iphoting/heroku-buildpack-php-tyler

Superseded by: https://github.com/CHH/heroku-buildpack-php Formerly, a custom Heroku buildpack that bundles Nginx + PHP (via FPM). PHP dependencies are handled by Composer.

heroku-buildpack-php-tyler's Issues

log (access and error) does not show up in "heroku logs"

heroku[router] logs are there but
app[web.1] log entries are missing after the entry "State changed from starting to up"

You wrote

All output to `/app/vendor/php/var/log/php-fpm.log` 
is sent to `heroku logs`. 
So, just run `heroku logs` too see them.

"heroku logs" doesn't bring them up. Tested with the "develop" branch and newrelic active.

User "orensol" suggested (via disqus):

The php-fpm.conf file has a a setting that makes the error log this:
php_admin_value[error_log] = /app/vendor/php/var/log/php-errors.log

Since this log file doesn't exist, it isn't tailed when boot.sh is run. 
So I guess you should either remove this setting, 
or create an empty file with that name in the build.

I stumbled about having not logs at all and thought this issue is not closed.

Consider using `port_in_redirect off;` in default nginx.conf.erb

The default is on, which means that if any redirects do end up happening, the value of ENV['PORT'] will be included in the URL returned to the browser. This URL is of course inaccessible to the browser. Thanks.

Upgrade New Relic instrumentation to v3.0.

New version announced: http://blog.newrelic.com/2012/10/23/new-relic-launches-php-agent-3-0-includes-multi-tenancy-support/

Update PHP to v5.4.12

EC2 ami does not exist & buckets urls

Hello,

I tried to compile php onto my machine, but it's not really working on heroku.

But when I run ./support/ec2-up.sh I get Client.InvalidAMIID.NotFound: The image id '[ami-04c9306d]' does not exist

Also the s3 url S3_URL="https://s3.amazonaws.com/${S3_BUCKET}" doesn't work anymore as they want us to prefix with the bucket like this : S3_URL="https://${S3_BUCKET}.s3.amazonaws.com/"

Edit:

The ami actually works if I use us-east-1, I was using eu-west-1

Also, shouldn't ec2-up.sh execute a ec2-run-instances --user-data-file ec2-build-php.sh ??

Update PHP to v5.5.5.

`support/package_php` fails when not executed in the same directory as the script is in.

Performance deviates from README.

Set HTTPS on only when SSL is enabled

Use a conditional for https://github.com/iphoting/heroku-buildpack-php-tyler/blob/master/conf/nginx.conf.erb#L98.

Solution:

map $http_x_forwarded_proto $https {
  default off;
  https on;
}
fastcgi_param HTTPS $https if_not_empty;

Update PHP to v5.5.2.

http://www.php.net/ChangeLog-5.php#5.5.2

Memcached::setSalsAuthData() not found

I got this error and tried to solve it by enabling memcache sasl (memcached.use_sasl = On). That didn't work. I tried to compile memcached with --enable-memcached-sasl, but that didn't work to. On support/vulcan-build-php.sh line 106 (sed -i -e '21 s/no, no/yes, yes/' ./config.m4) config.m4 is modified and that seems to enable sasl during the memcached compilation.

I also tried to compile libmemcached with --with-memcached_sasl, but that didn't work to. Does somebody use this buildpack with a memcache service authenticated by sasl?

PHP 5.4.11

New PHP version.

nginx and co. binaries not in the PATH

heroku create myphpinfo --buildpack https://github.com/iphoting/heroku-buildpack-php-tyler.git

results in Application Error, because (according to logs) it wasn't able to locate nginx or php-fpm

• php-fpm: not found
• nginx: not found

Possible Solution

symlink binaries to /app/bin or update $PATH variable with corresponding locations

PHP 5.4.10

New PHP version.

New Relic Server Monitoring Support

Since Server Monitoring is free to everyone, can you please add in the required daemons? More info is available here

Update PHP to v5.4.16.

How can i define Worker job and what worker job doing in your PHP buidpack

Your build pack very cool, but i don't know how to define Worker ( structure in Profile) and how to define what Worker i define in Procfile do ? ( what file we define and structure to define like Rake in Ruby on rail)

'user' and 'group' directive is ignored when FPM is not running as root

Consider to uncomment
https://github.com/iphoting/heroku-buildpack-php-tyler/blob/develop/conf/php-fpm.conf#L132-L133
to avoid the log-entry on startup:

'user' directive is ignored when FPM is not running as root
'group' directive is ignored when FPM is not running as root

erb not found

boot.sh can't produce nginx.conf because erb not found, do we need erb at all if there is sed available?, e.g.:

sed "s/\\\$PORT/\$PORT/g" /app/vendor/nginx/conf/nginx.conf.tpl > /app/vendor/nginx/conf/nginx.conf

Where nginx.conf.tpl could looks like this

php doesn't recognize environment variables

I might be missing, but php doesn't see environment variables with this buildpack, e.g.:

getenv("CLEARDB_DATABASE_URL");

or

getenv("BUILDPACK_URL");

doesn't return anything.

I also found that original buildpack puts environment variables into httpd.conf in boot.sh script, see: https://github.com/heroku/heroku-buildpack-php/blob/master/bin/compile#L50

CloudFlare Railgun™ support

May be an interesting feature to support:

http://www.cloudflare.com/railgun
http://blog.cloudflare.com/cloudflares-railgun-easier-than-ever
http://www.cloudflare.com/docs/railgun/installation.html

Update Nginx to v1.4.1.

Update pertains to a security issue which only affects the 1.3.x branch and v1.4.0.

See http://nginx.org/en/CHANGES-1.4.

WordPress Permalinks not working

Installed heroku-buildpack-php-tyler to my Heroku app which is running Wordpress and my permalinks no longer work. I can only access the homepage and /wp-admin. If i change my permalinks back to the default structure (?p=123), then the permalinks will work again, but I really could do with having them Pretty.

I found this post on the CHH version of the buildpack. I put the nginx.conf file in the root and pushed it but it didn't work, and I wasn't sure what to do about the composer.json stuff, but this didn't work (probably because I didn't have the Composer stuff).

I'm quite new to Heroku and nginx (as you can probably tell) so if you could it take nice and slow that'd be appreciated. Thanks.

Alternatively, how would I go how uninstalling/unsetting the buildpack? I'm building this site for a client and it might be easier just to have the default apache engine installed. For some reason whenever I run heroku config:unset BUILDPACK_URL=git://github.com/iphoting/heroku-buildpack-php-tyler.git, it just returns this: `Unsetting BUILDPACK_URL=git://github.com/iphoting/heroku-buildpack-php-tyler.git and restarting hnf-heroku... failed
! Resource not found``

Thanks for any help :)

Update PHP to v5.4.15.

Missing newrelic.so in packaged PHP on S3

Somewhere in the PHP build/packaging process, newrelic.so is left out.

Update Nginx to v1.4.0.

See: http://nginx.org/en/CHANGES-1.4.

Update Nginx to v1.4.2.

http://nginx.org/en/CHANGES-1.4

Update Nginx to v1.2.8.

R12 (Exit timeout) - in newrelic/3.1 branch

Looks like nginx is not exiting and therefore there comes an error:

2013-02-21T23:13:39+00:00 heroku[web.1]: Error R12 (Exit timeout) -> At least one process failed to exit within 10 seconds of SIGTERM
2013-02-21T23:13:39+00:00 heroku[web.1]: Stopping remaining processes with SIGKILL
2013-02-21T23:14:32+00:00 heroku[web.1]: Process exited with status 137
2013-02-21T23:14:33+00:00 heroku[web.1]: State changed from up to down
2013-02-21T23:15:20+00:00 heroku[router]: at=error code=H21 desc="Backend connection refused" method=GET path=/ host=*********de fwd="************" dyno=web.1 queue=0ms wait=3ms connect=4ms service= status=503 bytes=
...
2013-02-21T23:19:23+00:00 heroku[router]: at=error code=H21 desc="Backend connection refused" method=GET path=/ping ...
(later)
2013-02-21T23:22:33+00:00 heroku[web.1]: State changed from up to down
2013-02-21T23:22:33+00:00 heroku[web.1]: State changed from up to down
2013-02-21T23:22:33+00:00 heroku[web.1]: State changed from down to starting
2013-02-21T23:23:44+00:00 heroku[router]: at=error code=H20 desc="App boot timeout" method=GET path=/ping...

I had to "heroku ps:restart" the app by hand to solve the problem ;-( Will have an eye on this issue - maybe it's something more seriously cause I recently changed to this buildpack and never had issues like that before with the default php/apache buildpack.

error while loading shared libraries: libjpeg.so.62

While trying this buildpack with dokku I got this error:

remote: -----> Installing dependencies using Composer
remote:        Fetching composer.phar
remote:        http://getcomposer.org/composer.phar
remote:        Running: php composer.phar install
remote: /build/app/vendor/php/bin/php: error while loading shared libraries: libjpeg.so.62: cannot open shared object file: No such file or directory

Not sure if it is reproducible on heroku.

Include config snippets within nginx.conf.

Allow app-specific configuration snippets by way of includes into nginx.conf.

This way, buildpack nginx.conf improvements will flow downstream without much reconfiguration.

Use Vulcan to build & package PHP

Refer to https://github.com/marzocchi/vulcan-php.

Nginx 1.2.6

Version bump.

Set $PATH via .profile.d mechanism

A Heroku lab feature allows the slugc to source .profile.d/*.sh files.

There are considerations to be made when setting PATH using this method.

Should the build pack path be prepended or appended to the value specified by heroku config?

References:
http://blog.iphoting.com/blog/2012/05/24/running-php-on-heroku/#comment-732414477
http://stackoverflow.com/a/13804801/942074
https://devcenter.heroku.com/articles/labs-dot-profile-d

Update Nginx to v1.2.7.

New version released.

Specify `user` in php-fpm.conf

In order to get FPM to launch with dokku, I had to specify user = www-data in php-fpm.conf. It would be great to get that added - assuming it still works with heroku.

Update PHP to v5.4.17.

sh: php: not found

I just pushed a php application on heroku, everythings gets updated including composer dependencies. But i got following error after all

Writing lock file
Generating autoload files
sh: php: not found
Script php artisan optimize handling the post-install-cmd event returned with an error
[RuntimeException]
Error Output: sh: php: not found

Can you help me resolve this issue please. Thanks in advance.