Analysis of loaded web server logs using Hadoop Eco scheduling

Automates the task of periodically analyzing logs loaded into Object Storage using the Hadoop Eco scheduling function. You can easily set up Hadoop cluster execution in the desired situation for efficient resource utilization of Hadoop.

Basic information

Estimated time: 60 minutes
User environment
- Operating system: MacOS, Ubuntu
- Region: kr-central-2
Reference documents
- Automating web server log storage to Object Storage
- Web server log analysis using Hadoop Eco

Before you start

To proceed with the practice, you need to check the access key and VM access key pair.

Step 1. Check the log file uploaded to Object Storage

Create the log file example required for the hands-on according to the document. First, proceed with the corresponding hands-on tutorial, and then check whether the example log file was successfully uploaded to Object Storage.

Upload sample log file

Step 2. Create Data Catalog Resource

Data Catalog is a fully managed service that helps you identify and efficiently manage organization and user data assets within KakaoCloud. In this step, you must create the Catalog, Database, and Table that make up the Data Catalog.

Catalog is a fully managed central repository within the VPC. To use the Data Catalog service, first create a Catalog.

Item Setting
Name hands_on
VPC ${any}
Subnet ${public}
When the status of the created Catalog becomes Running, create a Database. The database of the Data Catalog is a container that stores tables.

Item Setting
Catalog hands_on
Name hands_on_db
Path:Bucket hands-on
Path:Directory nginx
Create a table, which is the metadata of the Data Catalog.
- Origin data table
- Schema
Item Setting
Database hands_on_db
Table name handson_log_original
Data storage path: Bucket name hands-on
Data storage path: Directory log/nginx
Data type CSV

Item	Setting
Name	hands_on
VPC	`${any}`
Subnet	`${public}`

Item	Setting
Catalog	hands_on
Name	hands_on_db
Path:Bucket	hands-on
Path:Directory	nginx

Item	Setting
Database	hands_on_db
Table name	handson_log_original
Data storage path: Bucket name	hands-on
Data storage path: Directory	log/nginx
Data type	CSV

Partition key	Column number	Field name	Data type
off	1	log	string
on	-	date_id	string
on	-	hour_id	string

Step 3. Create Hadoop Eco resource

Hadoop Eco is a KakaoCloud service for executing distributed processing tasks using open source frameworks. Here's how to create a Hadoop Eco resource:

Select the Hadoop Eco in the KakaoCloud Console. Click the [Create cluster] button and create a cluster corresponding to the information below.

Item Setting
Cluster Name hands-on
Cluster Version Hadoop Eco 2.0.0
Cluster Type Core Hadoop
Cluster Availability Standard
Administrator ID ${ADMIN_ID}
Administrator Password ${ADMIN_PASSWORD}
Set the master node and worker node instances.

Classification Master Node Worker Node
Instance Count 1 2
Instance Type m2a.xlarge m2a.xlarge
Volume Size 50GB 100GB
Set the key pair and network configuration (VPC, Subnet) to suit the environment in which users can access ssh. Next, click the [Create security group] button.

Item	Setting
Cluster Name	hands-on
Cluster Version	Hadoop Eco 2.0.0
Cluster Type	Core Hadoop
Cluster Availability	Standard
Administrator ID	`${ADMIN_ID}`
Administrator Password	`${ADMIN_PASSWORD}`

Classification	Master Node	Worker Node
Instance Count	1	2
Instance Type	m2a.xlarge	m2a.xlarge
Volume Size	50GB	100GB

Open the scheduling settings, select hive, and then enter a query. The query cleans the nginx default format log, computes user requests, and stores the results.

-- Set the database
USE hands_on_db;

-- Repair the table
MSCK REPAIR TABLE handson_table_original;

-- Create empty refined_json table with JSON format using JsonSerDe
CREATE EXTERNAL TABLE IF NOT EXISTS refined_json (
    remote_addr STRING,
    request_method STRING,
    request_url STRING,
    status STRING,
    request_time STRING,
    day_time STRING)
PARTITIONED BY (date_id STRING)
ROW FORMAT SERDE 'org.apache.hadoop.hive.serde2.JsonSerDe'
STORED AS TEXTFILE;

-- Populate refined_json table
INSERT INTO refined_json PARTITION (date_id)
SELECT
    split(log, ' ')[0] AS remote_addr,
    split(log, ' ')[5] AS request_method,
    split(log, ' ')[6] AS request_url,
    split(log, ' ')[8] AS status,
    split(log, ' ')[9] AS request_time,
    regexp_extract(split(log, ' ')[3], '\\d{2}:\\d{2}:\\d{2}', 0) AS day_time,
    date_id
FROM
    handson_table_original;

-- Create empty urlcount table with JSON format using JsonSerDe
CREATE EXTERNAL TABLE IF NOT EXISTS urlcount (
    request_url STRING,
    status STRING,
    count BIGINT)
PARTITIONED BY (date_id STRING)
ROW FORMAT SERDE 'org.apache.hadoop.hive.serde2.JsonSerDe'
STORED AS TEXTFILE;

-- Populate urlcount table
INSERT INTO urlcount PARTITION (date_id)
SELECT
    request_url,
    status,
    count(*) AS count,
    date_id
FROM
    refined_json
GROUP BY
    request_url,
    status,
    date_id;

Set up the cluster according to the conditions below.

Item	Setting
HDFS block size	128
HDFS replica count	2
Cluster configuration settings	Click [Enter directly] and enter the code below

Cluster configuration settings - Object Storage linkage
{
"configurations": [
    {
    "classification": "core-site",
    "properties": {
        "fs.swifta.service.kic.credential.id": "${ACCESS_KEY}",
        "fs.swifta.service.kic.credential.secret": "${ACCESS_SECRET_KEY}"
    }
    }
]
}

Set up monitoring agent, Data Catalog service integration, etc.

Item Description
Install monitoring agent Do not install
Service integration Data Catalog integration
Data catalog name Select [hands_on] created in [hands_on]
Check the entered information and create a cluster.

Item	Description
Install monitoring agent	Do not install
Service integration	Data Catalog integration
Data catalog name	Select [hands_on] created in [hands_on]

Step 4. Hadoop Eco scheduling cron job setup

Create a virtual machine to run the cron job.

Type Virtual Machine
Quantity 1
Name cron-vm
Image Ubuntu 20.04
Flavor m2a.large
Volume 20

info
The instance that will run the cron job sends a request to the external network. Therefore, you need to set up a Security Group and network environment that can send and receive requests to the external network.
Connect to the instance that will perform the cron job via ssh. You can try to connect to ssh by adding a public IP or using a Bastion host, etc.
```
ssh -i ${PRIVATE_KEY_FILE} ubuntu@${CRON_VM_ENDPOINT}
```
Install the jq package to easily handle JSON formatted data in the shell.
```
sudo apt-get update -y
sudo apt-get install -y jq
```
Connect to the instance via ssh, then create an environment variable file to load the logs by referring to the table below. The cluster ID can be checked in the cluster details.
```
cat << \EOF | sudo tee /tmp/env.sh
#!/bin/bash
export CLUSTER_ID="${CLUSTER_ID}"
export HADOOP_API_KEY="${HADOOP_API_KEY}"
export ACCESS_KEY="${ACCESS_KEY}"
export ACCESS_SECRET_KEY="${ACCESS_SECRET_KEY}"
EOF
```
Environment Variable Key Environment Variable Value
${CLUSTER_ID} Cluster ID
${HADOOP_API_KEY} Hadoop API Key
${ACCESS_KEY} Access key
${ACCESS_SECRET_KEY} Secret access key

Type	Virtual Machine
Quantity	1
Name	cron-vm
Image	Ubuntu 20.04
Flavor	m2a.large
Volume	20

Environment Variable Key	Environment Variable Value
`${CLUSTER_ID}`	Cluster ID
`${HADOOP_API_KEY}`	Hadoop API Key
`${ACCESS_KEY}`	Access key
`${ACCESS_SECRET_KEY}`	Secret access key

Write a script that requests cluster creation using the information in the environment variable file. For more information about the Hadoop Cluster API, see Hadoop Eco API.

cat << \EOF | sudo tee /tmp/exec_hadoop.sh
#!/bin/bash

. /tmp/env.sh

curl -X POST "https://hadoop-eco.kr-central-1.kakaoi.io/hadoop-eco/v1/cluster/${CLUSTER_ID}" \
    -H "Hadoop-Eco-Api-Key:${HADOOP_API_KEY}" \
    -H "Credential-ID:${ACCESS_KEY}" \
    -H "Credential-Secret:${ACCESS_SECRET_KEY}" \
    -H "Content-Type: application/json"
EOF

Use the cron package to automatically run the script at regular intervals.
```
sudo apt update -y
sudo apt install -y cron
```

Enter a command to run the script that creates the Hadoop cluster written in crontab every midnight.

cat << EOF > tmp_crontab
0 0 * * * /bin/bash /tmp/exec_hadoop.sh
EOF
sudo crontab tmp_crontab
rm tmp_crontab

Check if the cron job is registered.
```
sudo crontab -l
```
Run the command below and check if the cluster is created.
```
bash /tmp/exec_hadoop.sh
```
Check the cluster's task results. Check the result log in the bucket of Object Storage set in the Hadoop cluster creation step.
Check the task results saved in the bucket and log directory selected in Object Storage.

Before you start​

Step 1. Check the log file uploaded to Object Storage​

Step 2. Create Data Catalog Resource​

Step 3. Create Hadoop Eco resource​

Step 4. Hadoop Eco scheduling cron job setup​

Before you start

Step 1. Check the log file uploaded to Object Storage

Step 2. Create Data Catalog Resource

Step 3. Create Hadoop Eco resource

Step 4. Hadoop Eco scheduling cron job setup