Running Release Fails after "set include_erts: false"

[Faolain]

I followed the guide here at https://shovik.com/blog/8-deploying-phoenix-apps-with-docker and it was advised to not bundle the erts files, that is, set include_erts: false in the rel/config.exs. I was both able to run mix docker.build and mix docker.release but when it comes time to running the container it errors out stating

{"init terminating in do_boot",{load_failed,[file_io_server,heart,lists,error_handler,erl_eval,code_server,erl_lint,application_master,gen_server,kernel,file_server,application,filename,gen_event,gen,supervisor,code,ets,erl_parse,proc_lib,file,application_controller,error_logger]}}
init terminating in do_boot ()

Crash dump is being written to: erl_crash.dump...done

Shutting down..
Node is not running!

If I then instead set to include erts to true
set include_erts: true

then it is able to run the container/initiate the phoenix server. If it helps at all I'm running this on OSX without any other changes to the standard included Dockerfile build and release files.

[tlvenn]

Seems related to this issue: bitwalker/distillery#20

It should be fixed with Distillery 1.3.3 which is used by the latest version of mix_docker

[tlvenn]

@Faolain can you tell us if you still have the issue with the latest version of mix_docker ?

[Faolain]

Thanks for the fast response unfortunately although I updated there seems to be another issue before even being able to address the erts boolean. I'm not sure what has changed (is it the alpine image itself?) but now I am unable to even run mix docker.build at all even with set include_erts: false.

Step 10/12 : RUN mix do deps.get, deps.compile
 ---> Running in a6ec85f0ac7c
* Getting stripity_stripe (https://github.com/code-corps/stripity_stripe.git)
remote: Counting objects: 1873, done.
remote: Compressing objects: 100% (24/24), done.
remote: Total 1873 (delta 2), reused 0 (delta 0), pack-reused 1848
Receiving objects: 100% (1873/1873), 502.20 KiB | 0 bytes/s, done.
Resolving deltas: 100% (1112/1112), done.

06:07:30.756 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired


06:07:30.756 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired


06:07:30.756 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired


06:07:30.757 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired


06:07:30.764 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired


06:07:30.775 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired


06:07:30.790 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired


06:07:30.806 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired


06:07:30.810 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired


06:07:30.826 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired

Failed to fetch record for 'mix_docker' from registry (using cache)

06:07:30.839 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired


06:07:30.845 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired

{:failed_connect, [{:to_address, {'repo.hex.pm', 443}}, {:inet, [:inet], {:tls_alert, 'certificate expired'}}]}
Failed to fetch record for 'phoenix' from registry (using cache)
{:failed_connect, [{:to_address, {'repo.hex.pm', 443}}, {:inet, [:inet], {:tls_alert, 'certificate expired'}}]}
Failed to fetch record for 'postgrex' from registry (using cache)
{:failed_connect, [{:to_address, {'repo.hex.pm', 443}}, {:inet, [:inet], {:tls_alert, 'certificate expired'}}]}

06:07:30.854 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired

Failed to fetch record for 'guardian' from registry (using cache)
{:failed_connect, [{:to_address, {'repo.hex.pm', 443}}, {:inet, [:inet], {:tls_alert, 'certificate expired'}}]}

06:07:30.874 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired

Failed to fetch record for 'plug' from registry (using cache)
{:failed_connect, [{:to_address, {'repo.hex.pm', 443}}, {:inet, [:inet], {:tls_alert, 'certificate expired'}}]}

06:07:30.901 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired

Failed to fetch record for 'cors_plug' from registry (using cache)

06:07:30.908 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired

{:failed_connect, [{:to_address, {'repo.hex.pm', 443}}, {:inet, [:inet], {:tls_alert, 'certificate expired'}}]}
Failed to fetch record for 'ecto' from registry (using cache)
{:failed_connect, [{:to_address, {'repo.hex.pm', 443}}, {:inet, [:inet], {:tls_alert, 'certificate expired'}}]}

06:07:30.912 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired

Failed to fetch record for 'mochiweb_html' from registry (using cache)
{:failed_connect, [{:to_address, {'repo.hex.pm', 443}}, {:inet, [:inet], {:tls_alert, 'certificate expired'}}]}
Failed to check for new Hex version

06:07:30.929 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired


06:07:30.932 [error] SSL: :certify: ssl_handshake.erl:1605:Fatal error: certificate expired

** (Mix) Unknown package mochiweb_html in lockfile
{:failed_connect, [{:to_address, {'repo.hex.pm', 443}}, {:inet, [:inet], {:tls_alert, 'certificate expired'}}]}
The command '/bin/sh -c mix do deps.get, deps.compile' returned a non-zero code: 1
** (MatchError) no match of right hand side value: {%IO.Stream{device: :standard_io, line_or_bytes: :line, raw: false}, 1}
    lib/mix_docker.ex:179: MixDocker.system!/2
    lib/mix_docker.ex:20: MixDocker.build/1
    (mix) lib/mix/task.ex:296: Mix.Task.run_task/3
    (mix) lib/mix/cli.ex:58: Mix.CLI.run_task/2
    (elixir) lib/code.ex:363: Code.require_file/2

[teamon]

@Faolain This is definitely not a mix_docker issues - I just got the same SSL error in one project that is still using quite old version of mix_docker and custom Dockerfiles that hasn't been changes for a while.

I don't know yet what is causing this. I'll let you know as soon as I find anything.

[teamon]

In my case this was a clock sync issue:

Before fix

⌘ ~ λ date -u
Fri Apr 14 11:31:40 UTC 2017
⌘ ~ λ docker run --rm -it bitwalker/alpine-erlang:6.1 date -u
Thu Apr 13 20:37:38 UTC 2017

The fix - source

⌘ ~ λ docker run -it --rm --privileged --pid=host debian nsenter -t 1 -m -u -n -i date -u $(date -u +%m%d%H%M%Y)

After fix

⌘ ~ λ date -u
Fri Apr 14 11:48:37 UTC 2017
⌘ ~ λ docker run --rm -it bitwalker/alpine-erlang:6.1 date -u
Fri Apr 14 11:48:07 UTC 2017

[Faolain]

Thanks so much @teamon that fixed the date/SSL issue. Strange how that suddenly came up.

@tlvenn Sigh :/ even with the most recent mix_docker update with distillery at 1.3.4 the following still occurs when setting set include_erts: false

 {"init terminating in do_boot",{load_failed,[ets,kernel,erl_parse,gen_event,code,code_server,lists,gen_server,file_server,supervisor,application,proc_lib,application_master,gen,file,heart,file_io_server,erl_eval,application_controller,error_handler,filename,error_logger,erl_lint]}}
init terminating in do_boot ()

Crash dump is being written to: erl_crash.dump...done

[tlvenn]

Ya I have the same issue @Faolain

[Faolain]

@tlvenn, @teamon 's suggestion worked for me! I would say try that.

Also on a sidenote @teamon beside the erts issue I tried commenting on your blog post for clustering but unfortunately Discus threw my comment into spam. I was wondering how you take care of migrations as I am trying to do so with RPC but currently I believe after changing my hostname using the vm.args...I am no longer able to RPC connect to my node and run my migration scripts. Is there a way to set the RPC to a particular destination or a better way to run migrations instead of calling a bash script in post_start with distillery? Since Mix is not available this was the only way I saw how.

https://gist.github.com/Faolain/9312eb507d19c6992d5055ab3296e677

[tlvenn]

@Faolain do you mean you solved the original issue with set include_erts: false ?

Running migration with distellery is actually easy, have a look at: https://hexdocs.pm/distillery/running-migrations.html#content

[Faolain]

@tlvenn unfortunately no, only the clock sync issue teamon pointed out.

I was able to run the migration without having a cluster(changing the node name) but once I change the node away from whatever the default is(I'm assuming 127.0.0.1) the rpc call states that
Node is not running!

[tlvenn]

But what do you need RPC to run migrations ?

[Faolain]

I assume it is using RPC in the background, in any case, running the bin/app migrate within my application returns

/opt/app $ bin/hello migrate
{"init terminating in do_boot",{undef,[{'Elixir.Release.Tasks',migrate,[],[]},{i
nit,start_em,1,[]},{init,do_boot,3,[]}]}}
init terminating in do_boot ()

Crash dump is being written to: erl_crash.dump...done

so I'd love to know how @teamon ended up running migrations using a cluster

[tlvenn]

It works fine on my side but I had the above issue with set include_erts: false. As soon as I put it back to true, things started to work properly.

[mfcollins3]

I was trying this for the first time after reading this article. I had a similar error. Changing set include_erts: true in my config.exs file, it worked correctly as well. I generated the standard Dockerfiles using the mix docker.customize command.