Difference between revisions of "Run relax at Google Cloud Computing"

From relax wiki
Jump to navigation Jump to search
(Switch to the {{relax source}} and {{relax url}} templates for the code URLs.)
 
(39 intermediate revisions by 2 users not shown)
Line 1: Line 1:
 +
The [https://en.wikipedia.org/wiki/Google_Cloud_Platform Google Cloud Platform] is a cloud computing platform that can be used for cheaply running relax on a cluster.
 +
 
== Background ==
 
== Background ==
 
This follows the same idea as from [[Run_relax_at_cloud.sagemath.com|Run relax at cloud.sagemath.com]]
 
This follows the same idea as from [[Run_relax_at_cloud.sagemath.com|Run relax at cloud.sagemath.com]]
Line 60: Line 62:
 
gcloud compute zones list
 
gcloud compute zones list
 
gcloud compute machine-types list | grep us-central1-f
 
gcloud compute machine-types list | grep us-central1-f
 +
gcloud compute machine-types list | grep us-central1-f | sort -k3 -n
 +
gcloud compute images list
  
 
# Create instance
 
# Create instance
gcloud compute instances create relax --zone us-central1-f --machine-type n1-highcpu-8 --image ubuntu-14-04 --boot-disk-size 10GB  
+
gcloud compute instances create relax --zone us-central1-f --machine-type n1-highcpu-4 --image ubuntu-14-04 --boot-disk-size 10GB  
 
gcloud compute instances list
 
gcloud compute instances list
 +
 +
# Or
 +
gcloud compute instances create relax --zone us-central1-f --machine-type n1-highcpu-4 --image rhel-6-v20160921 --boot-disk-size 10GB
 +
 
</source>
 
</source>
  
Line 70: Line 78:
 
PROJ=relax-1
 
PROJ=relax-1
 
ZONE=us-central1-f
 
ZONE=us-central1-f
PRID=relax
+
PRID=instance-1
 
GCC=gcc
 
GCC=gcc
 
gcloud compute --project $PROJ ssh --zone $ZONE $PRID
 
gcloud compute --project $PROJ ssh --zone $ZONE $PRID
 +
 +
gcloud compute --project "relax-1" ssh --zone "us-central1-f" "instance-1"
 
</source>
 
</source>
  
Line 84: Line 94:
 
For '''rapid deployment''' use this command. This will install current version of relax and trunk.  
 
For '''rapid deployment''' use this command. This will install current version of relax and trunk.  
  
[http://svn.gna.org/viewcvs/*checkout*/relax/trunk/devel_scripts/deploy_google_computing_ubuntu.sh?content-type=text%2Fplain See the install script here]
+
See the list of deploy scripts here
<source lang="bash">
+
* {{relax url|path=devel_scripts/deploy_scripts/}}
URL=http://svn.gna.org/viewcvs/*checkout*/relax/trunk/devel_scripts/deploy_google_computing_ubuntu.sh
+
 
#eval "$(curl "$URL")"
+
Then download the script that matches the system.
wget $URL
+
{{#tag: source|
cat deploy_google_computing_ubuntu.sh
+
# Select a version script
source deploy_google_computing_ubuntu.sh
+
# Redhat Version 6
 +
VERS=deploy_google_computing_redhat_6_86_x64.sh
 +
VERS=deploy_google_computing_redhat_6_86_x64_upgrade_python.sh
 +
# Centos Version 6
 +
VERS=deploy_google_computing_centos_6_86_x64_upgrade_python.sh
 +
# Redhat Version 7
 +
VERS=deploy_google_computing_redhat_7_86_x64.sh
 +
# Ubuntu
 +
VERS=deploy_google_computing_ubuntu_14-04_86_x64.sh
 +
 
 +
# Then set download URL
 +
URL={{relax url|path=devel_scripts/deploy_scripts/$VERS|view=raw}}
 +
 
 +
# Download and readl
 +
sudo yum -y install wget
 +
wget $URL $VERS
 +
mv $VERS* $VERS
 +
cat $VERS
 +
 
 +
# Load functions
 +
source $VERS
  
# See functions
+
# Do installation
cat deploy_google_computing_ubuntu.sh
 
 
installandcheck
 
installandcheck
</source>
+
| lang="bash"
 +
}}
 +
 
  
 
'''Check installation of relax'''
 
'''Check installation of relax'''
 
{{#tag:source|
 
{{#tag:source|
mpirun --np 8 relax_trunk --multi='mpi4py'
+
mpirun --np 2 relax_trunk --multi='mpi4py'
 
relax_trunk --time -x
 
relax_trunk --time -x
 
|lang="bash"
 
|lang="bash"
 
}}
 
}}
 +
 +
== Test openmpi ==
 +
 +
For BASH shell<br>
 +
{{relax source|path=devel_scripts/openmpi_test_install_bash.sh|view=raw|text=See the script here}}
 +
{{#tag: source|
 +
URL={{relax url|path=devel_scripts/openmpi_test_install_bash.sh|view=raw}}
 +
wget $URL
 +
mv openmpi_test_install_bash.sh* openmpi_test_install_bash.sh
 +
 +
# Source functions
 +
source openmpi_test_install_bash.sh
 +
testopenmpi
 +
| lang="bash"
 +
}}
 +
 +
For TCSH shell<br>
 +
{{relax source|path=devel_scripts/openmpi_test_install_tcsh.sh|view=raw|text=See the script here}}
 +
{{#tag: source|
 +
alias get 'set noglob; wget {{relax url|path=devel_scripts/openmpi_test_install_tcsh.sh|view=raw}}; mv openmpi_test_install_tcsh.sh* openmpi_test_install_tcsh.sh;unset noglob'
 +
get
 +
 +
# Source functions
 +
source openmpi_test_install_tcsh.sh
 +
| lang="bash"
 +
}}
 +
  
 
==Transfer files back and forth to the server==
 
==Transfer files back and forth to the server==
Line 121: Line 179:
 
This will start a browser to authenticate you. When it ask for a ID, just write "test" and say no to cloning.
 
This will start a browser to authenticate you. When it ask for a ID, just write "test" and say no to cloning.
  
Then go to:
+
=== ssh-config with gcloud ===
* https://console.developers.google.com/compute
+
See https://cloud.google.com/sdk/gcloud/reference/compute/config-ssh
* In the line of relax, click SSH, and drop down to: '''View gcloud command'''.
 
  
It could look like this. Try this in your own terminal. This should make some SSH keys and bring you to the server
+
gcloud compute config-ssh makes SSHing to virtual machine instances easier by adding an alias for each instance to the user SSH configuration (~/.ssh/config) file.
<source lang="bash">
 
gcloud compute --project "relax-1" ssh --zone "us-central1-f" "relax"
 
</source>
 
 
 
This should bring you to the google server.
 
<source lang="bash">
 
lscpu
 
cat $HOME/.ssh/authorized_keys
 
exit
 
 
 
# At your own computer
 
ls -l $HOME/.ssh/google*
 
cat $HOME/.ssh/config
 
ssh-add -l
 
ssh-add $HOME/.ssh/google_compute_engine
 
ssh-add -l
 
</source>
 
 
 
=== Copy files ===
 
See: https://cloud.google.com/sdk/gcloud/reference/compute/copy-files
 
  
The ssh command was
 
 
<source lang="bash">
 
<source lang="bash">
bash
 
 
PROJ=relax-1
 
PROJ=relax-1
 
ZONE=us-central1-f
 
ZONE=us-central1-f
 
PRID=relax
 
PRID=relax
 
GCC=gcc
 
GCC=gcc
gcloud compute --project $PROJ ssh --zone $ZONE $PRID
 
exit
 
</source>
 
 
We make a test file on our local conouter and copy it
 
<source lang="bash">
 
touch test.txt
 
gcloud compute --project $PROJ copy-files test.txt $PRID:~ --zone $ZONE
 
 
# Try a directory
 
mkdir testdir
 
cp test.txt testdir
 
gcloud compute --project $PROJ copy-files testdir $PRID:~ --zone $ZONE
 
 
# Back again
 
cd testdir
 
gcloud compute --project $PROJ copy-files $PRID:~/testdir . --zone $ZONE
 
</source>
 
 
=== ssh-config with gcloud ===
 
See https://cloud.google.com/sdk/gcloud/reference/compute/config-ssh
 
 
gcloud compute config-ssh makes SSHing to virtual machine instances easier by adding an alias for each instance to the user SSH configuration (~/.ssh/config) file.
 
 
<source lang="bash">
 
 
gcloud compute --project $PROJ config-ssh
 
gcloud compute --project $PROJ config-ssh
 
cat $HOME/.ssh/config
 
cat $HOME/.ssh/config
 
# ssh in
 
ssh $PRID.$ZONE.$PROJ
 
exit
 
  
 
# Use sed to change value of your local computes ssh config file
 
# Use sed to change value of your local computes ssh config file
grep "$PRID.$ZONE.$PROJ" $HOME/.ssh/config
 
sed "s/$PRID.$ZONE.$PROJ/$GCC/g" $HOME/.ssh/config
 
 
sed -i.bak "s/$PRID.$ZONE.$PROJ/$GCC/g" $HOME/.ssh/config
 
sed -i.bak "s/$PRID.$ZONE.$PROJ/$GCC/g" $HOME/.ssh/config
 
ls -l $HOME/.ssh/config*
 
ls -l $HOME/.ssh/config*
Line 200: Line 204:
 
{{#tag:source|
 
{{#tag:source|
 
#First fast check
 
#First fast check
 +
cd $HOME/Downloads
 
touch test2.txt
 
touch test2.txt
scp test2.txt $GCC:
+
scp test2.txt $GCC:
 
ssh $GCC
 
ssh $GCC
 
ls
 
ls
 +
rm test2.txt
 
exit
 
exit
 +
rm test2.txt
  
 
# Or use rsync
 
# Or use rsync
mkdir dirtest
+
mkdir deletethisdir
rsync --progress -avzhe ssh ./dirtest $GCC:
+
rsync --progress -avzhe ssh ./deletethisdir $GCC:
 
ssh $GCC
 
ssh $GCC
 
ls
 
ls
rmdir dirtest
+
rmdir deletethisdir
 
exit
 
exit
 +
rmdir deletethisdir
 +
|lang="bash"
 +
}}
 +
 +
== Use relax on a test dataset from a tutorial ==
 +
This follows some example data from the tutorial here [[Tutorial_for_sorting_data_stored_as_numpy_to_on-resonance_R1rho_analysis|Tutorial for sorting data stored as numpy to on resonance R1rho_analysis]]
 +
 +
{{relax source|path=test_suite/shared_data/dispersion/Paul_Schanda_2015_Nov|text=Overview of directory}}
 +
 +
Copy data to local dir
 +
{{#tag:source|
 +
cd $HOME
 +
cp -r $HOME/relax_trunk/test_suite/shared_data/dispersion/Paul_Schanda_2015_Nov .
 +
cd Paul_Schanda_2015_Nov
 +
|lang="bash"
 +
}}
 +
 +
 +
{{relax source|path=test_suite/shared_data/dispersion/Paul_Schanda_2015_Nov/1_tutorial_prepare_data.py|text=Prepare data}}
 +
{{#tag:source|
 +
cat 1_tutorial_prepare_data.py
 +
python 1_tutorial_prepare_data.py
 +
ls -latr
 +
|lang="bash"
 +
}}
 +
 +
{{relax source|path=test_suite/shared_data/dispersion/Paul_Schanda_2015_Nov/2_tutorial_load_data.py|text=Run analysis}}
 +
{{#tag:source|
 +
cat 2_tutorial_load_data.py
 +
mpirun --np 2 relax_trunk --multi='mpi4py' 2_tutorial_load_data.py
 +
ls -latr
 +
|lang="bash"
 +
}}
 +
 +
{{relax source|path=test_suite/shared_data/dispersion/Paul_Schanda_2015_Nov/3_tutorial_clustered.py|text=Run clustered analysis}}
 +
 +
{{#tag:source|
 +
cat 3_tutorial_clustered.py
 +
mpirun --np 2 relax_trunk --multi='mpi4py' 3_tutorial_clustered.py
 +
ls -latr
 +
|lang="bash"
 +
}}
 +
 +
See results
 +
{{#tag:source|
 +
cd results_cluster/final
 +
ls -latr
 +
cat phi_ex.out
 +
cat kex.out
 
|lang="bash"
 
|lang="bash"
 
}}
 
}}
Line 226: Line 282:
 
===By terminal===
 
===By terminal===
 
<source lang="bash">
 
<source lang="bash">
 +
PROJ=relax-1
 +
ZONE=us-central1-f
 +
PRID=relax
 +
GCC=gcc
 +
 
# Get at list of projects
 
# Get at list of projects
 
gcloud compute --project $PROJ instances list
 
gcloud compute --project $PROJ instances list
Line 249: Line 310:
  
 
==Deploy google datalab to your server==
 
==Deploy google datalab to your server==
 +
{{warning|This does not work}}
 +
 
Read more here: http://techcrunch.com/2015/10/13/google-launches-cloud-datalab-an-interactive-tool-for-exploring-and-visualizing-data
 
Read more here: http://techcrunch.com/2015/10/13/google-launches-cloud-datalab-an-interactive-tool-for-exploring-and-visualizing-data
  
Line 255: Line 318:
 
See Location at: https://appengine.google.com/
 
See Location at: https://appengine.google.com/
  
Start your server:  
+
Stop your server:  
 
<source lang="bash">
 
<source lang="bash">
gcloud compute --project $PROJ instances start --zone $ZONE $PRID
+
gcloud compute --project $PROJ instances stop --zone $ZONE $PRID
ssh $GCC
 
 
</source>
 
</source>
  
Cloud Datalab is deployed as an Google App Engine application module in a Google Developers Console project.  
+
Cloud Datalab is deployed as an Google App Engine application module in a Google Developers Console project. <br>
 
[https://console.developers.google.com/flows/enableapi?apiid=compute_component&redirect=https://console.developers.google.com/&_ga=1.74861778.1584957856.1448496848 Click here to → Enable the Google Compute Engine API.]
 
[https://console.developers.google.com/flows/enableapi?apiid=compute_component&redirect=https://console.developers.google.com/&_ga=1.74861778.1584957856.1448496848 Click here to → Enable the Google Compute Engine API.]
  
Line 271: Line 333:
 
* Launch datalab
 
* Launch datalab
 
* Sign in
 
* Sign in
* Deploy to your project. This takes '''10 min.'''
+
* Deploy to your project.
  
 
== See also ==
 
== See also ==
 
[[Category:Installation]]
 
[[Category:Installation]]

Latest revision as of 12:47, 27 October 2017

The Google Cloud Platform is a cloud computing platform that can be used for cheaply running relax on a cluster.

Background

This follows the same idea as from Run relax at cloud.sagemath.com

Make a rapid deployment of relax on a google computer cluster, and calculate for 2 months with 8 CPU's, for free.

Pricing

For this example of Google Cloud Computing, we wan't to have access to as many computer cores as possible

At November 2015, google give free trial of 300$. https://cloud.google.com/free-trial/

Note, there are some limitations:

For this time, the most important limitation is 8 CPUs.

8 CPU's

For the price calculation of:

  • 1 server
  • using Free: Debian, CentOS, ...
  • with a regular VM Class
  • Instance type of (HIGH CPU) n1-highcpu-8 (vCPUs: 8, RAM: 7.20 GB)
  • 0 SSD
  • Datacenter location United States
  • Average days per month each server is running: 1 days, per month.

This 1 day price was of November 2015, $8.06. Constantly running is $171.70 per month.

This essentially gives you 2 months of free computation with 8 CPU's.

Start free trial

Please first see this video:

Go to: https://cloud.google.com/free-trial/ and click Start your free trial.

You need to sign up with a credit card. To prevent mis-abuse from robots and hackers.

Then go to: https://console.developers.google.com

  • You should already have a project called My First Project.
  • Create a new project in top right corner of the browser under the dropdown of My First Project.
  • Call the new project name relax, Project ID relax-1, App Engine location at us-central. NOTE: App engine HAS TO BE located in united states, if datalab should be enabled later.

Create a Virtual Machine instance

Links

# List all compute instances in the project
gcloud compute instances list
gcloud compute zones list
gcloud compute machine-types list | grep us-central1-f
gcloud compute machine-types list | grep us-central1-f | sort -k3 -n
gcloud compute images list

# Create instance
gcloud compute instances create relax --zone us-central1-f --machine-type n1-highcpu-4 --image ubuntu-14-04 --boot-disk-size 10GB 
gcloud compute instances list

# Or
gcloud compute instances create relax --zone us-central1-f --machine-type n1-highcpu-4 --image rhel-6-v20160921 --boot-disk-size 10GB

Then you can SSH into the instance

PROJ=relax-1
ZONE=us-central1-f
PRID=instance-1
GCC=gcc
gcloud compute --project $PROJ ssh --zone $ZONE $PRID

gcloud compute --project "relax-1" ssh --zone "us-central1-f" "instance-1"

Access Virtual Machine instance

Install packages at server

For rapid deployment use this command. This will install current version of relax and trunk.

See the list of deploy scripts here

Then download the script that matches the system.

# Select a version script
# Redhat Version 6
VERS=deploy_google_computing_redhat_6_86_x64.sh
VERS=deploy_google_computing_redhat_6_86_x64_upgrade_python.sh
# Centos Version 6
VERS=deploy_google_computing_centos_6_86_x64_upgrade_python.sh
# Redhat Version 7
VERS=deploy_google_computing_redhat_7_86_x64.sh
# Ubuntu
VERS=deploy_google_computing_ubuntu_14-04_86_x64.sh

# Then set download URL
URL=https://sourceforge.net/p/nmr-relax/code/ci/master/tree/devel_scripts/deploy_scripts/$VERS?format=raw

# Download and readl
sudo yum -y install wget
wget $URL $VERS
mv $VERS* $VERS
cat $VERS

# Load functions
source $VERS

# Do installation
installandcheck


Check installation of relax

mpirun --np 2 relax_trunk --multi='mpi4py'
relax_trunk --time -x

Test openmpi

For BASH shell
See the script here

URL=https://sourceforge.net/p/nmr-relax/code/ci/master/tree/devel_scripts/openmpi_test_install_bash.sh?format=raw
wget $URL
mv openmpi_test_install_bash.sh* openmpi_test_install_bash.sh

# Source functions
source openmpi_test_install_bash.sh
testopenmpi

For TCSH shell
See the script here

alias get 'set noglob; wget https://sourceforge.net/p/nmr-relax/code/ci/master/tree/devel_scripts/openmpi_test_install_tcsh.sh?format=raw; mv openmpi_test_install_tcsh.sh* openmpi_test_install_tcsh.sh;unset noglob'
get

# Source functions
source openmpi_test_install_tcsh.sh


Transfer files back and forth to the server

See here

Install gcloud on your local computer

See https://cloud.google.com/sdk/

On your own computer

cd $HOME/Downloads
curl https://sdk.cloud.google.com | bash
source $HOME/.bash_profile
gcloud init

This will start a browser to authenticate you. When it ask for a ID, just write "test" and say no to cloning.

ssh-config with gcloud

See https://cloud.google.com/sdk/gcloud/reference/compute/config-ssh

gcloud compute config-ssh makes SSHing to virtual machine instances easier by adding an alias for each instance to the user SSH configuration (~/.ssh/config) file.

PROJ=relax-1
ZONE=us-central1-f
PRID=relax
GCC=gcc
gcloud compute --project $PROJ config-ssh
cat $HOME/.ssh/config

# Use sed to change value of your local computes ssh config file
sed -i.bak "s/$PRID.$ZONE.$PROJ/$GCC/g" $HOME/.ssh/config
ls -l $HOME/.ssh/config*
grep "$GCC" $HOME/.ssh/config
ssh $GCC
exit

Use scp and rsync when gcloud has been ssh-config configured

#First fast check
cd $HOME/Downloads
touch test2.txt
scp test2.txt $GCC:
ssh $GCC
ls
rm test2.txt
exit
rm test2.txt

# Or use rsync
mkdir deletethisdir
rsync --progress -avzhe ssh ./deletethisdir $GCC:
ssh $GCC
ls
rmdir deletethisdir
exit
rmdir deletethisdir

Use relax on a test dataset from a tutorial

This follows some example data from the tutorial here Tutorial for sorting data stored as numpy to on resonance R1rho_analysis

Overview of directory

Copy data to local dir

cd $HOME
cp -r $HOME/relax_trunk/test_suite/shared_data/dispersion/Paul_Schanda_2015_Nov .
cd Paul_Schanda_2015_Nov


Prepare data

cat 1_tutorial_prepare_data.py 
python 1_tutorial_prepare_data.py
ls -latr

Run analysis

cat 2_tutorial_load_data.py
mpirun --np 2 relax_trunk --multi='mpi4py' 2_tutorial_load_data.py
ls -latr

Run clustered analysis

cat 3_tutorial_clustered.py 
mpirun --np 2 relax_trunk --multi='mpi4py' 3_tutorial_clustered.py 
ls -latr

See results

cd results_cluster/final
ls -latr
cat phi_ex.out
cat kex.out

STOP the google instance after test to stop the billing of computation

See: https://cloud.google.com/compute/docs/instances/stopping-or-deleting-an-instance

By browser

Go to:

By terminal

PROJ=relax-1
ZONE=us-central1-f
PRID=relax
GCC=gcc

# Get at list of projects
gcloud compute --project $PROJ instances list

# Start it, it takes 30 seconds.
gcloud compute --project $PROJ instances start --zone $ZONE $PRID
ssh $GCC
exit

# Stop it through terminal
gcloud compute --project $PROJ instances stop --zone $ZONE $PRID
# This should now not work
ssh $GCC

Upgrade account

To use 32 CPU's you need to upgrade your account

See

Deploy google datalab to your server

Warning  This does not work

Read more here: http://techcrunch.com/2015/10/13/google-launches-cloud-datalab-an-interactive-tool-for-exploring-and-visualizing-data

Note: Cloud Datalab uses AppEngine and Managed VMs beta, which is available in the US region only.
You cannot change the location after the project has been created.
See Location at: https://appengine.google.com/

Stop your server:

gcloud compute --project $PROJ instances stop --zone $ZONE $PRID

Cloud Datalab is deployed as an Google App Engine application module in a Google Developers Console project.
Click here to → Enable the Google Compute Engine API.

Visit:

Do the following

  • Launch datalab
  • Sign in
  • Deploy to your project.

See also