Table of Contents

1. TCP and UDP ​ports …
   1. Corporate Firewall
   2. IPTables configuration
   3. Install RPM packages
   4. ESGF user configuration
2. Configure host certificate and CA public key

This installation guide will provide instructions about how to install an ESGF data/compute node. In order to do it, the VM should have 1 core, 2GB of RAM memory and 20GB of Hard Disk.

For the installation process, it is highly recommendable to provide more than 1 core

Pre-requisites

TCP and UDP ​ports firewall configuration

Corporate Firewall

Port	Direction	Type	Application	Description
80	in	tcp	Tomcat	Web server access
443	in	tcp	Tomcat	SSL - Secure Web Server Access.
5432	in	tcp	Postgres	Postgres Access. (not external: by default bound ONLY TO LOCAL INTERFACE)
2811	in	tcp	GridFTP	user-configured GridFTP Server control channel
(60000-61000)	in/out	tcp	GridFTP	user-configured GridFTP Server data channel (or as defined in the global variable GLOBUS_TCP_PORT_RANGE)
2812	in	tcp	GridFTP	BDM-configured GridFTP Server control channel. May run together with the user-configured one though not recommended - system resource intensive!
(60000-61000)	in/out	tcp	GridFTP	BDM-configured GridFTP Server data channel. May run together with the user-configured one though not recommended - system resource intensive!
7512	out	tcp	MyProxy	MyProxy client access to the certificate repository
8984	-	tcp	esgf-search (Tomcat)	local connection to the Solr master instance (not external!)
8983	in/out	tcp	esgf-search (Tomcat)	Connection to remotes Solr slave instance. Used in distributed search (shard).
80	out	tcp	esg-publisher	Local connection to THREDDS server (e.g., to check catalogs) and other nodes (node-manager)
443	out	tcp	esg-publisher	Local secure connection to THREDDS server (e.g., to restart the application) and to the idp

IPTables configuration

Add the rules below to the IPTables configuration file, i.e. /etc/sysconfig/iptables

-A INPUT -m state --state NEW -m tcp -p tcp --dport 22 -j ACCEPT
-A INPUT -m state --state NEW -m tcp -p tcp --dport 80 -j ACCEPT
-A INPUT -m state --state NEW -m tcp -p tcp --dport 443 -j ACCEPT
-A INPUT -m state --state NEW -m tcp -p tcp --dport 2811 -j ACCEPT
-A INPUT -m state --state NEW -m tcp -p tcp --dport 2812 -j ACCEPT
-A INPUT -m state --state NEW -m tcp -p tcp --dport 8984 -j ACCEPT
-A INPUT -m state --state NEW -m tcp -p tcp --dport 8983 -j ACCEPT
-A INPUT -m state --state NEW -m tcp -p tcp --dport 60000:61000 -j ACCEPT

then, restart the IPTables services

$ services iptables restart

Install RPM packages

First, install the sourceforge RPM repository for the *ExtUtils* packages:

$ rpm -iv http://dag.wieers.com/packages/rpmforge-release/rpmforge-release-0.3.6-1.el4.rf.x86_64.rpm

after that, the ESGF required RPM packages :

$ yum install autoconf automake bison file flex gcc gcc-c++ gettext-devel libtool libuuid-devel libxml2 libxml2-devel libxslt libxslt-devel lsof make openssl-devel pam-devel pax readline-devel tk-devel wget zlib-devel perl-Archive-Tar perl-XML-Parser libX11-devel libtool-ltdl-devel e2fsprogs-devel.x86_64 gcc-gfortran libicu-devel.x86_64 libgtextutils-devel.x86_64 perl-ExtUtils-AutoInstall.noarch perl-ExtUtils-Depends.noarch perl-ExtUtils-CBuilder.x86_64 perl-ExtUtils-CChecker.x86_64 perl-ExtUtils-Config.noarch perl-ExtUtils-DynaGlue.noarch perl-ExtUtils-Embed.x86_64 perl-ExtUtils-F77.noarch perl-ExtUtils-FakeConfig.noarch perl-ExtUtils-FindFunctions.noarch perl-ExtUtils-H2PM.noarch perl-ExtUtils-Helpers.noarch perl-ExtUtils-InstallPaths.noarch perl-ExtUtils-MakeMaker.x86_64 perl-ExtUtils-MakeMaker-Coverage.noarch perl-ExtUtils-ParseXS.x86_64 perl-ExtUtils-PerlPP.noarch perl-ExtUtils-PkgConfig.noarch perl-ExtUtils-TBone.noarch perl-ExtUtils-XSBuilder.noarch

Please make sure that the ntp package is installed $ rpm -qa | grep ntp, otherwise instal it $ yum install ntp

ESGF user configuration

Fist, add a esgf user:

$ adduser esgf
...

After that, change the password:

$ passwd esgf
...

To finish, configure the esgf user with sudoers privileges. Add the following line to /etc/sudoers file:

esgf    ALL=(ALL)       ALL

Install the ESGF data/compute node

The instructions have been provided by the IPSL¹.

Do it as esgf user

$ whoami 
esgf
$ cd /usr/local/bin
$ wget -O esg-bootstrap  http://198.128.245.140/dist/esgf-installer/esg-bootstrap
$ diff <(md5sum esg-bootstrap | tr -s " " | cut -d " " -f 1) <(curl -s  http://198.128.245.140/dist/esgf-installer/esg-bootstrap.md5 | tr -s " " | cut -d " " -f 1) 
$ chmod 555 esg-bootstrap
$ esg-bootstrap --devel

In our case, we are going to configure only data and compute types:

$ sudo ./esg-node --type data compute --install
-------------------------------------------------------

During the installation, you will have to fill in several questionnaires:

Welcome to the ESGF Node installation program! :-)

What is the fully qualified domain name of this node? [vesgdev-data.ipsl.jussieu.fr]: data.meteo.unican.es
What is the admin password to use for this installation? (alpha-numeric only) []: *******
Please re-enter password: ********
What is the name of your organization? [jussieu]: unican
Please give this node a "short" name: []: data-unican
Please give this node a more descriptive "long" name []: data-unican
What is the namespace to use for this node? (set to your reverse fqdn - Ex: "gov.llnl") [fr.jussieu.ipsl]: es.unican.meteo
What peer group(s) will this node participate in? (if not sure, use default) [esgf-test]: esgf-test
What is the default peer to this node? [esgf-node1.llnl.gov]: data.meteo.unican.es
What is the hostname of the node do you plan to publish to? [esgf-node1.llnl.gov]: vesgdev-idx.ipsl.jussieu.fr
What email address should notifications be sent as? []: meteo@unican.es
Is the database external to this node? [y/N]: 
Please enter the database connection string...
 (form: postgresql://[username]@[host]:[port]/esgcet)
What is the database connection string? [postgresql://dbsuper@localhost:5432/esgcet]: postgresql://
entered: postgresql://dbsuper@localhost:5432/esgcet
What is the (low priv) db account for publisher? [esgcet]: esgcet
What is the db password for publisher user (esgcet)? []: *****

/usr/local/pgsql/bin/createuser -U postgres -P -s -e dbsuper
Enter password for new role: 
Enter it again:

Questionnaire interactif 3

Would you like a "system" or "user" publisher configuration: 
        -------------------------------------------
        *[1] : System
         [2] : User
        -------------------------------------------
         [C] : (Custom)
        -------------------------------------------

Looking for keystore [/esg/config/tomcat/keystore-tomcat]... (don't see one)... 
Keystore setup: 
Launching Java's keytool:
store_password = ******
Would you like to use the DN: (OU=ESGF.ORG, O=ESGF) ? [Y/n]: Y
Using keystore DN = CN=vesgdev-data.ipsl.jussieu.fr, OU=ESGF.ORG, O=ESGF
Enter key password for <my_esgf_node>
        (RETURN if same as keystore password):  
Do you wish to generate a Certificate Signing Request at this time? [Y/n] Y

Please enter the password for this keystore   : 
Please re-enter the password for this keystore: 

Create user credentials
Please enter username for tomcat [dnode_user]:  
dnode_user
Please enter password for user, "dnode_user" [********]:   73769edbd97410aacfb3560ebb817f882d141517
Would you like to add another user? [y/N]: 

Please Enter the IP address of this host [134.157.179.48]:> 

Using IP: 134.157.179.48
Please select the IDP Peer for this node: 
        -------------------------------------------
        *[1] : ESGF-PCMDI-9 -> pcmdi9.llnl.gov
         [2] : ESGF-PCMDI   -> pcmdi3.llnl.gov
         [3] : ESGF-JPL     -> esg-gateway.jpl.nasa.gov
         [4] : ESGF-ORNL    -> esg2-gw.ccs.ornl.gov
         [5] : ESGF-BADC    -> cmip-gw.badc.rl.ac.uk
         [6] : ESGF-DKRZ    -> ipcc-ar5.dkrz.de
         [7] : ESGF-PNNL    -> esg1-gw.pnl.gov
         [8] : ESGF-ANL     -> dev.esg.anl.gov
         [9] : ESGF-PCMDI-TEST3 -> esgf-node3.llnl.gov
        -------------------------------------------
         [C] : (Manual Entry)
        -------------------------------------------
select [1] > C
Please enter the IDP Peer's name [ESGF-PCMDI-9] ESGF-TEST
Please enter the IDP Peer's hostname [pcmdi9.llnl.gov] data.meteo.unican.es

You have selected: (Manual Entry)
ESGF-TEST -> data.meteo.unican.es

Is this correct? [Y/n] Y

Loading KeyStore /esg/config/tomcat/esg-truststore.ts...
Opening connection to vesgdev-idx.ipsl.fr:443...
Starting SSL handshake...
Do not currently have this certificate

Server sent 2 certificate(s):

 1 Subject CN=vesgdev-idx.ipsl.fr, OU=simpleCA-vesgdev-idx.ipsl.fr, OU=GlobusTest, O=Grid
   Issuer  CN=Globus Simple CA, OU=simpleCA-vesgdev-idx.ipsl.fr, OU=GlobusTest, O=Grid
   sha1    cf f9 20 2b ce a6 bc b0 5d b4 a7 bb 0c 08 18 99 14 47 a6 86 
   md5     bd 6d ab cb 0b 75 58 fb 54 52 89 60 8e 1b 44 b8 

 2 Subject CN=Globus Simple CA, OU=simpleCA-vesgdev-idx.ipsl.fr, OU=GlobusTest, O=Grid
   Issuer  CN=Globus Simple CA, OU=simpleCA-vesgdev-idx.ipsl.fr, OU=GlobusTest, O=Grid
   sha1    06 09 9b cc b6 70 6f 3e 59 00 34 b9 fa 0a ba 87 0b f1 16 10 
   md5     0b b0 a3 56 f6 a7 c7 32 7e 35 b5 b9 e3 bb cd 26 

Enter certificate to add to trusted keystore or 'q' to quit: [1] > Y

Creating directory /esg/content/thredds/esgcet
INFO       2013-08-02 16:48:46,144 Writing THREDDS ESG master catalog /esg/content/thredds/esgcet/catalog.xml
INFO       2013-08-02 16:48:46,173 Writing THREDDS root catalog /esg/content/thredds/catalog.xml
THREDDS dataset root directories (option=thredds_dataset_roots)
Each entry has the form 'path_identifier | absolute_directory_path':
Current value is: 

esg_dataroot | /esg/data

Enter lines, or <RETURN> to end
Add new line: 

# ESGF cronjob BEGIN ###
35 0,12 * * * ESG_USAGE_PARSER_CONF=/esg/config/gridftp/esg-bdm-usage-gridftp.conf /esg/tools/esg_usage_parser 
# ESGF cronjob END ###
Is this ok ? [Y/n]Y

# ESGF cronjob BEGIN ###
35 0,12 * * * ESG_USAGE_PARSER_CONF=/esg/config/gridftp/esg-bdm-usage-gridftp.conf /esg/tools/esg_usage_parser 
5 0,12 * * * ESG_USAGE_PARSER_CONF=/esg/config/gridftp/esg-server-usage-gridftp.conf /esg/tools/esg_usage_parser 
# ESGF cronjob END ###
Is this ok ? [Y/n]Y

Server sent 2 certificate(s):

 1 Subject CN=vesgdev-idx.ipsl.fr, OU=simpleCA-vesgdev-idx.ipsl.fr, OU=GlobusTest, O=Grid
   Issuer  CN=Globus Simple CA, OU=simpleCA-vesgdev-idx.ipsl.fr, OU=GlobusTest, O=Grid
   sha1    cf f9 20 2b ce a6 bc b0 5d b4 a7 bb 0c 08 18 99 14 47 a6 86 
   md5     bd 6d ab cb 0b 75 58 fb 54 52 89 60 8e 1b 44 b8 

 2 Subject CN=Globus Simple CA, OU=simpleCA-vesgdev-idx.ipsl.fr, OU=GlobusTest, O=Grid
   Issuer  CN=Globus Simple CA, OU=simpleCA-vesgdev-idx.ipsl.fr, OU=GlobusTest, O=Grid
   sha1    06 09 9b cc b6 70 6f 3e 59 00 34 b9 fa 0a ba 87 0b f1 16 10 
   md5     0b b0 a3 56 f6 a7 c7 32 7e 35 b5 b9 e3 bb cd 26 

Enter certificate to add to trusted keystore or 'q' to quit: [1] > 1

After that, you should restart the esg-node:

$ sudo ./esg-node restart

If you want to re-install it, you have to use the force option :

$ sudo ./esg-node --type data compute --install --force

Index peer configuration

Configure host certificate and CA public key

Do it as root user

First, you have to send the csr file located under /esg/config/tomcat/ directory to the CA.

$/esg/config/tomcat/data.meteo.unican.es-esg-node.csr

Then you should put the signed csr under the /etc/grid-security/ directory.

$ /etc/grid-security/data.meteo.unican.es-esg-node-globus.csr.signed

And, if the tomcat key is not in /etc/grid-security directory, copy it inside:

$ cd /etc/grid-security
$ cp /esg/conf/tomcat/hostkey.pem ./

Install the key pair in tomcat. You will be prompted to enter the cacert file; enter the url to the index node cacert.pem:

$ esg-node --install-keypair data.meteo.unican.es-esg-node-globus.csr.signed hostkey.pem
Please enter your Certificate Athority's certificate chain file(s): 
 [enter each cert file/url press return, press return with blank entry when done]
certfile> http://vesgint-idx.ipsl.jussieu.fr/cacert.pem
...................................
...................................

This process should fetch the CA cert to /etc/grid-security/certificates

Set auto fetch certs false, otherwise /etc/grid-security/certificates/* will be overwritten by esgf-prod peer groups certificates

$ esg-node --set-auto-fetch-certs false
$ esg-node restart

Register connects to desired node, fetches and stores their certificate to enable ingress SSL connections

$ esg-node --register vesgint-idx.ipsl.jussieu.fr

Then rebuild the Tomcat's trustsore

$ esg-node --rebuild-truststore

Data Publishing

​http://devel.esgf.org/wiki/ESGF_Data_Publishing

​http://devel.esgf.org/wiki/ESGF_Project_Configuration

​http://www2-pcmdi.llnl.gov/Members/bdrach/.personal/esg-publisher-configuration/