Hi, I just installed ezlaunchpad in my Arch Linux machine, the init command works well, but when I run ~/ez clean or create it doesn't do anything, what could cause this?
I'm running PHP 8.0.2, Docker 20.10.3 and docker-compose 1.28.2

If I activate yarn watch to watch filesystemchanges on css, html, js files the watcher does not react on any changes on editing files locally

how can we solve this

kind regards
Peter

Varnish does not work with ezplatform enterprise edition. Particularly with admin ui purge events. I've tested this using ezlaunchpad by initializing a ezplatform project and an ezplatform-ee project.

Test 1: ezplatform standard edition. Works as expected.

1. ~/ez init
2. Open homepage in tab 1, Varnish cache reports a miss. Refresh again to test that next reload is a hit.
3. Open admin ui in tab 2
4. Edit content
5. Refresh tab 1, Varnish cache has been purged and varnish is a MISS and X-Count-Hits is reset.

Test 2: ezplatform-ee edition. Varnish is broken. Purge not working. Logged in session is being cached.

1. ~/ez init ezystems/ezplatform-ee
2. Open Homepage in tab 1, Varnish cache reports a miss. X-Varnish: 123
3. Open Homepage in tab 2, Varnish cache reports hit. X-Varnish: 123
4. Open admin ui in tab 3
5. Edit ezplatform folder object
6. Refresh Homepage in tab 2, Varnish cache does not purge. Tab 2 cache using different X-Varnish ID. X-Varnish: 789, X-Cache-Hits: 1
7. Repeat step 5, X-Cache-Hits: 2
8. Log out of Admin UI in tab 3
9. Refresh tab 2, Varnish cache returns to same as tab 1. X-Varnish: 123

So it appears to me that ezplatform-ee is not purging content. When logged into the admin ui I can see that varnish is caching the homepage when in ezplatform project that is not happening as expected since Varnish shouldn't be caching content containing cookies. When I log out of admin ui the varnish cache returns to the original cache without the new content.

The other side affect I'm noticing ONLY when using the ezplatform-ee project is when running an ~/ez down && ~/ez create to rebuild the environments that leaving the admin ui logged in will cause varnish to cache a composer error message. If I log out of admin_ui then run the ez commands varnish doensn't cache that error. I believe it has something to do with the admin ui notifications trying to hit the site while it's down and varnish turns on it sees that request and caches it. However I've tested this on the ezplatform project and it doesn't happen there. Seems isolated to enterprise edition only.

~/.composer/cache must be create if it does not exist.

Hello,

how can i work with the SSH Key from my Mac in the Launchpad?

Thanks
Dominik

Hi,

If i destroy the containers with docker rm -f and rebuild all mysql data are lost.
Do i have to add an own mountpoint to store the data on the host?

kind regards

Peter

Hi,
is it possible to run multiple instances of eZ Launchpad. I.e. run projects on different ports?

When I execute the command ~/ez init, I got:

Traceback (most recent call last):
  File "site-packages/PyInstaller/loader/pyiboot01_bootstrap.py", line 200, in <module>
  File "<frozen importlib._bootstrap>", line 971, in _find_and_load
  File "<frozen importlib._bootstrap>", line 955, in _find_and_load_unlocked
  File "<frozen importlib._bootstrap>", line 665, in _load_unlocked
  File "/Users/distiller/compose-osx-release/compose/venv/lib/python3.6/site-packages/PyInstaller/loader/pyimod03_importers.py", line 631, in exec_module
  File "ctypes/macholib/dyld.py", line 18, in <module>
  File "posixpath.py", line 247, in expanduser
KeyError: 'getpwuid(): uid not found: 501'
[81899] Failed to execute script pyiboot01_bootstrap

In Process.php line 239:

  The command "docker-compose -p richardez -f /Users/richard/dev/aad/ez/provisioning/dev/docker-compose.yml -f /Users/richar
  d/dev/aad/ez/provisioning/dev/docker-compose-osx.yml build --no-cache" failed.

  Exit Code: 255(Unknown error)

  Working directory: /Users/richard/dev/aad/ez

  Output:
  ================


  Error Output:
  ================

Docker for Mac 18.06.1-ce-mac73 (26764)
MacOS 10.13.6

Hi,

After an upgrade from 1.7.0 to 1.8.6 with command ez self-update.
Can not use ez command anymore.

I had to chmod u+x /path/to/.ezlaunchpad/ez.phar

I dunno if it could be fix from update (or if it's really an issue) but may be can it be had to your doc if it happen for others :)

Thx for everything you do on ezlaunchpad !

Just noticed the composer_cache folder is no longer owned by wwwdata but by root.

If I try to import 14 GB dump i get:

ERROR 1114 (HY000) at line 8834: The table 'xxxx' is full

kind regards

Peter

Default deploy commands that were created by ezlaunchpad throws an error in platform.sh deploy logs.

bin/console cache:pool:clear cache.redis throws an error: `You have requested a non-existent service "cache.redis"

On command ~/ez init the solar installation breaks with following errormessage:

mkdir: cannot create directory '/ezsolr/server': Permission denied
cp: cannot stat 'vendor/ezsystems/ezplatform-solr-search-engine/lib/Resources/config/solr/*': No such file or directory
chmod: cannot access '/ezsolr/server/ez': No such file or directory

How to fix that?

kind regards

Peter

When installing eZ Enterprise, the php app/console ezplatform:install studio-clean must be run instead of the clean

eZ Launchpad + Legacy Bridge + macOs + NFS is not compatible.
The problem is that /ezpublish_legacy/kernel/classes/clusterfilehandlers/ezfsfilehandler.php uses php function 'flock' to gain exclusive access to ezmutex files. Unfortunately the php flock function calls the underlying flock system call which is not supported by NFS. Therefore no content cache files will be generated, and consequently, Legacy Content Panels will be empty.

If anyone has this problem, a dirty workaround if you are working in dev is to comment out the following lines in /ezpublish_legacy/kernel/classes/clusterfilehandlers/ezfsfilehandler.php:

// Lock the entry for exclusive access, if the entry does not exist
// it will be inserted with mtime=-1
if ( !$this->_exclusiveLock( $fname ) )
{
    // Cannot get exclusive lock, so return null.
    return null;
}

If you run ezlaunchpad outside your home, the export is not used and it fails

ezlaunchpad accept both command:

~/ez init ezsystems/ezplatform 2.5.9
~/ez init ezsystems/ezplatform v2.5.9

But when using "v2.5.9" as version, the "v" char make ezlaunchpad understanding the initialization as a Version 3 instalation.

ezlaunchpad should detect better the version to install.

Atte,
Vincent

On solr entry point, solr engine is run in 'foreground' mode, in this line:

due to this, the solr core creation loop present after this line is never executed, I'm right?

Thanks!

If you are using eZ Launchpad on Mac OS, and you want to migrate to Catalina, just wait.

https://www.firehydrant.io/blog/nfs-with-docker-on-macos-catalina/

We need to redo the mapping from within eZ Launchpad.

Thank you @Webarkitekt who is the one living on the Edge that got first this issue and warned us.

To fix that brand new issue, we will have to change the export path depending on the OS version and the mapping in the docker-composer-osx.yaml.

That will require a new variable but that will be transparent for eZ Launchpad user.

There is a workaround so far:

Change your /etc/exports and your provisioning/dev/docker-compose-osx.yml manually.

It would be very nice if eZ Launchpad supported xdebug.

Could this be implemented?

It's impossible to run EzPublishRestBundle functional tests, as the client which executes http requests obtains the http host from environment variable 'EZP_TEST_REST_HOST', which should have value 'nginx'.

Solution:
Edit docker-compose.yml to include the 'EZP_TEST_REST_HOST' environment variable with value 'nginx' as so:

version: '2.1'
services:
    ...
    engine:
        ...
        environment:
            ...
            - EZP_TEST_REST_HOST=nginx

Thx @florianalexandre for the idea.

• install composer in the dockerfile
• remove the code that manages that in eZ Launchad

Travis does not work when someone forked because of 'before_install' line about decryption of the tar.

Todo: do it only for a merge on master. (not useful otherwise)

Currently no domains can be setup for the services. Traefik would be a great addition to the current Launchpad setup.

E.g.

• solr.PROJECTNAME.local
• mailcatcher.PROJECTNAME.local
• redis.PROJECTNAME.local

When installing 2.5.x the schema.xml config from ezplatform-solr-search-engine v1.7.3 is incompatible with solr 7.7 layout configuration.

To avoid travis error on building docker we need to have the latest database and storage.
Keep ez docker:importdata

But we could add some commands like :

• ez docker:importdb
• ez docker:importstorage
• ez docker:import-storage-from-url
  also : (but not very secure) :
• ez docker:import-db-from-url

So we could use dropbox or gdrive to import storage. What do you think ?

Dear @Plopix ,

I have tested eZLaunchpad more this week, and in fact it is not working perfectly :

1. RAM is not managed correctly : As the engine is not based on an Apple M1 architecture (arm64v8 ?), the RAM is not correctly set up and we have exhausted memory at 256Mo even if we assign 8Go of Ram to docker)

See this issue for examples that have same kind of issue because of not correct architecture : docker/for-mac#5204

2. If I force to build a dedicated version for the Apple M1 architecture - all other library won't work, especially concerning PHP GD installation and configuration.

Available to let you test different option on my machine,

Yohann

Indeed auth is stored on the container directly, then it is lost after a rebuild.

We have to rerun composer command

The workflow main-ci.yaml is referencing action actions/checkout using references v1. However this reference is missing the commit a6747255bd19d7a757dbdda8c654a9f84db19839 which may contain fix to the some vulnerability.
The vulnerability fix that is missing by actions version could be related to:
(1) CVE fix
(2) upgrade of vulnerable dependency
(3) fix to secret leak and others.
Please consider to update the reference to the action.

eZ Platform 3.x has now ElasticSearch in the EE version.

The launchpad wizard should be updated to propose ElasticSearch OR Solr when on Enterprise.

• Update the Wizard
• Do the Docker config
• Do the recipe for install
• Tests

A fresh install of ezlaunchpad and a new init of ezplatform results in a solr collection error.

I've noticed that this happens off and on. See #61 .travis failure for example.

I believe this may have something to do with permissions. Ref #55

Solr is running

Creating new core 'collection1' using command:
http://localhost:8983/solr/admin/cores?action=CREATE&name=collection1&instanceDir=collection1


ERROR: Error CREATEing SolrCore 'collection1': Could not create a new core in /ezsolr/server/ez/collection1as another core is already defined there

Core collection1 created.
Solr is running
Re-indexing started for search engine: eZ Platform Solr Search Engine

Purging index...
Re-Creating index for 13 items across 1 iteration(s), using single (current) process:
 0/1 [>---------------------------]   0%08:07:10 ERROR     [console] Error thrown while running command "--env=prod ezplatform:reindex". Message: "Wrong HTTP status received from Solr: 404 on http://solr:8983/solr/collection1\nEzSystems\EzPlatformSolrSearchEngine\Gateway\Endpoint::__set_state(array(\n   'scheme' => 'http',\n   'user' => NULL,\n   'pass' => NULL,\n   'host' => 'solr',\n   'port' => 8983,\n   'path' => '/solr',\n   'core' => 'collection1',\n))\nEzSystems\EzPlatformSolrSearchEngine\Gateway\Message::__set_state(array(\n   'headers' => \n  array (\n    'version' => '1.1',\n    'status' => 404,\n    'Cache-Control' => 'must-revalidate,no-cache,no-store',\n
...
..
.

Add a new command init_ibexa to initialize a new ibexa 3.3.x project

Would be nice to include more local cache directories i.e. ".npm" the same way .composer is handled. Instead of hard coding .composer it could be included dynamically. Essentially create a dynamic array of variables based on the global.yml file.

After executing the ~/ez init command, the containers are well created but the following error appear:

Recreating richardez_engine_1 ... error

ERROR: for richardez_engine_1  Cannot start service engine: error while mounting volume '/var/lib/docker/volumes/richardez_nfsmount/_data': error while mounting volume with options: type='nfs' device=':/Users/richard/dev/aad/ez/' o='addr=host.docker.internal,lock': connection refused

ERROR: for engine  Cannot start service engine: error while mounting volume '/var/lib/docker/volumes/richardez_nfsmount/_data': error while mounting volume with options: type='nfs' device=':/Users/richard/dev/aad/ez/' o='addr=host.docker.internal,lock': connection refused
ERROR: Encountered errors while bringing up the project.

In Process.php line 239:

  The command "docker-compose -p richardez -f /Users/richard/dev/aad/ez/provisioning/dev/docker-compose.yml -f /Users/richard/dev/aad/ez/provis
  ioning/dev/docker-compose-osx.yml up -d" failed.

  Exit Code: 1(General error)

  Working directory: /Users/richard/dev/aad/ez

  Output:
  ================


  Error Output:
  ================

Docker for Mac 18.06.1-ce-mac73 (26764)
MacOS 10.13.6

What do you think @andrerom

Do we need to provide the support for 5.6?
We are talking about the PHP version useful for the CLI of eZ Launchpad, not for eZ Platform.

Removing PHP 5.6 support would allow us to:

• move to Symfony 4.1
• use strong type hint
• speed Travis

Actually, the question is for everybody, I feel like since OSX ships 7.1 by default, we could/should.

?

Solr Container contains permissions issue when running ez create.

Solr is unable to create the new data directory inside /ezsolr/server/ez/collection1

To avoid having to redo all steps if you forget to start docker server.

Probably most relevant for Windows and Mac, as this can happen if Docker server/app is stopped/paused.

Unsure which command is best to check against, but here is one possible:

$ docker system info
Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?

thanks @luckyluketest

It seems the limit for the TCP port prefix should be between 1 and 64.

With this option:

What is the TCP Port Prefix you want? [42]:
> 65

I had this error

ERROR: for solr  Cannot create container for service solr: invalid port specification: "65983"
ERROR: Encountered errors while bringing up the project.

see: #56

I can clear my cache via ez sf cache:clear & ez sf cache:clear --env=dev however when trying to clear the prod env cache I get a [ERROR] Your are not in a folder managed by eZ Launchpad. when running ez sf cache:clear --env=prod

Not a deal breaker, I can ez enter -> cd project/ezplatform -> bin/console cache:clear --env=prod but the ez sf shortcut is supper nice to have :)

I get a Communication link failure: 1153 Got a packet bigger than 'max_allowed_packet'

How can i Increese max_allowed_packet

Kind regards

Peter

This line is causing 2 issues here (Ubuntu 16.04):

• New package requirement: gnustep-base-runtime to have "defaults" command available
• After that installed, every launchpad execution we're getting: defaults read: couldn't read default

Don't know if in modern Ubuntu releases is also happening.

Thanks!

I replaced adminer with phpmyadmin and ez info isn't smart enough to pick up this change. It would be nice if ez info would be able to dynamically pick up additional containers added to the current project.

I try to get ezlaunchpad running on an Apple M1 and I am aware that Docker is not yet completely "M1 ready".
However I think this issue is not related to the Docker compatibility.

Varnish instantly crashes

varnish_1      | Error:
varnish_1      | Running VCC-compiler failed, signal 5
varnish_1      | VCL compilation failed

I think this might be related to Varnish not yet providing ARM packages

If you are using multi databases with platform.sh, getmysqlcredentials.php returns nothing

Documentation is harder to update because we did it "multilingual"

• Do you think that is useful? I could be full English and would simplify the addition etc.

• Doc for Blackfire usage/config

• Doc for MySQL usage/config

• Doc for X-Debug usage/config

More, other ideas for the doc?

In ez engine container, environment variable SYMFONY_TMP_DIR has a typo. /tmp/ezplatormcache/var/cache/dev

I'm not sure if this is causing any issues with ez sf cache:clear by not emptying the cache folder.

I've been running bin/console inside the engine container to work around this.