- CPU: Intel Atom® Processor
- Memory: 8GB
- OS: Ubuntu 16.04 LTS
- Built-in Storage: 16GB
- Graphics: Intel Gen9 Graphics Engine
- Supports:
- Intel® Movidius™ Neural Compute Stick
- Intel® RealSense™ depth sensor
Other than the OS upgrade, it seems to be ok to apply update to the system modules. To run the module update, open a terminal window and run the following commands:
sudo apt update
sudo apt upgrade -yIf it prompted for awscam update, go ahead and install the update by running the following command:
sudo apt install awscamTo access the device for the first time, for new device or as a result of factory reset, use aws_cam to login to the device. The password is aws_cam. It will prompt for password reset. Remember to first enter the existing password which is aws_cam and then choose a good password.
As the DeepLens is reaching EOL, it is better to disable the services integrating with AWS. To disable the services run the ./scripts/awscam-disable after login to the device. The script can be downloaded via copying the raw file or cloning the git repo locally after seeting up a different login user to be discussed further down.
sudo ./scripts/awscam-disableTo see a list of running services, do
systemctlThe default setting does not include the channels outside of AWS. To install some useful components such as tmux, samba, etc..., enable the following channels:
- Canonical-supported free and open-source software (main)
- Community-maintained free and open-source software (universe)
Open the software update app by pressiong windows key and type "Software" and run "Software and Updates". Then check the corresponding check boxes for the channels.
The in a terminal window, run the following commands to install the components:
- samba - expose DeepLens files to remote host which allows us to use better editor / IDE.
- tmux - maints remote session after disconnect
- libffi-dev, libsqlite3-dev - for compiling Python and running Jupyter Lab
sudo apt update
# useful tools
sudo apt install tmux samba
# libraries for software compliation
# These are for python and Jupyter Lab
sudo apt install libffi-dev libsqlite3-devNote:
- A sample tmux config can be copied to ${HOME}/.tmux.conf with mouse integration enabled as well as more convinent key binding than the default one.
cp ./misc/tmux.conf ~/.tmux.conf
The DeepLens came with Firewall turned on for security reasons. To allow easy access and development, open the following firewall ports:
- ssh
- samba
- port 8000-9999
Note: the samba alias might not be available until after the samba module is installed.
sudo ufw allow ssh
sudo ufw allow samba
sudo ufw allow 8000:9999/tcp
# to check the firewall settings, run
sudo ufw status numbered
# output:
Status: active
To Action From
-- ------ ----
[ 1] 10.105.168.217 on mlan0 DENY IN Anywhere
[ 2] 80 ALLOW IN Anywhere
[ 3] 4000 ALLOW IN Anywhere
[ 4] 22 ALLOW IN Anywhere
[ 5] 443 ALLOW IN Anywhere
[ 6] 8000:9999/tcp ALLOW IN Anywhere
[ 7] Samba ALLOW IN Anywhere
[ 8] 80 (v6) ALLOW IN Anywhere (v6)
[ 9] 4000 (v6) ALLOW IN Anywhere (v6)
[10] 22 (v6) ALLOW IN Anywhere (v6)
[11] 443 (v6) ALLOW IN Anywhere (v6)
[12] 8000:9999/tcp (v6) ALLOW IN Anywhere (v6)
[13] Samba (v6) ALLOW IN Anywhere (v6)As AWS is not providing any update to the OS and the OS was pretty much locked down, it is better not to install anything at the system level and not to pollute the dependencies for the aws_cam user. Misconfiguring the aws_cam user could lead to malfunction of network connectivity. It is better to create a new user account and install the components at the user level to avoid the risk damaging the system level dependencies.
In addition, there is not much storage space left with the built-in 16GB storage after the OS installation. To compile and install newer dependences, it is required to use some external storage such as the MicroSD card to build components from source. Thus, it is necessary use Linux native file system to store the source codes and build artifacts. It is recommended to use the 32GB SD card (that come with the DeepLense) or larger carpacity as external storage. Use the fdisk to create a native Linux partition and format it with e2fs. Then setup a persistance mount so the file paths can be predictable.
The following instructions simply delete the existing partition on the SD card and create a new Linux partition. For more sophisticated users, feel free to create the partitions on the SD card that fit for different use cases. If the SD card only contains a single partition, it can be simply using the (t)ype command instead of (d)elete and (n)ew a new partition.
# Switch to root user as we need to perform multiple actions
sudo -s
# List the block devices to find out the path of the block device
lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
mmcblk0rpmb 179:24 0 4M 0 disk
mmcblk0boot0 179:8 0 4M 1 disk
mmcblk2 179:32 0 29.7G 0 disk
└─mmcblk2p1 179:33 0 29.7G 0 part /media/aws_cam/3661-3134
mmcblk0boot1 179:16 0 4M 1 disk
mmcblk0 179:0 0 14.7G 0 disk
├─mmcblk0p2 179:2 0 14.6G 0 part /
└─mmcblk0p1 179:1 0 94M 0 part /boot/efi
# Create Linux partition with fdisk by deleting the existing Windows partition
# and create a new partition. Then set the file system type to type ID 83 Linux.
fdisk /dev/mmcblk2
Welcome to fdisk (util-linux 2.27.1).
Changes will remain in memory only, until you decide to write them.
Be careful before using the write command.
Command (m for help): p
Disk /dev/mmcblk2: 29.7 GiB, 31914983424 bytes, 62333952 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x00000000
Device Boot Start End Sectors Size Id Type
/dev/mmcblk2p1 8192 62333951 62325760 29.7G c W95 FAT32 (LBA)
Command (m for help): d
Selected partition 1
Partition 1 has been deleted.
Command (m for help): n
Partition type
p primary (0 primary, 0 extended, 4 free)
e extended (container for logical partitions)
Select (default p): p
Partition number (1-4, default 1): 1
First sector (2048-62333951, default 2048):
Last sector, +sectors or +size{K,M,G,T,P} (2048-62333951, default 62333951):
Created a new partition 1 of type 'Linux' and of size 29.7 GiB.
Command (m for help): p
Disk /dev/mmcblk2: 29.7 GiB, 31914983424 bytes, 62333952 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x00000000
Device Boot Start End Sectors Size Id Type
/dev/mmcblk2p1 2048 62333951 62331904 29.7G 83 Linux
Command (m for help): wRemember to take note of the UUID of the created partition. It is needed later to configure mount in the /etc/fstab file.
Once the partition is updated, use the mke2fs to format the partition to e2fs.
# format MicroSD card as e2fs
mkfs -t ext4 /dev/mmcblk2
Create a mount point in the root file system. For this example, it uses /mnt/ext which also can be changed as desired.
mkdir -p /mnt/ext
To automatically mount the during start up, add the following entry to /etc/fstab. Note the UUID should be replaced to the
UUID from the output of fdisk when the partition was created. The UUID can also be found with blkid command:
blkid
/dev/mmcblk0: PTUUID="7b61f71c-add4-43ab-8bc4-fc5b6db22e80" PTTYPE="gpt"
/dev/mmcblk0p1: UUID="EC97-640C" TYPE="vfat" PARTUUID="a6a77533-9f2c-4e49-9e20-fda8be912501"
/dev/mmcblk0p2: UUID="a0b62c85-25f6-4429-a9fc-b5560b0eb096" TYPE="ext4" PARTUUID="d855bc51-ed2b-426c-89cf-d149ca740e0c"
/dev/mmcblk2: PTTYPE="dos"
/dev/mmcblk2p1: UUID="cebcb1b8-927d-441e-a899-728c541d80c1" TYPE="ext2"```Add the following entry (update the UUID accordingly) to /etc/fstab. Note the nofail option which allows the OS to
contiue boot sequence in case the SD card is not present or damaged.
UUID=cebcb1b8-927d-441e-a899-728c541d80c1 /mnt/ext ext4 defaults,nofail 0 2
Mount the partition before proceed to the next step:
mount -a
# double check to make sure the partition is mounted
df -h
Filesystem Size Used Avail Use% Mounted on
udev 3.8G 0 3.8G 0% /dev
tmpfs 781M 39M 742M 5% /run
/dev/mmcblk0p2 15G 7.1G 6.5G 53% /
tmpfs 3.9G 336K 3.9G 1% /dev/shm
tmpfs 5.0M 4.0K 5.0M 1% /run/lock
tmpfs 3.9G 0 3.9G 0% /sys/fs/cgroup
/dev/mmcblk0p1 99M 50M 50M 50% /boot/efi
/dev/mmcblk2p1 30G 12G 17G 43% /mnt/ext
tmpfs 781M 60K 781M 1% /run/user/1002As previously mentioned, it is better to use a separate user so not to pollute the existing system library and causing
issues. This instruction use adduser wrapper script to create an OS user dwu with home directory set to /mnt/ext/dwu
, it is ok to chose any user name and other directory path on the SD card as desired.
# Assuming we are adding a new data working user `dwu`
adduser dwu --home /mnt/ext/dwu
# Add user to user groups
usermod -a -G adm dwu
usermod -a -G sudo dwu
usermod -a -G aws_cam dwu
The default SAMBA configuration doesn't expose users' home directory. Uncomment the following sections in /etc/samba/samba.conf
[homes]
comment = Home Directories
browseable = no
Then restart the SAMBA services
sudo systemctl restart smbd.service
sudo systemctl restart nmbd.serviceNow the user's home can be access from remote computer.
This repo contains some convinent scripts and configuration and can be cloned to the user account just created to make the following steps simpler.
# ssh to DeepLens as the recently created user, which should have the home directory set to the SD card
# clone this repo from github.com
git clone https://github.com/danielwu68/aws-deeplens-hack.git
cd aws-deeplens-hack
source ./scripts/env-setupAlso add the following to ${HOME}/.profile
source ${HOME}/aws-deeplens-hack/bin/env-setup
Pip in Python requires ssl library to acces the repo. Thus, it is required to build and install the OpenSSL library. The following instructions
assume the necessary environment variables were setup properly via sourcing the env-setup mentioned above. Also the NodeJS is required for
running Jupyter Lab. Download and extract the package with the commands listed below.
Note:
- At the time of writing, the followings are the most recent and compatible versions:
- OpenSSL: openss-3.0.11 (Long Term Support - LTS) and openssl-3.1.3 are available. This instruction will use the LTS openss-3.011.
- NodeJS: node-v16.20.2 which can be found at https://nodejs.org/en/download/releases
- Python: 3.11.5 can be found at https://www.python.org/ftp/python/
mkdir ~/build ~/install ~/Downloads
cd ~/Downloads
curl -kO https://www.openssl.org/source/openssl-3.0.11.tar.gz
curl -O https://www.python.org/ftp/python/3.11.5/Python-3.11.5.tar.xz
curl -O https://nodejs.org/download/release/v16.20.2/node-v16.20.2-linux-x64.tar.xz
# Extract directly into install folder for the ones come with binary
cd ~/install
tar xvf ~/Downloads/node-v16.20.2-linux-x64.tar.xz
# Extract into build folder for the ones need to be built
cd ~/build
tar xvf ~/Downloads/openssl-3.0.11.tar.gz
tar xvf ~/Downloads/Python-3.11.5.tar.xzcd ~/build/openssl-3.0.11
# configure the build with custom target directories
./config --prefix=${OPENSSL_HOME} --openssldir=${OPENSSL_HOME}/ssl
# start the build process
make
# optionally after successful build, run validation tests
make test
# install the artifects to target directory
make installcd ~/build/Python-3.11.5To enable SSL support, which is needed by pip, uncomment the following lines in `Modules/Setting``.
Note, this is specific to Python 3.11,5. It can be different for different Python version:
# To statically link OpenSSL:
_ssl _ssl.c $(OPENSSL_INCLUDES) $(OPENSSL_LDFLAGS) \
-l:libssl.a -Wl,--exclude-libs,libssl.a \
-l:libcrypto.a -Wl,--exclude-libs,libcrypto.a
_hashlib _hashopenssl.c $(OPENSSL_INCLUDES) $(OPENSSL_LDFLAGS) \
-l:libcrypto.a -Wl,--exclude-libs,libcrypto.a
# configure the build with custom target directory
./configure --prefix=${PYTHON_HOME}
# build and install
make \
OPENSSL_INCLUDES=-I${OPENSSL_HOME}/include \
OPENSSL_LDFLAGS="-L ${OPENSSL_LIB} -Wl,-rpath,${OPENSSL_LIB}" \
installThis example uses pipenv to create the Python virtual environment. If desired, other Python virtual environment can be used.
Please refer to specific Python virtual environment took kit for details.
pip3 install pipenv
cd ${HOME}/aws-deeplens-hack/jupyter
pipenv shell --python 3.11
pipenv install -r requirements-dev.txtFor convinence sake, use the tmux to start the Jupyter Lab. Use the URL display on the output to connect to the Jupyter Lab server. Please modify the URL so that it reflects the correct hostname or IP address if connect from remote machine.
tmux
cd ${HOME}/aws-deeplens-hack/jupyter
./start-server.sh