Deployer: Laravel, npm install, fresh & first deploy [Deployer\Exception\RuntimeException (-1)]

[jrean]

Hi,

deployer 5.1.3
deployer recipes dev-master
Local Mac Os Sierra
Remote Ubuntu 16.04

<?php

namespace Deployer;

require 'recipe/laravel.php';
require 'recipe/npm.php';
...

task('deploy', [
    'deploy:prepare',
    'deploy:lock',
    'deploy:release',
    'deploy:update_code',
    /* 'deploy:shared', */
    /* 'deploy:vendors', */
    /* 'deploy:writable', */
    /* 'artisan:storage:link', */
    /* 'artisan:view:clear', */
    /* 'artisan:cache:clear', */
    /* 'artisan:config:cache', */
    /* 'artisan:optimize', */
    /* 'deploy:symlink', */
    /* 'deploy:unlock', */
    /* 'cleanup', */
]);

after('deploy:update_code', 'npm:install');

For an unknown reason I receive the following trace:

➤ Executing task deploy:update_code
✔ Ok [19s 945ms]

➤ Executing task npm:install
[xxx] > which npm
[xxx] < /usr/bin/npm
[xxx] > cd /var/www/test-deploy/releases/1 && /usr/bin/npm install
➤ Executing task deploy:failed
• done on [xxx]
✔ Ok [1ms]

[Deployer\Exception\RuntimeException (-1)]
  The command "cd /var/www/test-deploy/releases/1 && /usr/bin/npm install" failed.
  Exit Code: -1 (Unknown error)
  Host Name: xxx
  ================

Exception trace:
 () at /Volumes/xxx/code/xxx/xxx/vendor/deployer/deployer/src/Ssh/Client.php:95
 Deployer\Ssh\Client->run() at /Volumes/xxx/code/xxx/xxx/vendor/deployer/deployer/src/functions.php:297
 Deployer\run() at /Volumes/xxx/code/xxx/xxx/vendor/deployer/recipes/npm.php:23
 Deployer\Deployer::Deployer\{closure}() at n/a:n/a
 call_user_func() at /Volumes/xxx/code/xxx/xxx/vendor/deployer/deployer/src/Task/Task.php:82
 Deployer\Task\Task->run() at /Volumes/xxx/code/xxx/xxx/vendor/deployer/deployer/src/Executor/SeriesExecutor.php:63
 Deployer\Executor\SeriesExecutor->run() at /Volumes/xxx/code/xxx/xxx/vendor/deployer/deployer/src/Console/TaskCommand.php:142
 Deployer\Console\TaskCommand->execute() at /Volumes/xxx/code/xxx/xxx/vendor/symfony/console/Command/Command.php:264
 Symfony\Component\Console\Command\Command->run() at /Volumes/xxx/code/xxx/xxx/vendor/symfony/console/Application.php:869
 Symfony\Component\Console\Application->doRunCommand() at /Volumes/xxx/code/xxx/xxx/vendor/deployer/deployer/src/Console/Application.php:132
 Deployer\Console\Application->doRunCommand() at /Volumes/xxx/code/xxx/xxx/vendor/symfony/console/Application.php:223
 Symfony\Component\Console\Application->doRun() at /Volumes/xxx/code/xxx/xxx/vendor/symfony/console/Application.php:130
 Symfony\Component\Console\Application->run() at /Volumes/xxx/code/xxx/xxx/vendor/deployer/deployer/src/Deployer.php:320
 Deployer\Deployer::run() at phar:///usr/local/bin/dep/bin/dep:119
 require() at /usr/local/bin/dep:4

When I ssh on the remote server and cd into /var/www/test-deploy/release, I can see the node_modules directory with the packages inside.

If I run the very first deploy without the npm task, everything is running smooth and the second deploy will be successfully achieved with npm install and npm run production tasks.

Can I get some light, please?

The only hint I found is Deployer\Deployer::Deployer\{closure}() at n/a:n/a

Note: Using the yarn recipe (require 'recipe/yarn.php';) instead of npm works properly.

[qerim]

@jrean this was bugging me for a while.

Then I remembered I am running a DigitalOcean 512MB System, so you need to 'create' more RAM:

sudo fallocate -l 1G /swapfile
sudo chmod 600 /swapfile
sudo mkswap /swapfile
sudo swapon /swapfile

It works now :)

[jrean]

@qerim, smart!
I didn't think about it and it makes complete sense.
@antonmedv do you still consider it as a bug. Is there something that could be improved on the package side to give more details when such error happens?

Thxs

[antonmedv]

I think will be cool to detect stuff like this and inform user what to do. Will be completely awesome.

[itelmenko]

@jrean , Can you post result config? With npm run production (or yarn)

[qerim]

@antonmedv I don't think this should be classed as Deployer bug.

You could detect this issue by checking if the server has less than 1GB RAM and issue a message instructing them to either allocate virtual RAM or install more RAM.

[staabm]

what do you guys think about showing the actual used RAM by php besides the timing information like

➤ Executing task deploy:update_code
✔ Ok [19s 945ms, 63MB memory]

➤ Executing task npm:install
[xxx] > which npm
[xxx] < /usr/bin/npm
[xxx] > cd /var/www/test-deploy/releases/1 && /usr/bin/npm install
➤ Executing task deploy:failed
• done on [xxx]
✔ Ok [1ms, 112MB memory]

?

Thats not the perfect solution for the problem but will give the user a hint about memory consumption.