Skip to content

sfproductlabs/scrp

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

46 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Scrp

A fully resumable horizontally (infinitely) scalable webscraper in Go. Think 1000's of machines scraping sites in a distributed way. Based on Docker Swarm, Cassandra, colly, gRPC, and my other boilerplate.

Note

Why are you even here? Maybe you could probably just use colly... Especially if you don't care about scalability... or use a shell script, for example:

#!/bin/bash
url="http://www.cityfeet.com/cont/api/search/listings-spatial"
cookie="ASP.NET_SessionId=x335iekckm5tqxcq12psv1p2; __RequestVerificationToken_L2NvbnQ1=FTTyjLMPpvjTLNYvWo5a5yFqhos830-fpyjtxwr4vsVnG8P7_bf5zEEpH4JjY2KfIKgHMuuotd9IyW4iUmSeYRHnLzQ1"
DATE=`date +%Y-%m-%d`
#for i in $(cat query.txt); do
for i in {1..9}; do
    body="{'location':{'name':'San Francisco, CA','bb':[37.708131,-122.51777,37.863424,-122.3570311],'lat':37.7857775,'lng':-122.43740055,'state':'CA','city':'San Francisco','id':'3-19282','level':3},'lt':1,'pt':0,'sort':null,'partnerId':null,'lc':[],'mode':2,'portfolio':-1,'tt':0,'ignoreLocation':false,'KeyWord':null,'rent':{'type':1,'basis':0},'term':'San Francisco, CA','PageNum':$i,'PageSize':30,'state':{'\$type':'Cityfeet.Core.Listing.MultiSearchState, Core','ProviderPosition':{'PDS':$((30 * ($i -1))),'CF':0}}}"
    content="$(curl -v -s "$url" --header "Cookie: $cookie" --header "Content-Type: application/json" --data "$body" --cookie "$cookie")"
    echo "$content" > ./data/city-feet-com-listings-spatial-$DATE-$i.json
    sleep 5
done

Why

I built this to distribute scraping across multiple servers, so as to go undetected. I could have used proxies, but wanted to reuse the code for other distributed apps.

Local Execution

Installing (local)

Run:

docker-compose up

Then (on linux - you can use brew on mac):

#apt install go
#./prereq.sh
#./build.sh

Scrape Instructions (local)

Add backend.local to your /etc/hosts file:

bash -c 'echo "127.0.0.1 backend.local" >> /etc/hosts'

First run the server on all the nodes (use the testlocal.sh script for brevity):

#GOCQL_HOST_LOOKUP_PREFER_V4=true /usr/bin/nice -n 5 ./gsvc localhost false false ./.setup/keys/rootCa.crt ./.setup/keys/cassandra-client.crt ./.setup/keys/cassandra-client.key

Notice the parameters:

[0] - cassandra-databases (comma-separated, no spaces)
[1] - cassandra-retry (should we retry execution on the cassandra cluster)
[2] - cassandra-veify (should we verify the cassandra service)
[3] - cassandra-rootca  (only use this if you need)
[4] - cassandra-client-cert (only use this if you need)
[5] - cassandra-client-key (only use this if you need)

Then send a request via the client:

./gcli backend.local:50551 https://en.wikipedia.org/wiki/List_of_HTTP_status_codes

Or something a little more complex (with domain filter & regex [note you can split regex into multiple filters using ||]):

./gcli backend.local:50551 https://en.wikipedia.org/wiki/List_of_HTTP_status_codes en.wikipedia.org,wikipedia.org ".*List.*status_codes$"

Or without the domain filter and just the regex (use the _ operator to skip):

./gcli backend.local:50551 https://en.wikipedia.org/wiki/List_of_HTTP_status_codes _ ".*List.*status_codes$"

Running on Docker Swarm

TL;DR Be your own google bot cluster in 5 minutes.

This will set you up with a cluster running 30 machines on Hetzner Cloud (change the first 20 lines to suit your own cloud provider). I use this on my own production servers. Nothing else. This assumes you've setup a project in Hetzner and an API key. It should be a fresh environment. We may delete ALL the machines in Hetzner if things don't work, so start with a fresh project and use the RIGHT API key. On your local/desktop/development machine (from the scrp github repository):

sudo apt install hcloud-cli
hcloud ssh-key create --name andy --public-key-from-file ~/.ssh/id_rsa.pub  
hcloud network create --ip-range=10.1.0.0/16 --name=aftnet
hcloud network add-subnet --ip-range=10.1.0.0/16 --type=server --network-zone=eu-central aftnet
for n in {1..30}; do (hcloud server create --name scrp$RANDOM$RANDOM$RANDOM$RANDOM --type cx11 --image debian-9 --datacenter nbg1-dc3 --network aftnet --ssh-key andy 2>&1 >/dev/null &) ; done
watch -n 5 "echo "Press Ctrl-c to exit when your server count meets the desired amount. You will need to copy and paste just the following instructions to proceed." && hcloud server list | grep 'running' | awk 'END {print NR}'"

Wait until all your servers have been created then continue:

rm *.txt
hcloud server list -o columns=name -o noheader > scrps-names.txt
hcloud server list -o columns=ipv4 -o noheader > scrps-ips.txt
cat scrps-names.txt | xargs -I {} hcloud server describe -o json {} | jq -r '.private_net[0].ip' >> scrps-vips.txt
hcloud server create --name cassandra1 --type cx41 --image debian-9 --datacenter nbg1-dc3 --network aftnet --ssh-key andy
hcloud server describe -o json cassandra1 | jq -r '.private_net[0].ip' > cassandra-vip.txt
hcloud server create --name manager1 --type cx11 --image debian-9 --datacenter nbg1-dc3 --network aftnet --ssh-key andy
hcloud server describe -o json manager1 | jq -r '.private_net[0].ip' > manager-vip.txt
scp -o StrictHostKeyChecking=no *.txt root@$(hcloud server list -o columns=ipv4,name -o noheader | grep manager1 | awk '{print $1}'):~/
scp -o StrictHostKeyChecking=no ansible/* root@$(hcloud server list -o columns=ipv4,name -o noheader | grep manager1 | awk '{print $1}'):~/
scp -o StrictHostKeyChecking=no scrp-docker-compose.yml root@$(hcloud server list -o columns=ipv4,name -o noheader | grep manager1 | awk '{print $1}'):~/
scp -o StrictHostKeyChecking=no .setup/schema* root@$(hcloud server list -o columns=ipv4,name -o noheader | grep manager1 | awk '{print $1}'):~/

If it stuffs up run DANGEROUS it will delete all your servers for the project:

hcloud server list -o columns=name -o noheader | xargs -P 8 -I {} hcloud server delete {}

If not get on the manager node ssh -l root -A $(hcloud server list -o columns=ipv4,name -o noheader | grep manager1 | awk '{print $1}') and run (copy and paste the whole thing):

apt-get update && \
apt-get upgrade -y && \
apt-get install apt-transport-https ca-certificates curl gnupg-agent software-properties-common -y && \
curl -fsSL https://download.docker.com/linux/debian/gpg | sudo apt-key add - && \
apt-key fingerprint 0EBFCD88 && \
add-apt-repository "deb [arch=amd64] https://download.docker.com/linux/debian $(lsb_release -cs) stable" && \
apt-get update && \
apt-get install docker-ce docker-ce-cli containerd.io ansible -y && \
docker swarm init --advertise-addr=ens10 && \
docker swarm join-token worker | xargs | sed -r 's/^.*(docker.*).*$/\1/' > join.sh && \
chmod +x join.sh && \
printf "\n[defaults]\nhost_key_checking = False\n" >> /etc/ansible/ansible.cfg && \
printf "\n[cassandras]\n" >> /etc/ansible/hosts && \
cat cassandra-vip.txt >> /etc/ansible/hosts && \
printf "\n[managers]\n" >> /etc/ansible/hosts && \
cat manager-vip.txt >> /etc/ansible/hosts && \
printf "\n[dockers]\n" >> /etc/ansible/hosts && \
cat manager-vip.txt >> /etc/ansible/hosts && \
cat scrps-vips.txt >> /etc/ansible/hosts && \
cat cassandra-vip.txt >> /etc/ansible/hosts && \
printf "\n[scrps]\n" >> /etc/ansible/hosts && \
cat scrps-vips.txt >> /etc/ansible/hosts && \
ansible dockers -a "uptime" && \
printf "\n            $(cat join.sh | awk '{print $0}')" >> swarm-init.yml && \
ansible-playbook swarm-init.yml && \
ansible dockers -a "docker stats --no-stream" && \
docker node ls && \
docker node update --label-add cassandra=true cassandra1 && \
docker network create -d overlay --attachable forenet --subnet 192.168.9.0/24 && \
ansible-playbook cassandras-init.yml && \
docker secret create schema.1.cql schema.1.cql && \
docker stack deploy -c scrp-docker-compose.yml scrp

Finally run a query (see above for more complex examples):

docker run -it --net=forenet sfproductlabs/scrp /app/scrp/gcli scrp_scrp:50551 https://httpbin.org/delay/2

Deploy to a swarm

Important: First make sure you deploy the schema to cassandra somewhere.

Ex. cqlsh --ssl -f ./.setup/schema.1.cql

Checkout and use the swarm-config example then on your docker swarm manager:

docker stack deploy -c scrp-docker-compose.yml scrp

Then follow the logs to see if you need to update anything:

docker service logs scrp_scrp -f

Then issue a query to the swarm (as above):

docker run -it --net=forenet sfproductlabs/scrp /app/scrp/gcli scrp_scrp:50551 https://httpbin.org/delay/2

Deploying swarm on Hetzner

$ = Your client/development machine, run them from this git repository root
# = As root on the dockermanager
d# = As root on the docker swarm drone
Setting up Hetzner

Remember to run the $ commands from the git repository root

  • Install hetzner cli:
$sudo apt install hcloud-cli
  • Go to the cloud console and create a project (important! make sure it's a new one, we will be deleting every server in here when we are done)
  • Then click on project->access->api tokens->generate token
  • Setup access on your local machine to the datacenters/project:
$hcloud context create scrp
  • Make sure there are no servers here (yet)
$hcloud server list  
  • Add your local machine to ssh auth
$hcloud ssh-key create --name andy --public-key-from-file ~/.ssh/id_rsa.pub  
  • Choose a server-type
$hcloud server-type list
$hcloud image list
$hcloud datacenter list
  • Create a network
$hcloud network create --ip-range=10.1.0.0/16 --name=aftnet
$hcloud network add-subnet --ip-range=10.1.0.0/16 --type=server --network-zone=eu-central aftnet
  • Create 30 servers (replace the type with your instance preference Ex. cx41)
$for n in {1..30}; do (hcloud server create --name scrp$RANDOM$RANDOM$RANDOM$RANDOM --type cx11 --image debian-9 --datacenter nbg1-dc3 --network aftnet --ssh-key andy &) ; done
  • In a SEPARATE terminal see the status of your booting machines (you can delete them all using the command below if something bad happens):
$watch -n 5 "echo "Press Ctrl-c to exit when your server count meets the desired amount" && hcloud server list | grep 'running' | awk 'END {print NR}'"
  • Get a list of them. IMPORTANT. This will be used to delete the servers later. Check them!
$rm *.txt
$hcloud server list -o columns=name -o noheader > scrps-names.txt
$hcloud server list -o columns=ipv4 -o noheader > scrps-ips.txt
$cat scrps-names.txt | xargs -I {} hcloud server describe -o json {} | jq -r '.private_net[0].ip' >> scrps-vips.txt
  • Create a cassandra server (16GB ram):
$hcloud server create --name cassandra1 --type cx41 --image debian-9 --datacenter nbg1-dc3 --network aftnet --ssh-key andy
$hcloud server describe -o json cassandra1 | jq -r '.private_net[0].ip' > cassandra-vip.txt
  • Create a manager node, copy some files to it and login: Addtional step required only on a mac:
eval `ssh-agent`
ssh-add ~/.ssh/id_rsa

Now create a manager, and get to it:

$hcloud server create --name manager1 --type cx11 --image debian-9 --datacenter nbg1-dc3 --network aftnet --ssh-key andy
$hcloud server describe -o json manager1 | jq -r '.private_net[0].ip' > manager-vip.txt
$scp -o StrictHostKeyChecking=no *.txt root@$(hcloud server list -o columns=ipv4,name -o noheader | grep manager1 | awk '{print $1}'):~/
$scp -o StrictHostKeyChecking=no ansible/* root@$(hcloud server list -o columns=ipv4,name -o noheader | grep manager1 | awk '{print $1}'):~/
$scp -o StrictHostKeyChecking=no scrp-docker-compose.yml root@$(hcloud server list -o columns=ipv4,name -o noheader | grep manager1 | awk '{print $1}'):~/
$scp -o StrictHostKeyChecking=no .setup/schema* root@$(hcloud server list -o columns=ipv4,name -o noheader | grep manager1 | awk '{print $1}'):~/
$ssh -l root -A $(hcloud server list -o columns=ipv4,name -o noheader | grep manager1 | awk '{print $1}')
Initializing a Docker Swarm

https://docs.docker.com/engine/install/debian/

From the #docker manager1 (last ssh command above) as root (it's important to make sure this runs perfectly) run:

apt-get update && \
apt-get upgrade -y && \
apt-get install apt-transport-https ca-certificates curl gnupg-agent software-properties-common -y && \
curl -fsSL https://download.docker.com/linux/debian/gpg | sudo apt-key add - && \
apt-key fingerprint 0EBFCD88 && \
add-apt-repository "deb [arch=amd64] https://download.docker.com/linux/debian $(lsb_release -cs) stable" && \
apt-get update && \
apt-get install docker-ce docker-ce-cli containerd.io ansible -y && \
docker swarm init --advertise-addr=ens10 && \
docker swarm join-token worker | xargs | sed -r 's/^.*(docker.*).*$/\1/' > join.sh && \
chmod +x join.sh

Now we can setup docker on all the client machines using ansible (still in the docker manager1):

printf "\n[defaults]\nhost_key_checking = False\n" >> /etc/ansible/ansible.cfg

printf "\n[cassandras]\n" >> /etc/ansible/hosts
cat cassandra-vip.txt >> /etc/ansible/hosts

printf "\n[managers]\n" >> /etc/ansible/hosts
cat manager-vip.txt >> /etc/ansible/hosts

printf "\n[dockers]\n" >> /etc/ansible/hosts
cat manager-vip.txt >> /etc/ansible/hosts
cat scrps-vips.txt >> /etc/ansible/hosts
cat cassandra-vip.txt >> /etc/ansible/hosts

printf "\n[scrps]\n" >> /etc/ansible/hosts
cat scrps-vips.txt >> /etc/ansible/hosts

Test the machines are contactable:

ansible dockers -a "uptime"

If that worked, install docker on all the machines:

printf "\n            $(cat join.sh | awk '{print $0}')" >> swarm-init.yml
ansible-playbook swarm-init.yml

Test the dockers are up:

ansible dockers -a "docker stats --no-stream"
docker node ls

Now deploy the swarm stack:

docker node update --label-add cassandra=true cassandra1
ansible-playbook cassandras-init.yml
docker network create -d overlay --attachable forenet --subnet 192.168.9.0/24 
docker secret create schema.1.cql schema.1.cql
docker stack deploy -c scrp-docker-compose.yml scrp

Give it a few minutes to boot, the scrps will take a while and likely fail a few times before they finally connect to cassandra, to debug example:

docker service ps scrp_cassandra --no-trunc
docker service logs scrp_cassandra -f

Finally run a query (see above for more complex examples):

docker run -it --net=forenet sfproductlabs/scrp /app/scrp/gcli scrp_scrp:50551 https://httpbin.org/delay/2
Getting my data out

If you are new to cassandra heres a quick intro:

$ssh -l root -A $(hcloud server list -o columns=ipv4,name -o noheader | grep cassandra1 | awk '{print $1}')
#docker ps
#docker exec -it c41 bash #replace c41 with your container
## cqlsh --ssl
cqlsh> select count(*) from scrp.content;
cqlsh> COPY scrp.content  TO 'content.csv' WITH HEADER = FALSE;
Deleting Machines
  • DELETE THEM. Yes. Let's get used to it, and make sure we know what we're doing. Double check everything before executing these commands.
$cat scrps-names.txt | xargs -I {} hcloud server delete {}

or DANGEROUS (but great for cleaning up, will include cassandra), in parallel:

hcloud server list -o columns=name -o noheader | xargs -P 8 -I {} hcloud server delete {}
Misc

Example commands (https://docs.hetzner.cloud/):

$source <(hcloud completion bash)   # bash
$source <(hcloud completion zsh)   # zsh
$hcloud server list
$hcloud ssh-key create --name demo --public-key-from-file ~/.ssh/id_rsa.pub         
$hcloud server create --name demoserver --type cx11 --image debian-9 --ssh-key demo                 
$hcloud server list             
$hcloud server list | grep -E "[0-9]+.[0-9]+.[0-9]+.[0-9]+" | sed -r 's/.*(\w[0-9]+\.[0-9]+\.[0-9]+\.[0-9]+).*$/\1/' > scrps-ips.txt
$hcloud server list | grep -E '^[^ID]' | sed -r 's/^[0-9]+ +([^ ]+).*$/\1/ig' > scrps-names.txt
Issues

Sometimes I get a network issue running a new docker container(out of swarm mode using docker run) in a node in a drained state, try and run this to fix things:

docker node update --availability active manager1

For example:

docker node update --availability active manager1 && docker run -it --net=forenet sfproductlabs/scrp /app/scrp/gcli scrp_scrp:50551 https://www.bf.com/?country=us bf.com,www.bf.com