The demo-nodejs-api seems to have some issues installing and running under the latest LTS release of Node.js (10.x): https://github.com/geerlingguy/demo-nodejs-api

Basically, if you leave a Node.js app alone for more than 6 months or so, half the dependencies will go unmaintained and/or change names entirely (I'm looking at you, jade/pug!).

The example works under Ubuntu 14.04, but I am upgrading all my examples to at least 16.04 (in this case I would like to just go straight to 18.04), and I can't install Node.js 4.x or lower (which is compatible with the current nodejs-api demo app linked above) on Ubuntu 18.04... nor should I.

So for this ticket, I need to do some upstream work in the Node.js API demo app, to get it either rewritten or updated to work on modern Node.js, then test the full process on the deployments-rolling example.

TASK [Install Drupal dependencies with Composer.] ****************************** fatal: [default]: FAILED! => {"msg": "Failed to set permissions on the temporary files Ansible needs to create when becoming an unprivileged user (rc: 1, err: chown: changing ownership of '/var/tmp/ansible-tmp-1580524622.7705758-209663871137752/': Operation not permitted\nchown: changing ownership of '/var/tmp/ansible-tmp-1580524622.7705758-209663871137752/AnsiballZ_command.py': Operation not permitted\n}). For information on working around this, see https://docs.ansible.com/ansible/become.html#becoming-an-unprivileged-user"}

running against the provided vagrant vm setup - wrote the playbook on my own, ran into some issues, so i grabbed the playbook.yml provided in the drupal dir.

This appears to be a long-standing issue with Ansible, as seen here:

https://github.com/ansible/ansible/issues/15988

I commented on that thread with what I found, but I will copy+paste it here as well:

Attempting to do the ad-hoc command of:

ansible multi -b -B 3600 -a "yum -y update"

only returns:

192.168.60.4 | SUCCESS | rc=0 >>
Loaded plugins: fastestmirror
Loading mirror speeds from cached hostfile
 * base: mirror.den1.denvercolo.net
 * extras: mirrors.cat.pdx.edu
 * updates: mirrors.centos.webair.com
No packages marked for update

192.168.60.5 | SUCCESS | rc=0 >>
Loaded plugins: fastestmirror
Loading mirror speeds from cached hostfile
 * base: mirror.nodesdirect.com
 * extras: repos.mia.quadranet.com
 * updates: mirror.atlantic.net
No packages marked for update

192.168.60.6 | SUCCESS | rc=0 >>
Loaded plugins: fastestmirror
Loading mirror speeds from cached hostfile
 * base: mirror.den1.denvercolo.net
 * extras: mirror.jax.hugeserver.com
 * updates: mirrors.centos.webair.com
No packages marked for update

when it should be returning something like:

background launch...
192.168.60.6 | success >> {
    "ansible_job_id": "763350539037",
    "results_file": "/root/.ansible_async/763350539037",
    "started": 1
}
... [other hosts] ...

• [x] p.23 I get a deprecation warning when I try to use ANSIBLE_HOSTS, and am recommended ANSIBLE_INVENTORY instead
• [x] p.35 remove "owner=" and "group=" parameters for "state=link". This should have been caught by a technical reviewer. The owner and group assignment changes the original file, and not the symlink!
• [x] p.35 technically 'state=absent' can be used to remove a symlink as well
• [x] p.36,37 about getting an ansible_job_id: have you tried running this? I'm on Ansible 2.7.5. Not getting any feedback about any ansible_job_id
• [x] p.44 would really be nice to explain why fireball mode requires ZeroMQ... but accelerated mode doesnt require anything
• [x] p.46 explanation of 'plays' in first paragraph almost makes it look like 'play' = 'task'. Reading further in to the book, I think we figured that 'play' = "the entire set of tasks in a playbook" instead
• [x] p.50 are 'with_items', and '{{ item }}' special? In this case it appears that "item" is like a special reserved variable. Would be great to mention that.
• [x] p.54 "Real-world" Node.js app server: how real-world is copying files over from our master to the server, instead of say, using git checkout?
• [x] p.56 shell script: 'yum ... install node' does not work. Contrast this against the playbook on p.57, where it's "name=npm" (and not "name=node")
• [x] p.56 last paragraph: 'and works with variables other nifty Ansible features we’ll discuss later:' -> missing 'and' in 'and works with variables [and] other nifty Ansible features'
• [x] p.57
  • why does the code segment start at line 10, when fragment on p.55 ends at line 5? this is just confusing and raises questions (did we miss any lines?)
  • as mentioned, line 27 uses "name=npm", as opposed to the expected "name=node" (or "name=nodejs"?) when you look at the script on p.56
  • it would be nice to avoid quotes (see lines 12 and 17) unless really necessary. Otherwise it's confusing
  • would be better for playbook to follow the same order of "import Remi's key", then "install Remi repo" as per script it's supposed to replace. Otherwise, tell me why you're reversing the order!
  • contrary to what you say, installing the EPEL repo also requires importing the key; you just don't see it. But if you look carefully at the 'msg' field in the JSON returned by ansible, there will be a message about importing the key!
• [x] p.59 similar comment about using quotes (lines 33, 36) unnecessarily in the playbook
• [x] p.60 unnecessary quotes for line 48
• [x] p.65
  • a point about using custom Vagrant boxes: personally I would prefer using plain vanilla distro boxes (i.e. "ubuntu/xenial64" instead of "geerlingguy/ubuntu1604"). When I'm setting up a real box for production, I'm not going to be using a customized install. I'll most likely be using and starting from a plain vanilla distro install! It appears that "ubuntu/xenial64" will have a problem with installing 'mysql-server', but I guess we'll just have to deal with it?
  • s/Python/ansible/ in last paragraph
• [x] p.66 line numbers are wrong (p.65 already ends on line 60, but you start at line 60 again)
• [x] p.70 MySQL no longer installs a 'test' database by default?
• [x] p.73 please stop switching back and forth between 'shell:' and 'command:' for no reason. It's just confusing
• [x] p.74 line 166: according to our own testing, it appears that this permission-setting neither resolves, nor achieves anything?
• [x] p.77 playbook, including https://github.com/geerlingguy/ansible-for-devops/blob/master/solr/provisioning/playbook.yml, is missing a necessary 'become: yes'!
• [x] p.79
  • use 'remote_src: yes' instead of 'copy: no'. See doc.
  • 'command:' works instead of 'shell:'
• [x] p.80 would have appreciated a warning at the start to provision the VM with at least 512MB of RAM! or else Solr will fail
• [x] p.85 'instead of using lineinfile with a large list of items' -> u mean instead of using many 'lineinfile' tasks?
• [x] p.86 s/var_proxy/proxy_vars/ -> more intuitive!
• [x] p.86 method of handling /etc/environment seems counter-intuitive; if the vars and the task is all in the same playbook/file, why not have "proxy_state" be part of the "with_items" list? The only win this gives you by having a separate 'proxy_state' in a separate var is an easy (but not selective) on and off switch... for all proxying. But really I think the big win here would be if the task were imported from a shared playbook.

It appears that Ansible has created a new set of modules that no longer rely on dopy and add more functionality. The current one used in the book will be deprecated in 1.12. It may be beneficial to move to using the newer module(s).

Hello I'm getting this error in vagrant version 2.3.0

Command I used:

vagrant init geerlingguy/centos7 vagrant up

Bringing machine 'default' up with 'virtualbox' provider...
==> default: Box 'geerlingguy/centos7' could not be found. Attempting to find and install...
    default: Box Provider: virtualbox
    default: Box Version: >= 0
The box 'geerlingguy/centos7' could not be found or
could not be accessed in the remote catalog. If this is a private
box on HashiCorp's Vagrant Cloud, please verify you're logged in via
`vagrant login`. Also, please double-check the name. The expanded
URL and error message are shown below:

URL: ["https://vagrantcloud.com/geerlingguy/centos7"]
Error: schannel: next InitializeSecurityContext failed: Unknown error (0x80092012) - The revocation function was unable to check revocation for the certificate.

After following the setup guide on page 68 to 72, I can't seem to be able to connect or ping the created vms through their private network ips. My problem is similar to https://leanpub.com/ansible-for-devops/feedback#comment-2101413414 or this one but neither of the solution provided their is working for me.

I am using:

• OS X El Capitan
• Virtualbox: Version 5.0.26 r108824
• Vagrant 1.8.5
• ansible 2.1.0.0

I do suspect that there is something wrong with the vagrant box cause I've tried other boxes(e.g.scotch/box) with the same vagrant file configuration and the private ips works fine.

Please, any help or guidance will be really appreciated. Also, let me know if you need more information about my setup.

Thanks,

From failed build: https://travis-ci.org/geerlingguy/ansible-for-devops/jobs/475166316

TASK [geerlingguy.ruby : Install Bundler.] *************************************
fatal: [localhost]: FAILED! => {"changed": false, "cmd": "/usr/local/bin/gem install --no-user-install --no-document bundler", "msg": "ERROR:  Error installing bundler:\n\tbundler requires Ruby version >= 2.3.0.", "rc": 1, "stderr": "ERROR:  Error installing bundler:\n\tbundler requires Ruby version >= 2.3.0.\n", "stderr_lines": ["ERROR:  Error installing bundler:", "\tbundler requires Ruby version >= 2.3.0."], "stdout": "", "stdout_lines": []}

I need to get the 14.04 LTS examples upgraded to 18.04 anyways... this ain't 2014 anymore!

TASK [Install Drupal.] *********************************************************
fatal: [default]: FAILED! => {"changed": true, "cmd": ["drush", "si", "-y", "--site-name=Drupal Test", "--account-name=admin", "--account-pass=admin", "--db-url=mysql://drupaltest:1234@localhost/drupaltest"], "delta": "0:00:00.195786", "end": "2017-05-11 00:12:07.375858", "failed": true, "rc": 255, "start": "2017-05-11 00:12:07.180072", "stderr": "PHP Fatal error:  Class Symfony\\Component\\Console\\Style\\SymfonyStyle contains 4 abstract methods and must therefore be declared abstract or implement the remaining methods (Symfony\\Component\\Console\\Output\\OutputInterface::isQuiet, Symfony\\Component\\Console\\Output\\OutputInterface::isVerbose, Symfony\\Component\\Console\\Output\\OutputInterface::isVeryVerbose, ...) in /var/www/drupal-8.2.x-dev/vendor/symfony/console/Style/SymfonyStyle.php on line 33\n [error]  Drush command terminated abnormally due to an unrecoverable error.\nError: Class Symfony\\Component\\Console\\Style\\SymfonyStyle contains 4 abstract methods and must therefore be declared abstract or implement the remaining methods (Symfony\\Component\\Console\\Output\\OutputInterface::isQuiet, Symfony\\Component\\Console\\Output\\OutputInterface::isVerbose, Symfony\\Component\\Console\\Output\\OutputInterface::isVeryVerbose, ...) in /var/www/drupal-8.2.x-dev/vendor/symfony/console/Style/SymfonyStyle.php, line 33 ", "stderr_lines": ["PHP Fatal error:  Class Symfony\\Component\\Console\\Style\\SymfonyStyle contains 4 abstract methods and must therefore be declared abstract or implement the remaining methods (Symfony\\Component\\Console\\Output\\OutputInterface::isQuiet, Symfony\\Component\\Console\\Output\\OutputInterface::isVerbose, Symfony\\Component\\Console\\Output\\OutputInterface::isVeryVerbose, ...) in /var/www/drupal-8.2.x-dev/vendor/symfony/console/Style/SymfonyStyle.php on line 33", " [error]  Drush command terminated abnormally due to an unrecoverable error.", "Error: Class Symfony\\Component\\Console\\Style\\SymfonyStyle contains 4 abstract methods and must therefore be declared abstract or implement the remaining methods (Symfony\\Component\\Console\\Output\\OutputInterface::isQuiet, Symfony\\Component\\Console\\Output\\OutputInterface::isVerbose, Symfony\\Component\\Console\\Output\\OutputInterface::isVeryVerbose, ...) in /var/www/drupal-8.2.x-dev/vendor/symfony/console/Style/SymfonyStyle.php, line 33 "], "stdout": "", "stdout_lines": []}

I know nothing about Drupal so I'm kind of useless at debugging this. I can tell you that changing drupal_core_version to other tags in the 8.2.x branch didn't help.

Book version: 1.21 Chapter 4 introduces pre_tasks (page 63) with a brief explanation that they're run before the main set of tasks. It's not clear (at this point in the book anyway) how that's any different to just adding a task at the beginning of the tasks section. Assuming the reasoning is covered later in the book, it would be useful to reference that section, or provide a simple explanation at this point of the book.

Using Ansible 2.4.2.0; Darwin

When running the nodejs provisioning playbook, it keeps failing when it reaches this task:

https://github.com/geerlingguy/ansible-for-devops/blob/ed1f0325bab17b9115b355403b9579e68634de66/nodejs/provisioning/playbook.yml#L40

The error message :

[Errno 2] No such file or directory
fatal: [domain.com]: FAILED! => {"changed": false, "cmd": "forever list", "rc": 2}

The full task:

- name: "Check list of running Node.js apps."
  command: forever list
  register: forever_list
  changed_when: false

I've triple-checked that the app.js file is there. Why does it not find it ?

Minor Inline Code Formatting Issues

When viewing the PDF, incline code at the right side of paragraphs can go past margins, sometimes off the page. Haven't discovered any hard issues, since even in the case where main.yml is cut off, most people could figure out what it is supposed to be. My concern is that in future editions (or other books), inline code blocks might be added that extend past the page and miss key details.

Reproduce

Using the 15 April 2022 PDF from Leanpub, view the Examples pages. Tested with Firefox and Brave, issue is on both so it seems to be a PDF issue, not a PDF renderer issue. EPUBs don't have this issue.

If you're using Markdown -> Pandoc, my suspicion is that whatever filter is running to convert the inline code blocks from Markdown to equivalent LaTeX doesn't know when to break a "word" up if there's no natural separation like a space or an underscore (This is a guess, I haven't worked much with Pandoc and don't know your pipeline).

Examples

• Chapter 10, Page 279, file name slightly extends past margins
• Chapter 11, Page 315, command extends past margins (unclear if this is the full command)
• Chapter 12, Page 329, command extends past margins
• Chapter 12, Page 334, IP address continues past margins
• Chapter 15, Page 400, end of main.yml is not visible

Negative Examples

I've actually found a hyphenated inline code block - I don't know if this was a manual or automatic fix since the hyphen doesn't exist inside the EPUB.

• Chapter 9, Page 229, both underscores and spaces are split appropriately
• Chapter 10, Page 273, multi-word command is split correctly at end of paragraph. Suggests possible issue is that the filter doesn't know how to split a "word" and can only hyphenate things that are clearly multiple words (underscore or space separated).
• Chapter 10, Page 292

Possible fix

I've had issues with Markdown -(Pandoc)-> PDF before too and haven't gotten a great solution. Not sure exactly what your pipeline is, but this answer on tex.stackexchange with fvextra might help prevent these issues: https://tex.stackexchange.com/a/412316

The drupal playbook uses a syntax like this:

    - name: Disable the default site.
      command: >
        a2dissite 000-default
        removes=/etc/apache2/sites-enabled/000-default.conf
      notify: restart apache

From the way I do interpret the YAML specs and the ansible documentation this would result in a command string of a2dissite 000-default removes=/etc/apache2/sites-enabled/000-default.conf

eg combining all latter strings while the intent is to make a2dissite 000-default the command and removes= an ansible option

I'm not sure if I am missing something here, this "exploits" some side effect/undocumented features or that's a bug

if string in blue_values:
    return True
else:
    return False

could be replaced by

return string in blue_values

I feel it looks simpler and better that way, the conditional block is not necessary there.

Cheers, loving the book so far!

The second paragraph mentions '... you can automating unit, functional, and integration tests ...', but I believe the correct verb tense should be 'automate'.

I found an Intel based Mac and got by #510. Then I got it with #506. Now I'm getting the following errors. If guest additions matter, suggestions on how properly configure them? (or avoid them all together if that's my problem.)

Except for adding virtualbox__intnet: true to the end of each private network config line, I've only copied and pasted the Vagrantfile code from Ch3.

Got different reports about installed GuestAdditions version:
Virtualbox on your host claims:   6.1.32
VBoxService inside the vm claims: 7.0.2
Going on, assuming VBoxService is correct...
Got different reports about installed GuestAdditions version:
Virtualbox on your host claims:   6.1.32
VBoxService inside the vm claims: 7.0.2
Going on, assuming VBoxService is correct...
Got different reports about installed GuestAdditions version:
Virtualbox on your host claims:   6.1.32
VBoxService inside the vm claims: 7.0.2
Going on, assuming VBoxService is correct...
Restarting VM to apply changes...
==> db: Attempting graceful shutdown of VM...
==> db: Booting VM...
==> db: Waiting for machine to boot. This may take a few minutes...
==> db: Machine booted and ready!
==> db: Checking for guest additions in VM...
    db: No guest additions were detected on the base box for this VM! Guest
    db: additions are required for forwarded ports, shared folders, host only
    db: networking, and more. If SSH fails on this machine, please install
    db: the guest additions and repackage the box to continue.
    db: 
    db: This is not an error message; everything may continue to work properly,
    db: in which case you may ignore this message.
==> db: Setting hostname...
==> db: Configuring and enabling network interfaces...
(base) name@C02NX7BWG3QD ch3 % ansible multi -a "hostname"
192.168.60.4 | UNREACHABLE! => {
    "changed": false,
    "msg": "Failed to connect to the host via ssh: ssh: connect to host 192.168.60.4 port 22: Operation timed out",
    "unreachable": true
}
192.168.60.5 | UNREACHABLE! => {
    "changed": false,
    "msg": "Failed to connect to the host via ssh: ssh: connect to host 192.168.60.5 port 22: Operation timed out",
    "unreachable": true
}
192.168.60.6 | UNREACHABLE! => {
    "changed": false,
    "msg": "Failed to connect to the host via ssh: ssh: connect to host 192.168.60.6 port 22: Operation timed out",
    "unreachable": true
}