true;
}
private int processItem( String path ) {
int retVal = 0;
if( fullLog ) println "Processing folder: ${path}, ${function} with ${value}.";
def RC;
ItemHandle folder = repo.folder( path );
Map<String, List<String>> props;
boolean hasRqrd = true;
if( !noValue( mustHave ) ) {
props = folder.getProperties( mustHave );
if( props.size() > 0 ) hasRqrd = true; else hasRqrd = false; // I like this better than ternary operator, you?
}
if( !hasRqrd ) return retVal;
switch( function ) {
case 'delete':
if( !dryRun ) RC = repo.delete( path );
retVal++;
break;
case 'download':
if( folder.isFolder() ) {
Folder item = folder.info();
item.children.find() { kid ->
if( kid.uri.endsWith('.jar') ) {
DownloadableArtifact DA = repo.download( path + kid.uri );
InputStream dlJar = DA.doDownload(); // Open Source
FileWriter lclJar = new FileWriter( targetDir + kid.uri, false ); // Open Dest
for( id in dlJar ) { lclJar.write( id ); } // Copy contents
if( fullLog ) println( "Downloaded ${path + kid.uri} to ${targetDir + kid.uri}." );
retVal++;
return true;
}
}
}
break;
case 'mark':
props = folder.getProperties( value );
if( props.size() == 0 ) {
PropertiesHandler item = folder.properties();
PropertiesHandler PH = item.addProperty( value, 'true' );
if( !dryRun ) RC = PH.doSet();
retVal++;
}
break;
case 'clear':
props = folder.getProperties( value );
if( props.size() == 1 ) {
if( !dryRun ) RC = folder.deleteProperty( value ); // Null return is success, go figure!
retVal++;
}
break;
default:
println( "Unknown function $function with $value encountered on ${path}.")
}
if( retVal > 0 ) println "Completed $function on $path with ${(value==null)?mustHave:value}.";
return retVal;
}
private <T> T withClient( @ClosureParams( value = SimpleType, options = "org.jfrog.artifactory.client.Artifactory" ) Closure<T> closure ) {
def client = ArtifactoryClient.create( "${webServer}artifactory", userName, password )
try {
return closure( client )
} finally {
client.close()
}
}
}