Argos

From GMOD
Revision as of 21:25, 15 September 2009 by Clements (Talk | contribs)

Jump to: navigation, search

Description

Argos, a.k.a. Flybase-NG, a.k.a. biodb, is designed to provide automatic replication, installation and updates of genome and organism databases and information servers, including FlyBase and euGenes. It should be not too difficult to add other organism/genome services to this replication structure.

Its main value is a collection of pre-tested and implemented common database/information service tools needed for organism database systems, which can be automatically distributed and updated to any computer.

The replication includes scripts, configurations, data, and Unix binaries for all needed programs except Perl, Java and rsync. Rsync is used as the primary distribution program.

This is server distribution system is still in development. It will be possible to use this for automated updates of mirrored servers. Uses of an automated server distribution system include local use for load distribution (apache backhand module is included for this), world-wide mirror sites for rapid local access, institution/company mirror servers for local projects and data mining. An automatable mirroring system differs from the method of providing software and data downloads by FTP in that packages of data and software in this system are kept up-to-date without human intervention. Similar package management systems such as RPM, pacman and others are well developed tools but don't quite meet the needs of this bio-database distribution.

The basic system structure is:

  common/
     java/ ; perl/ -- language packages
     servers/ -- major programs (blast, dbms, internet servers)
     systems/ -- operating system binaries of programs, packages
  docs/      -- general documents
  logs/      -- server logs
  template/  -- template information system structure
  flybase/   -- implemented genome information system structures
  eugenes/
  daphnia/

This design allows segregation common infrastructure from project-specific parts. Projects may contain any needed software along with data, web docs, database files, etc. A common symbolic link folder in each project is used to access the common software structure.

Per-package installations and updates are available, to allow customer choices of packages. This includes logic to update infrastructure software from different source sites, and focuses on using rsync as primary distribution/update tool (ftp, http, others are possible; rsync has needed file-system aware updating methods).

Evaluation of RPM, pacman, cluster-backup/mirror tools, grid packaging tools found none were quite right, so a 'quick hack' perl installation program has been built.


Developer notes

Current developers are Don Gilbert, Nihar Sheth and Victor Strelets for FlyBase-NG and euGenes uses. We hope others will try it and join us in using and developing it. Email us at argos@eugenes.org or flybase-ng@flybase.net.

Contents in SourceForge for this project are installation and configuration files. Subversion is not designed for storage and distribution of bulk data, program binaries, and the many package installations included in Argos repositories.

For this, argos/install/packages.conf configurations point to source servers for fetching ready-to-use packages, similar to the distribution system used by Globus.org for grid computing packages that are distributed to multiple grid node computers.

Also it is presumed that each implemented service will maintain software and documents separately from Argos, as the open-source software collected into the Argos commons are separately maintained, but installed for use with Argos.

GMOD developers can add new package sets to the argos/install/packages.conf which point to rsync servers for the packages.

Demo & Screenshots

A slide set outlines Argos/FlyBase-NG: flybase-ng-may03.pdf


Flybase Servers

These are overviews of FlyBase's server system structures.

Last generation

-- flybase last-generation file system structure (1996-2002) 
-- all runs via Apache web server (no separate standalone servers beyond FTP)
-- plus nightly cron job to FTP mirror data and software from main server
-- installs and runs in any Unix path

(any)/flybase-server/  

 cgi-bin/     -- web CGI programs
  
 bin-local -> sun-sparc-solaris-bin/   -- OS specific programs
 
 data -> ~ftp/flybase/data    -- primary public data files
 
 indices/ -- data search index files
 
 logs/    -- web logs
 
 mirror/  -- FTP mirroring software
 
 server/    -- public web pages and other accessory data
   .etc/    -- accessory data, miscellany
   .srs/    -- search engine
   aberrations/ -- subject class folders seen in public interface
   anatomy/ 
   genes/      
   others...
   
 source/    -- program source code
 
 sun-sparc-solaris-bin/  -- OS specific binaries
 dec-alpha-osf4-bin/
 sgi-irix-bin/
 linux-i86-elf-bin/  
 
 -- installation documents
 Install*
 Readme

Next Generation

-- flybase+ next-generation information system structure 
 -- 2003- in progress

/bio/biodb/
 common/
   java/ ; perl/ -- language packages
   servers/ -- major programs (blast, dbms, internet servers)
   systems/ -- operating system binaries of programs, packages
    
 docs/
 logs/
  
 myorg/     -- template information system structure

 flybase/   -- implemented information system structures
 eugenes/
 daphnia/
 
  
-- segregate common infrastructure from project-specific parts

-- want per-package installations and updates 
  -- allow customer choices of packages
  -- need to find/make package distribution management utility
     -- should include logic to update infrastructure software from source sites
  -- focus on 'rsync' now as main distribution tool
  -- evaluting of RPM, pacman, cluster-backup/mirror tools, 
     grid packaging tools - none quite right
  -- want CVS management of biodb structure, package info, configs 
     but not main programs, data, binaries

-- per-project packages should be fairly flexible in structure, content
  -- need to specify necessary infrastructure packages
  -- need to include security/authentication options (private and 
  public sections and methods)
   
-- retain daily mirror-ability of current server, 
  for distribution & updating of active servers to others
  and for local clusters to manage high-volume traffic
   -- more issues now with rdbms and other stand-alone server updates
   -- fixed unix path currently is used to solve configuration needs 
   -- need install/update script to allow path choices, auto-restart servers
 
   
-- current structure details  
/bio/biodb:
   
	common:
  	
		java:
			axis:
			lsid:
			lucene:
 			ogsa:
			xindice:
			
		perl:
			lsid:
 			
		servers:
		  -- hope to use 'plain vanilla' copies of these tools so updates are easy
		     and customers can replace with own versions
		  -- customize per project and via configurations 
			apache:
			berkeleydb:
			blast:
			ldap:
			mirror:
			mysql:
			postgresql:
			rsync:
			srs:
			tomcat:
			wuftpd:
 			
		source:
			apache:
				fbapache_1.3.26.tar.gz
				mod_backhand-1.2.2.tar.gz
				mod_layout-3.2.tar.gz
				mod_throttle312.tgz
			postgresql-7.3.2.tar.gz
			berkeleydb, mysql, rsync, blast, ...
 			
		system-local
    -- symlink to active systems binaries for common symlinks
     		
		systems:
		-- compiled binaries for common servers
			apple-powerpc-darwin:
			intel-linux:
			sgi-irix:
			sun-sparc-solaris:
			rsync.exclude.local
 
	docs:
 
	logs:
 	
 	rsync.exclude
  	
  	
 	myorg:
 -- template folder
		cgi-bin:
		common
		conf:
			apache.conf
			apache.conf.in
		data:
		etc:
		indices:
		secure:
		tmp:
		web:
		webapps:
 		rsync.exclude.local
		
		
	daphnia:
  -- another genome organism

	eugenes:
 -- multi organism system
 
	flybase:
 		cgi-bin:
		-- standard web CGI programs
 		  
		common
		-- symlink to common programs
 		
 		conf:
			apache.conf
			apache.conf.in
			apache.conf.local
			cvsweb.conf
  			
		data:
		-- public data files area, symlink to anon-FTP folder
			aberrations:
			allied-data:
			docs:
			extdb:
			genes:
			images:
			maps:
			news:
			nomenclature:
			refs:
			work:
 
		dbs:
			blast:
			srs:
 
		etc:   
 		-- miscellany data, others; some should be relocated
			cytodb:
			expdb:
			gmod-fb:
			gnomap:
			icons:
			insitus:
 			jdata:
			jlib:
			kevin:
			other:
			people:
			perlbio:
			pix:
	   		plib:
 			prefs:
			sean:
			stockxgene:
			templates:
 			tomcat
			transmolmaps:
			transseq:
 
 		
		indices:
		-- data search/database files (? add postgres/mysql data here)
 			blast:
			srs:

 		fbwebstart.sh
 		rsync.exclude.local
 		tmp:
 
		usenet:
 			bionet:
 			
 		web:
		--  public web interface folders
			aberrations:
			allied-data:
 			alt-views:
  			anatomy:
	 		annotfb:
			clones:
			docs:
			fbservlet:
			gbrowse_fb
			genes:
			genome-projects:
			images:
			index.html
			maps:
			people:
			pep:
			refs:
			robots.txt
			search:
			sequences:
			stocks:
 			tmp
			transposons:
  
  
   		webapps:
  		-- standard web Servlet programs
 			cvservlet:
 			fbchado:

Requirements

A current Unix computer, with several free Gigabytes of disk space, depending on which system packages are to be installed. The following software needs to be pre-installed on the system. Argos includes all other packages needed for its operation, drawn from common open-source software tools and packages used for bioinformatics databases and information systems.

These packages need to be and commonly are preinstalled

The Argos system will replicate updates to compiled programs for these operating systems, obviating need for any human-attended compiling and installation. Unix systems that have binary package support are:

  • Apple MacOSX (v10.2 build)
  • Intel Linux (kernel 2.4 build)
  • Sun Solaris (v8 build)

In this alpha 0.3 (june 2003) release, installation of common Argos packages uses ~ 200 MB of disk. Installation of a full FlyBase service uses ~ 2.5 GB of disk. Installation of a full euGenes service uses ~ 4 GB of disk.

Documentation

Quick start:

Fetch argos/install/installng.pl and run from a command line:

>perl installng.pl


Summary of steps to installation of an Argos server system

  1. Fetch the install script from a command line withrsync rsync://flybase.net/biodb/install/installng.pl . (or use web link on this page)
  2. Run perl installng.pl for summary help.
  3. Run perl installng.pl -root=/usr/local/biodb -install to create root folder and fetch the installation package (location for -root= is your choice; change below steps to match)
  4. Edit /usr/local/biodb/install/install.conf.local to set configuration. Change package set, paths and ports as desired in this install.conf.
  5. Run /usr/local/biodb/install/installng.pl -install to add the full set of packages. Packages selected from packages.conf will by copied from servers.
  6. Run /usr/local/biodb/install/run-apache to start servers
  7. Run /usr/local/biodb/install/installng.pl -update to update server periodically.

Downloads

  • Argos-based servers:http://eugenes.org/argos/ for euGenes genome database, and other services in development.
  • Main package distribution: [rsync://eugenes.org/argos]
  • With project-specific packages distributed from other servers, as specified in the argos/install/packages.conf