[Mayan EDMS: 2289] Docker image for Mayan EDMS NG available

[e***@gmail.com]

The Docker image is available from the Docker Hub at:
https://hub.docker.com/r/mayanedmsng/mayanedmsng/

Every procedure is the same as with the upstream version with the only
change that the repository reference must be mayanedmsng/mayanedmsng
instead of mayanedms/mayanedms.

Once change this image has is that is includes a fourth worker to handle
the OCR and parsing queues. This worker is run with the lowest possible
process priority (19). Therefore OCR results might take longer than with
the upstream image. The advantage is that the OCR will not slowdown the
execution of the rest of the program. The total image layers were reduced
from 54 to 44. Total image size remained the same 1.32GB and 490.8MB
compressed (the actual download).

Other changes according to Docker's best practices were made (pinning
installs to a specific version, not running dist-upgrade, etc).

The repository for this image is here:
https://gitlab.com/e.riggs/mayan-edms-docker/commits/master

I will continue working on getting the size of the image down. I'll also
keep trying using other base images instead of Ubuntu and or update to a
more recent version of Ubuntu, as time allows.

[Lin Pro]

Post by e***@gmail.com
https://hub.docker.com/r/mayanedmsng/mayanedmsng/

This is great. Thank you for your contribution.
I have just recently learned that:
"In recent releases, a few things have happened in the Docker world. Swarm
mode got integrated into the Docker Engine in 1.12, and has brought with it
several new tools. Among others, it’s possible to make use of
docker-compose.yml files to bring up stacks of Docker containers, without
having to install Docker Compose."

"The command is called docker stack, and it looks exactly the same to
docker-compose. Here’s a comparison:
$ docker-compose -f docker-compose up
$ docker stack deploy -c docker-compose.yml somestackname"

The side effect is that the YML file has to be of Version 3, not Version
"2" or "2.1".
"The Docker Stack functionality, is included with the Docker engine. You
don’t need to install additional packages (i.e docker-compose, a python
project) to use it. Deploying docker stacks is part of the swarm mode. It
supports the same kinds of compose files, but the handling happens in Go
code, inside of the Docker Engine"

Do you think that it would be worth pursuing in the near future that
MAYAN-NG should adhere to that version 3 YML specification and be able to
be brought up with "docker stack deploy"?

Regards
Lin

[l***@gmail.com]

I'll leave that up to Eric out resident Docker expert. But if the file
format is the same I don't see much problem with supporting that too. One
idea we had a while back was to created a deployer for Mayan that would
take care of installing Docker Engine, pull the Mayan image and deploy the
container. The script would also have data backup and restore commands.

Post by Lin Pro

Post by e***@gmail.com
https://hub.docker.com/r/mayanedmsng/mayanedmsng/

This is great. Thank you for your contribution.
"In recent releases, a few things have happened in the Docker world. Swarm
mode got integrated into the Docker Engine in 1.12, and has brought with it
several new tools. Among others, it’s possible to make use of
docker-compose.yml files to bring up stacks of Docker containers, without
having to install Docker Compose."
"The command is called docker stack, and it looks exactly the same to
$ docker-compose -f docker-compose up
$ docker stack deploy -c docker-compose.yml somestackname"
The side effect is that the YML file has to be of Version 3, not Version
"2" or "2.1".
"The Docker Stack functionality, is included with the Docker engine. You
don’t need to install additional packages (i.e docker-compose, a python
project) to use it. Deploying docker stacks is part of the swarm mode. It
supports the same kinds of compose files, but the handling happens in Go
code, inside of the Docker Engine"
Do you think that it would be worth pursuing in the near future that
MAYAN-NG should adhere to that version 3 YML specification and be able to
be brought up with "docker stack deploy"?
Regards
Lin

[RW Shore]

If you're talking a Windows deployer, fine, since I don't use M$ stuff.
I've got 8 services running in a Docker swarm, and I use Docker from the
fedora repos, so a deployer for my environment would be unusable.

Post by l***@gmail.com
I'll leave that up to Eric out resident Docker expert. But if the file
format is the same I don't see much problem with supporting that too. One
idea we had a while back was to created a deployer for Mayan that would
take care of installing Docker Engine, pull the Mayan image and deploy the
container. The script would also have data backup and restore commands.

Post by Lin Pro

Post by e***@gmail.com
https://hub.docker.com/r/mayanedmsng/mayanedmsng/

This is great. Thank you for your contribution.
"In recent releases, a few things have happened in the Docker world.
Swarm mode got integrated into the Docker Engine in 1.12, and has brought
with it several new tools. Among others, it’s possible to make use of
docker-compose.yml files to bring up stacks of Docker containers, without
having to install Docker Compose."
"The command is called docker stack, and it looks exactly the same to
$ docker-compose -f docker-compose up
$ docker stack deploy -c docker-compose.yml somestackname"
The side effect is that the YML file has to be of Version 3, not Version
"2" or "2.1".
"The Docker Stack functionality, is included with the Docker engine. You
don’t need to install additional packages (i.e docker-compose, a python
project) to use it. Deploying docker stacks is part of the swarm mode. It
supports the same kinds of compose files, but the handling happens in Go
code, inside of the Docker Engine"
Do you think that it would be worth pursuing in the near future that
MAYAN-NG should adhere to that version 3 YML specification and be able to
be brought up with "docker stack deploy"?
Regards
Lin

--
---
You received this message because you are subscribed to the Google Groups
"Mayan EDMS" group.
To unsubscribe from this group and stop receiving emails from it, send an
For more options, visit https://groups.google.com/d/optout.

[e***@gmail.com]

Haven't worked with version 3 of docker compose files. Sounds interesting,
I'll give it a go.

Post by l***@gmail.com
I'll leave that up to Eric out resident Docker expert. But if the file
format is the same I don't see much problem with supporting that too. One
idea we had a while back was to created a deployer for Mayan that would
take care of installing Docker Engine, pull the Mayan image and deploy the
container. The script would also have data backup and restore commands.

Post by Lin Pro

Post by e***@gmail.com
https://hub.docker.com/r/mayanedmsng/mayanedmsng/

This is great. Thank you for your contribution.
"In recent releases, a few things have happened in the Docker world.
Swarm mode got integrated into the Docker Engine in 1.12, and has brought
with it several new tools. Among others, it’s possible to make use of
docker-compose.yml files to bring up stacks of Docker containers, without
having to install Docker Compose."
"The command is called docker stack, and it looks exactly the same to
$ docker-compose -f docker-compose up
$ docker stack deploy -c docker-compose.yml somestackname"
The side effect is that the YML file has to be of Version 3, not Version
"2" or "2.1".
"The Docker Stack functionality, is included with the Docker engine. You
don’t need to install additional packages (i.e docker-compose, a python
project) to use it. Deploying docker stacks is part of the swarm mode. It
supports the same kinds of compose files, but the handling happens in Go
code, inside of the Docker Engine"
Do you think that it would be worth pursuing in the near future that
MAYAN-NG should adhere to that version 3 YML specification and be able to
be brought up with "docker stack deploy"?
Regards
Lin

[l***@gmail.com]

Awesome work! Thanks Eric.

Post by e***@gmail.com
https://hub.docker.com/r/mayanedmsng/mayanedmsng/
Every procedure is the same as with the upstream version with the only
change that the repository reference must be mayanedmsng/mayanedmsng
instead of mayanedms/mayanedms.
Once change this image has is that is includes a fourth worker to handle
the OCR and parsing queues. This worker is run with the lowest possible
process priority (19). Therefore OCR results might take longer than with
the upstream image. The advantage is that the OCR will not slowdown the
execution of the rest of the program. The total image layers were reduced
from 54 to 44. Total image size remained the same 1.32GB and 490.8MB
compressed (the actual download).
Other changes according to Docker's best practices were made (pinning
installs to a specific version, not running dist-upgrade, etc).
https://gitlab.com/e.riggs/mayan-edms-docker/commits/master
I will continue working on getting the size of the image down. I'll also
keep trying using other base images instead of Ubuntu and or update to a
more recent version of Ubuntu, as time allows.

[Raul Garcia Sanchez]

I just tried to start run that image. However, it stucks in a rebooting
loop.
Do you know what I can do?
Is the image working?

Thanks

Post by e***@gmail.com
https://hub.docker.com/r/mayanedmsng/mayanedmsng/
Every procedure is the same as with the upstream version with the only
change that the repository reference must be mayanedmsng/mayanedmsng
instead of mayanedms/mayanedms.
Once change this image has is that is includes a fourth worker to handle
the OCR and parsing queues. This worker is run with the lowest possible
process priority (19). Therefore OCR results might take longer than with
the upstream image. The advantage is that the OCR will not slowdown the
execution of the rest of the program. The total image layers were reduced
from 54 to 44. Total image size remained the same 1.32GB and 490.8MB
compressed (the actual download).
Other changes according to Docker's best practices were made (pinning
installs to a specific version, not running dist-upgrade, etc).
https://gitlab.com/e.riggs/mayan-edms-docker/commits/master
I will continue working on getting the size of the image down. I'll also
keep trying using other base images instead of Ubuntu and or update to a
more recent version of Ubuntu, as time allows.

[Raul]

I was able to get it working by building it by myself with docker-compose.
Nice jobs guys :)

Post by Raul Garcia Sanchez
I just tried to start run that image. However, it stucks in a rebooting
loop.
Do you know what I can do?
Is the image working?
Thanks

Post by e***@gmail.com
https://hub.docker.com/r/mayanedmsng/mayanedmsng/
Every procedure is the same as with the upstream version with the only
change that the repository reference must be mayanedmsng/mayanedmsng
instead of mayanedms/mayanedms.
Once change this image has is that is includes a fourth worker to handle
the OCR and parsing queues. This worker is run with the lowest possible
process priority (19). Therefore OCR results might take longer than with
the upstream image. The advantage is that the OCR will not slowdown the
execution of the rest of the program. The total image layers were reduced
from 54 to 44. Total image size remained the same 1.32GB and 490.8MB
compressed (the actual download).
Other changes according to Docker's best practices were made (pinning
installs to a specific version, not running dist-upgrade, etc).
https://gitlab.com/e.riggs/mayan-edms-docker/commits/master
I will continue working on getting the size of the image down. I'll also
keep trying using other base images instead of Ubuntu and or update to a
more recent version of Ubuntu, as time allows.

[e***@gmail.com]

Awesome! Make sure you use Docker named volumes for the data. It is a
current limitation that I hope to correct soon.

Post by Raul
I was able to get it working by building it by myself with docker-compose.
Nice jobs guys :)

Post by Raul Garcia Sanchez
I just tried to start run that image. However, it stucks in a rebooting
loop.
Do you know what I can do?
Is the image working?
Thanks

Post by e***@gmail.com
https://hub.docker.com/r/mayanedmsng/mayanedmsng/
Every procedure is the same as with the upstream version with the only
change that the repository reference must be mayanedmsng/mayanedmsng
instead of mayanedms/mayanedms.
Once change this image has is that is includes a fourth worker to handle
the OCR and parsing queues. This worker is run with the lowest possible
process priority (19). Therefore OCR results might take longer than with
the upstream image. The advantage is that the OCR will not slowdown the
execution of the rest of the program. The total image layers were reduced
from 54 to 44. Total image size remained the same 1.32GB and 490.8MB
compressed (the actual download).
Other changes according to Docker's best practices were made (pinning
installs to a specific version, not running dist-upgrade, etc).
https://gitlab.com/e.riggs/mayan-edms-docker/commits/master
I will continue working on getting the size of the image down. I'll also
keep trying using other base images instead of Ubuntu and or update to a
more recent version of Ubuntu, as time allows.

[David Kornahrens]

Hey Eric,
Nice work. I just finished looking over some of your commits in the
gitlab. We currently are running the latest MayanEDMS and have a rather
large inventory loaded in. Couple of questions.

1. Are we able to upgrade to NG without losing the data? Anyway to
migrate the data?
2. We have offloading media to S3 using the Django offload method from
Mayan.
3. We haven't heard from Roberto in awhile, are you going to be taking
over the development of Mayan?

Thanks,
David

Post by e***@gmail.com
Awesome! Make sure you use Docker named volumes for the data. It is a
current limitation that I hope to correct soon.

Post by Raul
I was able to get it working by building it by myself with docker-compose.
Nice jobs guys :)

Post by Raul Garcia Sanchez
I just tried to start run that image. However, it stucks in a rebooting
loop.
Do you know what I can do?
Is the image working?
Thanks

Post by e***@gmail.com
https://hub.docker.com/r/mayanedmsng/mayanedmsng/
Every procedure is the same as with the upstream version with the only
change that the repository reference must be mayanedmsng/mayanedmsng
instead of mayanedms/mayanedms.
Once change this image has is that is includes a fourth worker to
handle the OCR and parsing queues. This worker is run with the lowest
possible process priority (19). Therefore OCR results might take longer
than with the upstream image. The advantage is that the OCR will not
slowdown the execution of the rest of the program. The total image layers
were reduced from 54 to 44. Total image size remained the same 1.32GB and
490.8MB compressed (the actual download).
Other changes according to Docker's best practices were made (pinning
installs to a specific version, not running dist-upgrade, etc).
https://gitlab.com/e.riggs/mayan-edms-docker/commits/master
I will continue working on getting the size of the image down. I'll
also keep trying using other base images instead of Ubuntu and or update to
a more recent version of Ubuntu, as time allows.

[e***@gmail.com]

Hi.

There are database changes between our release and version 2.7.3 of Mayan
upstream. The database changes in our release correspond to unpublished
changes already in Mayan upstream. But, if you migrate your database and
Mayan EDMS version 2.8 doesn't pull all of our changes you could end up
with an incompatible database. If you have a large number of critical
documents I suggest you wait until Mayan upstream revise and merge our
changes.

1. The data migration will happen automatically. It is performed by the
Docker image itself when it encounters and existing database.
2. Both Docker images (the official one and ours) support S3 object storage
out of the box.
3. It is not our intention to take over the project. That said, if Roberto
asks us to, I think we would. I already manage several department
installations of Mayan in our company as part of my day job. Maintaining
the Mayan project would not be outside of what I and Michael already do
daily. We try not to bother Roberto much, things in Puerto Rico are very
bad. I will not go into details, it is not my place. Some of what they are
dealing with =>
http://www.foxnews.com/politics/2018/03/09/storm-dealt-puerto-rico-knockout-blow-after-decades-fiscal-failure-corruption.html

Post by David Kornahrens
Hey Eric,
Nice work. I just finished looking over some of your commits in the
gitlab. We currently are running the latest MayanEDMS and have a rather
large inventory loaded in. Couple of questions.
1. Are we able to upgrade to NG without losing the data? Anyway to
migrate the data?
2. We have offloading media to S3 using the Django offload method from
Mayan.
3. We haven't heard from Roberto in awhile, are you going to be taking
over the development of Mayan?
Thanks,
David

Post by e***@gmail.com
Awesome! Make sure you use Docker named volumes for the data. It is a
current limitation that I hope to correct soon.

Post by Raul
I was able to get it working by building it by myself with
docker-compose.
Nice jobs guys :)

Post by Raul Garcia Sanchez
I just tried to start run that image. However, it stucks in a rebooting
loop.
Do you know what I can do?
Is the image working?
Thanks

Post by e***@gmail.com
https://hub.docker.com/r/mayanedmsng/mayanedmsng/
Every procedure is the same as with the upstream version with the only
change that the repository reference must be mayanedmsng/mayanedmsng
instead of mayanedms/mayanedms.
Once change this image has is that is includes a fourth worker to
handle the OCR and parsing queues. This worker is run with the lowest
possible process priority (19). Therefore OCR results might take longer
than with the upstream image. The advantage is that the OCR will not
slowdown the execution of the rest of the program. The total image layers
were reduced from 54 to 44. Total image size remained the same 1.32GB and
490.8MB compressed (the actual download).
Other changes according to Docker's best practices were made (pinning
installs to a specific version, not running dist-upgrade, etc).
https://gitlab.com/e.riggs/mayan-edms-docker/commits/master
I will continue working on getting the size of the image down. I'll
also keep trying using other base images instead of Ubuntu and or update to
a more recent version of Ubuntu, as time allows.

[Raul]

Hi Eric,

yeah it seems like /var/lib/docker/volumes/mayan_data/_data is not being
used currently...
Is this what you meant? How can I fix this?

Post by e***@gmail.com
Awesome! Make sure you use Docker named volumes for the data. It is a
current limitation that I hope to correct soon.

Post by Raul
I was able to get it working by building it by myself with docker-compose.
Nice jobs guys :)

Post by Raul Garcia Sanchez
I just tried to start run that image. However, it stucks in a rebooting
loop.
Do you know what I can do?
Is the image working?
Thanks

Post by e***@gmail.com
https://hub.docker.com/r/mayanedmsng/mayanedmsng/
Every procedure is the same as with the upstream version with the only
change that the repository reference must be mayanedmsng/mayanedmsng
instead of mayanedms/mayanedms.
Once change this image has is that is includes a fourth worker to
handle the OCR and parsing queues. This worker is run with the lowest
possible process priority (19). Therefore OCR results might take longer
than with the upstream image. The advantage is that the OCR will not
slowdown the execution of the rest of the program. The total image layers
were reduced from 54 to 44. Total image size remained the same 1.32GB and
490.8MB compressed (the actual download).
Other changes according to Docker's best practices were made (pinning
installs to a specific version, not running dist-upgrade, etc).
https://gitlab.com/e.riggs/mayan-edms-docker/commits/master
I will continue working on getting the size of the image down. I'll
also keep trying using other base images instead of Ubuntu and or update to
a more recent version of Ubuntu, as time allows.

[Raul]

Hi Eric,

it seems like everything is the correct volume except for the db.sqllite3
file.
Do you know how I can fix this?

Post by e***@gmail.com
Awesome! Make sure you use Docker named volumes for the data. It is a
current limitation that I hope to correct soon.

Post by Raul
I was able to get it working by building it by myself with docker-compose.
Nice jobs guys :)

Post by Raul Garcia Sanchez
I just tried to start run that image. However, it stucks in a rebooting
loop.
Do you know what I can do?
Is the image working?
Thanks

Post by e***@gmail.com
https://hub.docker.com/r/mayanedmsng/mayanedmsng/
Every procedure is the same as with the upstream version with the only
change that the repository reference must be mayanedmsng/mayanedmsng
instead of mayanedms/mayanedms.
Once change this image has is that is includes a fourth worker to
handle the OCR and parsing queues. This worker is run with the lowest
possible process priority (19). Therefore OCR results might take longer
than with the upstream image. The advantage is that the OCR will not
slowdown the execution of the rest of the program. The total image layers
were reduced from 54 to 44. Total image size remained the same 1.32GB and
490.8MB compressed (the actual download).
Other changes according to Docker's best practices were made (pinning
installs to a specific version, not running dist-upgrade, etc).
https://gitlab.com/e.riggs/mayan-edms-docker/commits/master
I will continue working on getting the size of the image down. I'll
also keep trying using other base images instead of Ubuntu and or update to
a more recent version of Ubuntu, as time allows.

[e***@gmail.com]

The db.sqlite3 file is created and used only when using SQLite. Since you
used the docker compose, the database backend being used is PostgreSQL.

Execute: docker ps

You should have two containers. One for the project and another for the
database.

Post by Raul
Hi Eric,
it seems like everything is the correct volume except for the db.sqllite3
file.
Do you know how I can fix this?

Post by e***@gmail.com
Awesome! Make sure you use Docker named volumes for the data. It is a
current limitation that I hope to correct soon.

Post by Raul
I was able to get it working by building it by myself with
docker-compose.
Nice jobs guys :)

Post by Raul Garcia Sanchez
I just tried to start run that image. However, it stucks in a rebooting
loop.
Do you know what I can do?
Is the image working?
Thanks

Post by e***@gmail.com
https://hub.docker.com/r/mayanedmsng/mayanedmsng/
Every procedure is the same as with the upstream version with the only
change that the repository reference must be mayanedmsng/mayanedmsng
instead of mayanedms/mayanedms.
Once change this image has is that is includes a fourth worker to
handle the OCR and parsing queues. This worker is run with the lowest
possible process priority (19). Therefore OCR results might take longer
than with the upstream image. The advantage is that the OCR will not
slowdown the execution of the rest of the program. The total image layers
were reduced from 54 to 44. Total image size remained the same 1.32GB and
490.8MB compressed (the actual download).
Other changes according to Docker's best practices were made (pinning
installs to a specific version, not running dist-upgrade, etc).
https://gitlab.com/e.riggs/mayan-edms-docker/commits/master
I will continue working on getting the size of the image down. I'll
also keep trying using other base images instead of Ubuntu and or update to
a more recent version of Ubuntu, as time allows.

[Raul]

You are right thanks :)

Post by e***@gmail.com
The db.sqlite3 file is created and used only when using SQLite. Since you
used the docker compose, the database backend being used is PostgreSQL.
Execute: docker ps
You should have two containers. One for the project and another for the
database.

Post by Raul
Hi Eric,
it seems like everything is the correct volume except for the db.sqllite3
file.
Do you know how I can fix this?

Post by e***@gmail.com
Awesome! Make sure you use Docker named volumes for the data. It is a
current limitation that I hope to correct soon.

Post by Raul
I was able to get it working by building it by myself with
docker-compose.
Nice jobs guys :)

Post by Raul Garcia Sanchez
I just tried to start run that image. However, it stucks in a
rebooting loop.
Do you know what I can do?
Is the image working?
Thanks

Post by e***@gmail.com
https://hub.docker.com/r/mayanedmsng/mayanedmsng/
Every procedure is the same as with the upstream version with the
only change that the repository reference must be mayanedmsng/mayanedmsng
instead of mayanedms/mayanedms.
Once change this image has is that is includes a fourth worker to
handle the OCR and parsing queues. This worker is run with the lowest
possible process priority (19). Therefore OCR results might take longer
than with the upstream image. The advantage is that the OCR will not
slowdown the execution of the rest of the program. The total image layers
were reduced from 54 to 44. Total image size remained the same 1.32GB and
490.8MB compressed (the actual download).
Other changes according to Docker's best practices were made (pinning
installs to a specific version, not running dist-upgrade, etc).
https://gitlab.com/e.riggs/mayan-edms-docker/commits/master
I will continue working on getting the size of the image down. I'll
also keep trying using other base images instead of Ubuntu and or update to
a more recent version of Ubuntu, as time allows.