Infrastructure - galaxyproject/usegalaxy-playbook GitHub Wiki

Galaxy Infrastructure

aka, what does what?

Resources are generally divided into two domains, usegalaxy.org resources for stuff related to running Test/Main, and galaxyproject.org resources for all the other services.

Note: there is a separate playbook, infrastructure-playbook that is responsible for managing most of the non-Galaxy infrastructure, and even some of the low level stuff on Galaxy VMs and nodes. In specific, checking the inventory files of the various Ansible "environments" should give you some insight on what does what, if it's not covered in this document.

usegalaxy.org resources

The following hosts run services for usegalaxy.org and test.galaxyproject.org.

TACC Hosts

These hosts live under the .tacc.utexas.edu domain (but most have aliases under one or more of our domains). Unless needed for the task you are performing, you should use your own TACC account. See Getting Set Up At TACC.

Admin privileges (for those with access) for all hosts in this section is via sudo. On Galaxy servers, code is installed as gxsrc and run as the appropriate (g2test/g2main) Galaxy user.

host login purpose aliases
galaxy-jump your TACC user "Jump" host to resources inside of TACC, running Ansible, etc.
galaxy-test g2test
gxsrc
Galaxy Test web, jobs, and database server test.galaxyproject.org
galaxy-control RabbitMQ and Slurm controller
galaxy-cvmfs CVMFS Stratum 1 server cvmfs1-tacc0.galaxyproject.org
galaxy-main g2main
gxsrc
Galaxy Main load balancer galaxy-main.usegalaxy.org
galaxy-main1 g2main
gxsrc
Galaxy Main web workers, tusd galaxy-main1.usegalaxy.org
galaxy-main2 g2main
gxsrc
Galaxy Main staging web workers, nginx upload module galaxy-main2.usegalaxy.org
galaxy-main3 g2main
gxsrc
Galaxy Main job handlers, workflow schedulers
galaxy-main4 g2main
gxsrc
Galaxy Main web celery
galaxy-db g2main Galaxy Main database server
galaxy-vgp g2main
gxsrc
Galaxy VGP server vgp.usegalaxy.org
  • The per-host Galaxy Main aliases can be useful for testing the state of each VM independently
  • The VMs run on TACC's VMWare cluster
  • VMs are configured with varying amounts of memory/CPU by purpose
  • Galaxy logs are in journald, accessible on the appropriate VM via your TACC user with journalctl -u galaxy-*
  • Galaxy can be restarted by the gxsrc user e.g. systemctl restart galaxy-main-gunicorn@0 or galaxyctl-main graceful

CVMFS Hosts

These hosts live under the .galaxyproject.org domain. CVMFS contents are also reachable via http://datacache.galaxyproject.org/.

host login purpose virtualization host aliases
cvmfs0-psu0 g2test (update data)
root (admin)
CVMFS release manager (stratum 0) server for the data.galaxyproject.org CVMFS repo westvleteren
cvmfs0-psu1 g2test (update test)
g2main (update main)
CVMFS release manager (stratum 0) server for the test.galaxyproject.org and main.galaxyproject.org CVMFS repos smart3
cvmfs1-psu0 g2test (snapshot test/data)
g2main (snapshot main)
root (admin)
CVMFS replica server (stratum 1) server for all *.galaxyproject.org CVMFS repos thumper
cvmfs1-tacc0 g2test (snapshot test/data)
g2main (snapshot main)
CVMFS replica server (stratum 1) server for all *.galaxyproject.org CVMFS repos galaxy-cvmfs.tacc.utexas.edu
cvmfs1-iu0 g2test (snapshot test/data)
g2main (snapshot main)
centos (admin)
CVMFS replica server (stratum 1) server for all *.galaxyproject.org CVMFS repos Jetstream @ IU

Distributed Computing Hosts

host login purpose notes
jetstream2.galaxyproject.org rocky Pulsar
Slurm controller
Jetstream2
jetstream2-vgp.galaxyproject.org rocky Pulsar for vgp.usegalaxy.org Jetstream @ IU
jeltz.galaxyproject.org root datacache Guest of smart1
login3.stampede3.tacc.utexas.edu xcgalaxy Pulsar
Test/Main megablast/lastz and manual submissions
TACC Stampede2, login from galaxy-jump only
vm030.bridges2.psc.edu xcgalaxy Pulsar
Large memory tools (Trinity, Unicycler, etc.)
PSC Bridges

galaxyproject.org resources

These hosts live under the .galaxyproject.org domain.

Virtualization hosts

All virtualization hosts run SmartOS unless otherwise noted.

host guests notes
smart1 nat1
jeltz
eddie
sentry
Hosts NAT for private network hosts, datacache server, Test Toolshed, Sentry
smart2 radegast Hosts Main Toolshed
smart3 duvel
stats
cvmfs0-psu1
Hosts InfluxDB/Grafana, Galaxy Test CVMFS repo
smart5 leeroy
jenkins0
Hosts Jenkins server/worker
thumper.p cvmfs1-psu0
Hosts CVMFS stratum 1
westvleteren.p cvmfs0-psu0
westvleteren
Backups
CVMFS data repo
orval.p orval depot.galaxyproject.org

Service Hosts

Our standalone AWS account (infrastructure@...) is being phased out in favor of the Penn State Cloud Program due to contract advantages and simplified billing. At this time, you must have a Penn State Access Account (i.e. you must be a PSU employee) to gain access to the AWS Dashboard for these services.

host login purpose virtualization host notes
newtrail root (admin) LDAP
Kerberos KDC
DNS
galaxyproject.org email
EC2
artifact root (admin) LDAP
Kerberos KDC
DNS
galaxyproject.org email
Website proxies
EC2 Proxies galaxyproject.org, training.galaxyproject.org
duvel root (admin) LDAP
Kerberos KDC
DNS
smart3
radegast shed
gxsrc
root (admin)
Toolshed smart2 toolshed.g2.bx.psu.edu
eddie shed
gxsrc
root (admin)
Test Toolshed smart1 testtoolshed.g2.bx.psu.edu
orval sites
depot
bioarchive
singularity
root (admin)
depot orval.p depot.galaxyproject.org
stats root (admin) InfluxDB/Grafana smart3 stats.galaxyproject.org
sentry root (admin) Sentry smart1 sentry.galaxyproject.org
rodenbach root (admin) DHCP
TFTP
PSU local network services
metal
leeroy jenkins
root (admin)
Jenkins server smart5 jenkins.galaxyproject.org
jenkins0 jenkins
root (admin)
Jenkins worker smart5

Non-host Cloud Services

name login purpose provider
wheels galaxy-wheels IAM user wheels.galaxyproject.org S3/CloudFront
docs galaxy-docs IAM user docs.galaxyproject.org S3/CloudFront
galaxyproject.org galaxy-hub IAM user galaxyproject.org S3/EC2 (artifact)
training training.galaxyproject.org S3/GitHub Pages (live)
S3/EC2 (artifact) (archive)
biostar.usegalaxy.org biostar.usegalaxy.org archive S3/EC2 (artifact)
status [email protected] Galaxy services status statuspage.io
⚠️ **GitHub.com Fallback** ⚠️