OliPelz · June 5, 2019 12:14
diff --git a/1_How to create bigger - huge Docker images (>100gb) in CentOS 7.x (using Docker 1.12.5) b/1_How to create bigger - huge Docker images (>100gb) in CentOS 7.x (using Docker 1.12.5)
 The default Docker daemon on CentOS 7.x from the official package repository has two very annoying limitations when trying to create bigger images.

 First is the default base image is 10gb space and second there is another barrier with the thin layer which is set to 100gb.

 I needed a Docker image containing a big database which was about 120gb.

 When I COPY'ed the huge database in the Dockerfile to build the image, I got a 

 No space left on device

 error when I was hitting the 10gb barrier. This can be controlled by the Docker Base device size which is set by default to
 10gb and can be controlled by the dockerd daemon parameter dm.basesize.

 After using a systemd drop-in config file with the changed dockerd parameter it went past this barrier but again after hitting 
 a COPY amount of 100gb I got more weired errors:

 ApplyLayer exit status 1 stdout:  stderr: input/output error

 and
 docker: Error response from daemon: devmapper: Thin Pool has 0 free data blocks which is less than minimum required

 so I learned that I had to expand the thin layer. Sadly, you have to wipe your Docker data dir first :( in order to do so.

 So here I provide the files and stuff I did in the end to make things work!

 Please note that in this example I have changed my default Docker data dir to /data/docker-root from /var/lib/docker because of free space issues on my server!
diff --git a/2_dockerd_modifications_on_systemd b/2_dockerd_modifications_on_systemd
 mkdir -p /etc/systemd/system/docker.service.d
 echo "[Service]
 ExecStart=
 ExecStart=/usr/bin/dockerd --storage-opt dm.basesize=200G -g /data/docker-root" > /etc/systemd/system/docker.service.d/docker.conf
 # reload 
 systemctl daemon-reload
 systemctl restart docker
diff --git a/3_expand_thin_layer b/3_expand_thin_layer
 # first wipe off your docker data dir - handle with care - i am not responsible for your data loss
 # please note: this is not the default docker data dir ;)
 systemctl stop docker
 rm -rf /data/docker-root/*
 mkdir /data/docker-root/devicemapper/devicemapper -p
 # create a new thin layer of 200GB which is double the size of default Docker installation on CentOS 7
 dd if=/dev/zero of=/data/docker-root/devicemapper/devicemapper/data bs=1G count=0 seek=200
 # this initalizes everything
 systemctl start docker
 docker info | grep -i space
 # output should now be 200gb instead of 100gb


diff --git a/4_the_end b/4_the_end
 Now everything works and I was able to create huge images >100gb!!!

 Here are some useful links I used to accomplish the above:

 https://linuxconfig.org/how-to-move-docker-s-default-var-lib-docker-to-another-directory-on-ubuntu-debian-linux
 http://www.projectatomic.io/blog/2016/03/daemon_option_basedevicesize/
 https://jpetazzo.github.io/2014/01/29/docker-device-mapper-resize/
	The default Docker daemon on CentOS 7.x from the official package repository has two very annoying limitations when trying to create bigger images.

	First is the default base image is 10gb space and second there is another barrier with the thin layer which is set to 100gb.

	I needed a Docker image containing a big database which was about 120gb.

	When I COPY'ed the huge database in the Dockerfile to build the image, I got a

	No space left on device

	error when I was hitting the 10gb barrier. This can be controlled by the Docker Base device size which is set by default to
	10gb and can be controlled by the dockerd daemon parameter dm.basesize.

	After using a systemd drop-in config file with the changed dockerd parameter it went past this barrier but again after hitting
	a COPY amount of 100gb I got more weired errors:

	ApplyLayer exit status 1 stdout: stderr: input/output error

	and
	docker: Error response from daemon: devmapper: Thin Pool has 0 free data blocks which is less than minimum required

	so I learned that I had to expand the thin layer. Sadly, you have to wipe your Docker data dir first :( in order to do so.

	So here I provide the files and stuff I did in the end to make things work!

	Please note that in this example I have changed my default Docker data dir to /data/docker-root from /var/lib/docker because of free space issues on my server!
	mkdir -p /etc/systemd/system/docker.service.d
	echo "[Service]
	ExecStart=
	ExecStart=/usr/bin/dockerd --storage-opt dm.basesize=200G -g /data/docker-root" > /etc/systemd/system/docker.service.d/docker.conf
	# reload
	systemctl daemon-reload
	systemctl restart docker
	# first wipe off your docker data dir - handle with care - i am not responsible for your data loss
	# please note: this is not the default docker data dir ;)
	systemctl stop docker
	rm -rf /data/docker-root/*
	mkdir /data/docker-root/devicemapper/devicemapper -p
	# create a new thin layer of 200GB which is double the size of default Docker installation on CentOS 7
	dd if=/dev/zero of=/data/docker-root/devicemapper/devicemapper/data bs=1G count=0 seek=200
	# this initalizes everything
	systemctl start docker
	docker info \| grep -i space
	# output should now be 200gb instead of 100gb
	Now everything works and I was able to create huge images >100gb!!!

	Here are some useful links I used to accomplish the above:

	https://linuxconfig.org/how-to-move-docker-s-default-var-lib-docker-to-another-directory-on-ubuntu-debian-linux
	http://www.projectatomic.io/blog/2016/03/daemon_option_basedevicesize/
	https://jpetazzo.github.io/2014/01/29/docker-device-mapper-resize/