On Linux
The following is a recommended way to configure Linux and install HarperDB. These instructions should work reasonably well for any public cloud or on-premises Linux instance.
These instructions assume that the following has already been completed:
- 1.Linux is installed
- 2.Basic networking is configured
- 3.A non-root user account dedicated to HarperDB with sudo privileges exists
- 4.An additional volume for storing HarperDB files is attached to the Linux instance
- 5.Traffic to ports 22 (SSH), 9925 (HarperDB HTTP or HTTPS), and 9926 (HarperDB Custom Functions) is permitted
For this example, we will use an AWS Ubuntu Server 18.04 LTS m5.large EC2 Instance with an additional General Purpose SSD EBS volume and the default “ubuntu” user account.
Logical Volume Manager (LVM) can be used to stripe multiple disks together to form a single logical volume. If striping disks together is not a requirement, skip these steps.
Find disk that already has a partition
used_disk=$(lsblk -P -I 259 | grep "nvme.n1.*part" | grep -o "nvme.n1")
Create array of free disks
declare -a free_disks
mapfile -t free_disks < <(lsblk -P -I 259 | grep "nvme.n1.*disk" | grep -o "nvme.n1" | grep -v "$used_disk")
Get quantity of free disks
free_disks_qty=${#free_disks[@]}
Construct pvcreate command
cmd_string=""
for i in "${free_disks[@]}"
do
cmd_string="$cmd_string /dev/$i"
done
Initialize disks for use by LVM
pvcreate_cmd="pvcreate $cmd_string"
$pvcreate_cmd
Create volume group
vgcreate_cmd="vgcreate hdb_vg $cmd_string"
$vgcreate_cmd
Create logical volume
lvcreate -n hdb_lv -i $free_disks_qty -l 100%FREE hdb_vg
Run
lsblk and note the device name of the additional volumelsblk
Create an ext4 filesystem on the volume (The below commands assume the device name is nvme1n1. If you used LVM to create logical volume, replace /dev/nvme1n1 with /dev/hdb_vg/hdb_lv)
sudo mkfs.ext4 -L hdb_data /dev/nvme1n1
Mount the file system and set the correct permissions for the directory
mkdir /home/ubuntu/hdb
sudo mount -t ext4 /dev/nvme1n1 /home/ubuntu/hdb
sudo chown -R ubuntu:ubuntu /home/ubuntu/hdb
sudo chmod 775 /home/ubuntu/hdb
Create a fstab entry to mount the filesystem on boot
echo "LABEL=hdb_data /home/ubuntu/hdb ext4 defaults,noatime 0 1" | sudo tee -a /etc/fstab
If a swap file or partition does not already exist, create and enable a 2GB swap file
sudo dd if=/dev/zero of=/swapfile bs=128M count=16
sudo chmod 600 /swapfile
sudo mkswap /swapfile
sudo swapon /swapfile
echo "/swapfile swap swap defaults 0 0" | sudo tee -a /etc/fstab
Increase the open file limits for the ubuntu user
echo "ubuntu soft nofile 500000" | sudo tee -a /etc/security/limits.conf
echo "ubuntu hard nofile 1000000" | sudo tee -a /etc/security/limits.conf
Install Node Version Manager (nvm)
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.38.0/install.sh | bash
Load nvm (or logout and then login)
export NVM_DIR="$HOME/.nvm"
[ -s "$NVM_DIR/nvm.sh" ] && \. "$NVM_DIR/nvm.sh"
[ -s "$NVM_DIR/bash_completion" ] && \. "$NVM_DIR/bash_completion"
nvm install <the node version>
Install HarperDB
npm install -g harperdb
harperdb install --TC_AGREEMENT "yes" --ROOTPATH "/home/ubuntu/hdb" --OPERATIONSAPI_NETWORK_PORT "9925" --HDB_ADMIN_USERNAME "HDB_ADMIN" --HDB_ADMIN_PASSWORD "abc123!"
HarperDB will automatically start after installation. If you wish HarperDB to start when the OS boots, you have two options
You can set up a crontab:
(crontab -l 2>/dev/null; echo "@reboot PATH=\"/home/ubuntu/.nvm/versions/node/v18.13.0/bin:$PATH\" && harperdb run") | crontab -
Or you can create a systemd script at
/etc/systemd/system/harperdb.servicePasting the following contents into the file:
[Unit]
Description=HarperDB
[Service]
Type=forking
Restart=always
User=ubuntu
Group=ubuntu
WorkingDirectory=/home/ubuntu
ExecStart=/bin/bash -c 'PATH="/home/ubuntu/.nvm/versions/node/v18.13.0/bin:$PATH"; harperdb'
[Install]
WantedBy=multi-user.target
And then running the following:
systemctl daemon-reload
systemctl enable harperdb
Last modified 6d ago