Mirroristas/bigchaindb
2
0
Fork 0
mirror of https://github.com/bigchaindb/bigchaindb.git synced 2024-10-13 13:34:05 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #254 from bigchaindb/deploy-clients-on-aws

Browse Source 
Deploy clients on AWS
This commit is contained in:
Troy McConaghy 2016-05-09 10:06:21 +02:00
commit 4a6c3ba891
5 changed files with 184 additions and 138 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
.gitignore vendored
View File

@ -72,3 +72,4 @@ target/
deploy-cluster-aws/conf/rethinkdb.conf deploy-cluster-aws/conf/rethinkdb.conf
deploy-cluster-aws/hostlist.py deploy-cluster-aws/hostlist.py
deploy-cluster-aws/confiles/ deploy-cluster-aws/confiles/
deploy-cluster-aws/client_confile

148
deploy-cluster-aws/awsdeploy.sh Executable file
View File

@ -0,0 +1,148 @@
#! /bin/bash
# The set -e option instructs bash to immediately exit
# if any command has a non-zero exit status
set -e
USAGE="usage: ./awsdeploy.sh <number_of_nodes_in_cluster> <pypi_or_branch> <servers_or_clients>"
if [ -z "$1" ]; then
echo $USAGE
echo "No first argument was specified"
echo "It should be a number like 3 or 15"
exit 1
else
NUM_NODES=$1
fi
if [ -z "$2" ]; then
echo $USAGE
echo "No second argument was specified, so BigchainDB will be installed from PyPI"
BRANCH="pypi"
else
BRANCH=$2
fi
if [ -z "$3" ]; then
echo $USAGE
echo "No third argument was specified, so servers will be deployed"
WHAT_TO_DEPLOY="servers"
else
WHAT_TO_DEPLOY=$3
fi
if [[ ("$WHAT_TO_DEPLOY" != "servers") && ("$WHAT_TO_DEPLOY" != "clients") ]]; then
echo "The third argument, if included, must be servers or clients"
exit 1
fi
# Check for AWS private key file (.pem file)
if [ ! -f "pem/bigchaindb.pem" ]; then
echo "File pem/bigchaindb.pem (AWS private key) is missing"
exit 1
fi
# Check for the confiles directory
if [ ! -d "confiles" ]; then
echo "Directory confiles is needed but does not exist"
echo "See make_confiles.sh to find out how to make it"
exit 1
fi
# Auto-generate the tag to apply to all nodes in the cluster
TAG="BDB-"$WHAT_TO_DEPLOY"-"`date +%m-%d@%H:%M`
echo "TAG = "$TAG
# Change the file permissions on pem/bigchaindb.pem
# so that the owner can read it, but that's all
chmod 0400 pem/bigchaindb.pem
# The following Python script does these things:
# 0. allocates more elastic IP addresses if necessary,
# 1. launches the specified number of nodes (instances) on Amazon EC2,
# 2. tags them with the specified tag,
# 3. waits until those instances exist and are running,
# 4. for each instance, it associates an elastic IP address
# with that instance,
# 5. writes the shellscript add2known_hosts.sh
# 6. (over)writes a file named hostlist.py
# containing a list of all public DNS names.
python launch_ec2_nodes.py --tag $TAG --nodes $NUM_NODES
# Make add2known_hosts.sh executable then execute it.
# This adds remote keys to ~/.ssh/known_hosts
chmod +x add2known_hosts.sh
./add2known_hosts.sh
# Rollout base packages (dependencies) needed before
# storage backend (RethinkDB) and BigchainDB can be rolled out
fab install_base_software
if [ "$WHAT_TO_DEPLOY" == "servers" ]; then
# (Re)create the RethinkDB configuration file conf/rethinkdb.conf
python create_rethinkdb_conf.py
# Rollout storage backend (RethinkDB) and start it
fab install_rethinkdb
fi
# Rollout BigchainDB (but don't start it yet)
if [ "$BRANCH" == "pypi" ]; then
fab install_bigchaindb_from_pypi
else
cd ..
rm -f bigchaindb-archive.tar.gz
git archive $BRANCH --format=tar --output=bigchaindb-archive.tar
# TODO: the archive could exclude more files besides the .gitignore ones
# such as the docs. See http://tinyurl.com/zo6fxeg
gzip bigchaindb-archive.tar
mv bigchaindb-archive.tar.gz deploy-cluster-aws
cd deploy-cluster-aws
fab install_bigchaindb_from_git_archive
rm bigchaindb-archive.tar.gz
fi
# Configure BigchainDB on all nodes
if [ "$WHAT_TO_DEPLOY" == "servers" ]; then
# The idea is to send a bunch of locally-created configuration
# files out to each of the instances / nodes.
# Assume a set of $NUM_NODES BigchaindB config files
# already exists in the confiles directory.
# One can create a set using a command like
# ./make_confiles.sh confiles $NUM_NODES
# (We can't do that here now because this virtual environment
# is a Python 2 environment that may not even have
# bigchaindb installed, so bigchaindb configure can't be called)
# Transform the config files in the confiles directory
# to have proper keyrings, api_endpoint values, etc.
python clusterize_confiles.py confiles $NUM_NODES
# Send one of the config files to each instance
for (( HOST=0 ; HOST<$NUM_NODES ; HOST++ )); do
CONFILE="bcdb_conf"$HOST
echo "Sending "$CONFILE
fab set_host:$HOST send_confile:$CONFILE
done
# Initialize BigchainDB (i.e. Create the RethinkDB database,
# the tables, the indexes, and genesis glock). Note that
# this will only be sent to one of the nodes, see the
# definition of init_bigchaindb() in fabfile.py to see why.
fab init_bigchaindb
# Start BigchainDB on all the nodes using "screen"
fab start_bigchaindb
else
# Deploying clients
# The only thing to configure on clients is the api_endpoint
# It should be the public DNS name of a BigchainDB server
fab send_client_confile:client_confile
# Start sending load from the clients to the servers
fab start_bigchaindb_load
fi
# cleanup
rm add2known_hosts.sh

15
deploy-cluster-aws/fabfile.py vendored
View File

@ -146,6 +146,15 @@ def send_confile(confile):
run('bigchaindb show-config') run('bigchaindb show-config')
@task
@parallel
def send_client_confile(confile):
put(confile, 'tempfile')
run('mv tempfile ~/.bigchaindb')
print('For this node, bigchaindb show-config says:')
run('bigchaindb show-config')
# Initialize BigchainDB # Initialize BigchainDB
# i.e. create the database, the tables, # i.e. create the database, the tables,
# the indexes, and the genesis block. # the indexes, and the genesis block.
@ -164,6 +173,12 @@ def start_bigchaindb():
sudo('screen -d -m bigchaindb -y start &', pty=False) sudo('screen -d -m bigchaindb -y start &', pty=False)
@task
@parallel
def start_bigchaindb_load():
sudo('screen -d -m bigchaindb load &', pty=False)
# Install and run New Relic # Install and run New Relic
@task @task
def install_newrelic(): def install_newrelic():

130
deploy-cluster-aws/startup.sh
View File

@ -1,130 +0,0 @@
#! /bin/bash
# The set -e option instructs bash to immediately exit
# if any command has a non-zero exit status
set -e
function printErr()
{
echo "usage: ./startup.sh <tag> <number_of_nodes_in_cluster> <pypi_or_branch>"
echo "No argument $1 supplied"
}
if [ -z "$1" ]; then
printErr "<tag>"
exit 1
fi
if [ -z "$2" ]; then
printErr "<number_of_nodes_in_cluster>"
exit 1
fi
TAG=$1
NUM_NODES=$2
# If they don't include a third argument (<pypi_or_branch>)
# then assume BRANCH = "pypi" by default
if [ -z "$3" ]; then
echo "No third argument was specified, so BigchainDB will be installed from PyPI"
BRANCH="pypi"
else
BRANCH=$3
fi
# Check for AWS private key file (.pem file)
if [ ! -f "pem/bigchaindb.pem" ]; then
echo "File pem/bigchaindb.pem (AWS private key) is missing"
exit 1
fi
# Check for the confiles directory
if [ ! -d "confiles" ]; then
echo "Directory confiles is needed but does not exist"
echo "See make_confiles.sh to find out how to make it"
exit 1
fi
# Change the file permissions on pem/bigchaindb.pem
# so that the owner can read it, but that's all
chmod 0400 pem/bigchaindb.pem
# The following Python script does these things:
# 0. allocates more elastic IP addresses if necessary,
# 1. launches the specified number of nodes (instances) on Amazon EC2,
# 2. tags them with the specified tag,
# 3. waits until those instances exist and are running,
# 4. for each instance, it associates an elastic IP address
# with that instance,
# 5. writes the shellscript add2known_hosts.sh
# 6. (over)writes a file named hostlist.py
# containing a list of all public DNS names.
python launch_ec2_nodes.py --tag $TAG --nodes $NUM_NODES
# Make add2known_hosts.sh executable then execute it.
# This adds remote keys to ~/.ssh/known_hosts
chmod +x add2known_hosts.sh
./add2known_hosts.sh
# (Re)create the RethinkDB configuration file conf/rethinkdb.conf
python create_rethinkdb_conf.py
# Rollout base packages (dependencies) needed before
# storage backend (RethinkDB) and BigchainDB can be rolled out
fab install_base_software
# Rollout storage backend (RethinkDB) and start it
fab install_rethinkdb
# Rollout BigchainDB (but don't start it yet)
if [ "$BRANCH" == "pypi" ]; then
fab install_bigchaindb_from_pypi
else
cd ..
rm -f bigchaindb-archive.tar.gz
git archive $BRANCH --format=tar --output=bigchaindb-archive.tar
# TODO: the archive could exclude more files besides the .gitignore ones
# such as the docs. See http://tinyurl.com/zo6fxeg
gzip bigchaindb-archive.tar
mv bigchaindb-archive.tar.gz deploy-cluster-aws
cd deploy-cluster-aws
fab install_bigchaindb_from_git_archive
rm bigchaindb-archive.tar.gz
fi
# Configure BigchainDB on all nodes
# The idea is to send a bunch of locally-created configuration
# files out to each of the instances / nodes.
# Assume a set of $NUM_NODES BigchaindB config files
# already exists in the confiles directory.
# One can create a set using a command like
# ./make_confiles.sh confiles $NUM_NODES
# (We can't do that here now because this virtual environment
# is a Python 2 environment that may not even have
# bigchaindb installed, so bigchaindb configure can't be called)
# Transform the config files in the confiles directory
# to have proper keyrings, api_endpoint values, etc.
python clusterize_confiles.py confiles $NUM_NODES
# Send one of the config files to each instance
for (( HOST=0 ; HOST<$NUM_NODES ; HOST++ )); do
CONFILE="bcdb_conf"$HOST
echo "Sending "$CONFILE
fab set_host:$HOST send_confile:$CONFILE
done
# Initialize BigchainDB (i.e. Create the RethinkDB database,
# the tables, the indexes, and genesis glock). Note that
# this will only be sent to one of the nodes, see the
# definition of init_bigchaindb() in fabfile.py to see why.
fab init_bigchaindb
# Start BigchainDB on all the nodes using "screen"
fab start_bigchaindb
# cleanup
rm add2known_hosts.sh
# rm -rf temp_confs

28
docs/source/deploy-on-aws.md
View File

@ -147,24 +147,36 @@ You can look inside those files if you're curious. In step 2, they'll be modifie
Step 2 is to launch the nodes ("instances") on AWS, to install all the necessary software on them, configure the software, run the software, and more. Step 2 is to launch the nodes ("instances") on AWS, to install all the necessary software on them, configure the software, run the software, and more.
Here's an example of how one could launch a BigchainDB cluster of three (3) nodes tagged `wrigley` on AWS: Here's an example of how one could launch a BigchainDB cluster of three (3) nodes on AWS:
```text ```text
# in a Python 2.5-2.7 virtual environment where fabric, boto3, etc. are installed # in a Python 2.5-2.7 virtual environment where fabric, boto3, etc. are installed
cd bigchaindb cd bigchaindb
cd deploy-cluster-aws cd deploy-cluster-aws
./startup.sh wrigley 3 pypi ./awsdeploy.sh 3
``` ```
The `pypi` on the end means that it will install the latest (stable) `bigchaindb` package from the [Python Package Index (PyPI)](https://pypi.python.org/pypi). That is, on each node, BigchainDB is installed using `pip install bigchaindb`. `awsdeploy.sh` is a Bash script which calls some Python and Fabric scripts. The usage is:
`startup.sh` is a Bash script which calls some Python and Fabric scripts. The usage is:
```text ```text
./startup.sh <tag> <number_of_nodes_in_cluster> <pypi_or_branch> ./awsdeploy.sh <number_of_nodes_in_cluster> [pypi_or_branch] [servers_or_clients]
``` ```
The first two arguments are self-explanatory. The third argument can be `pypi` or the name of a local Git branch (e.g. `master` or `feat/3752/quote-asimov-on-tuesdays`). If you don't include a third argument, then `pypi` will be assumed by default. **<number_of_nodes_in_cluster>** (Required)
If you're curious what the `startup.sh` script does, the source code has lots of explanatory comments, so it's quite easy to read. Here's a link to the latest version on GitHub: [`startup.sh`](https://github.com/bigchaindb/bigchaindb/blob/master/deploy-cluster-aws/startup.sh) The number of nodes you want to deploy. Example value: 5
**[pypi_or_branch]** (Optional)
Where the nodes should get their BigchainDB source code. If it's `pypi`, then BigchainDB will be installed from the latest `bigchaindb` package in the [Python Package Index (PyPI)](https://pypi.python.org/pypi). That is, on each node, BigchainDB will be installed using `pip install bigchaindb`. You can also put the name of a local Git branch; it will be compressed and sent out to all the nodes for installation. If you don't include the second argument, then the default is `pypi`.
**[servers_or_clients]** (Optional)
If you want to deploy BigchainDB servers, then the third argument should be `servers`.
If you want to deploy BigchainDB clients, then the third argument should be `clients`.
The third argument is optional, but if you want to include it, you must also include the second argument. If you don't include the third argument, then the default is `servers`.
---
If you're curious what the `awsdeploy.sh` script does, [the source code](https://github.com/bigchaindb/bigchaindb/blob/master/deploy-cluster-aws/awsdeploy.sh) has lots of explanatory comments, so it's quite easy to read.
It should take a few minutes for the deployment to finish. If you run into problems, see the section on Known Deployment Issues below. It should take a few minutes for the deployment to finish. If you run into problems, see the section on Known Deployment Issues below.