phabricator on redhat's openshift - php

I installed phabricator on openshift using a quickstart from github
https://github.com/CodeBlock/phabricator-openshift-quickstart
I got it running up fine, but I now have two issues:
1.) A setup issue that says apc.stat is enabled and that must be disabled in
/var/lib/openshift/my-user-hash/php/configuration/etc/php.ini
however I cannot access that with sudo command(sudo permission denied), even if I open it normally, I didn't find any apc.stat settings in the php.ini
2.) I can't figure out how set the local path for tracking repositories for diffusion. It says
I must give a path which should be read-writable by phabricator, I tried to give the persistent storage location ..data/ , but it gives me an exception as follows :
Unhandled Exception ("CommandException")
Command failed with error #1!
COMMAND
(cd '../data/' && HOME='/var/lib/openshift/my-user-hash/app-root/runtime/repo/phabricator/support/empty/' git cat-file --batch)
STDOUT
(empty)
STDERR
sh: line 0: cd: ../data/: No such file or directory
how do I fix this?

1) According to Num Duong answer, it seems like u currently could not resolve this issue, probably u should wait for openshift php.ini permissions policy changes.
Anyway this is minor non blocking issue.
2) Modify config file and re-deploy to openshift.
Look for available phabricator options here
U need smth like this: 'repository.default-local-path' => getenv('OPENSHIFT_DATA_DIR'),
P.S. Anyway there is one particular issue with phabricator on openshift that possibly never got resolved: cloning git repositories from OpenShift instances into Phabricator do not
work by SSH, due to permissions regarding the ssh configuration.

Related

Running Rscript from within PHP: cannot load shared libraries

I am trying to run R code from within PHP. I am using RHEL 7.4.
However, because Apache runs as a different user than installed the R packages, it cannot load packages:
My PHP code has this line:
echo shell_exec("(Rscript -e 'library(lme4, lib.loc = \"/var/www/R/lib\")')2>&1");
which produces this error in the browser:
Error: package or namespace load failed for 'lme4' in dyn.load(file, DLLpath = DLLpath, ...):
unable to load shared object '/var/www/R/lib/Rcpp/libs/Rcpp.so':
/var/www/R/lib/Rcpp/libs/Rcpp.so: failed to map segment from shared object: Permission denied
Execution halted
The file /var/www/R/lib/Rcpp/libs/Rcpp.so has permissions 775, so I don't think the permissions are actually the problem.
How can I tell Apache / PHP where to look for the shared libraries?
The server is not public-facing, so I would take any solution from properly accessing the shared library, to giving user apache sudo access with NOPASSWD to run as my user (or something similarly reckless, since that didn't actually work...).
Thank you!
This appears due in part to an SELinux setting in Red Hat. I'm not entirely sure whether this is Red Hat-centric, or specific to the configuration at my university (who installed Red Hat in the first place).
Originally:
$ getenforce
Enforcing
Fix:
$ sudo setenforce 0
$ getenforce
Permissive
Persistent fix: modify /etc/selinux/config to have the line
SELINUX=Permissive
Credit to this StackOverflow thread.

Twig Cache Error

I'm getting a Twig error ...
Uncaught exception 'RuntimeException' with message 'Unable to write in the cache directory (./cache/twig/69).
but only ever in that 1 hexadecimal directory listed above. If I then set the permissions on that folder to 0777 and also set the permissions on the files it contains, it works correctly but only for a few days. I then have to do the same fix and so on and so on. None of the other directories in the twig folder are affected. Any ideas?
Check the permission of the folder. It should have write access.
Check SELinux policy
sestatus
To disable SELinux
sudo setenforce 0
Are you running commands from the CLI with a different user than the webserver? This often occurs if you run console commands as e.g. root and they create cache directories. Try clearing the cache completely after each CLI call.

How to deploy Laravel 5.1 application to CentOS 6.4 VPS (or any!) server (PHP 5.5.32)

I have a Laravel 5.1 application that I'm trying to deploy on my CentOS 6.4 VPS server. The problem is that I am getting Error 500 when trying to load the page. When I check the error log it shows me this:
PHP Fatal error: Uncaught exception 'UnexpectedValueException' with message 'The stream or file "/home/pams/pams_todo_list/storage/logs/laravel.log" could not be opened: failed to open stream: Permission denied' in /home/pams/pams_todo_list/vendor/monolog/monolog/src/Monolog/Handler/StreamHandler.php:87
I have changed the permissions in pams_todo_list/storage and pams_todo_list/vendors to 777 instead of 755 as a test, but no love.
This problem has been solved, thanks to FFForever and lagbox on the Laravel IRC channel.
My owner permissions were wrong. I had inadvertently uploaded the project using the root account. The fix was two fold. First, I setup suPHP as my php handler (I do have root access to this server) through the EasyApache3 UI in WHM. Then I set the owner:
chown -R pams:pams pams_todo_list
Now it works. I cannot say if it would have worked without setting up suPHP, however it was strongly recommended to set that up, so I did.

Drush command(s) run on Drupal 6 sites, but not Drupal 7?

I am running into an issue with a customer's Drupal sites. He has a number of D6 installs, and a new D7 that he's just starting on. All of these sites are on the same shared hosting package.
The problem is when running certain drush commands only on the D7 site. There are no issues on the D6 sites. The specific error for drush up on the D7 site follows:
foo#bar [~/www/foo]# drush up
Command pm-update needs a higher bootstrap level to run - you will need to invoke drush [error]
from a more functional Drupal environment to run this command.
Command pm-update needs the following modules installed/enabled to run: update. [error]
The drush command 'up' could not be executed. [error]
Drush was not able to start (bootstrap) the Drupal database. [error]
Hint: This may occur when Drush is trying to:
* bootstrap a site that has not been installed or does not have a configured database. In
this case you can select another site with a working database setup by specifying the URI
to use with the --uri parameter on the command line. See `drush topic docs-aliases` for
details.
* connect the database through a socket. The socket file may be wrong or the php-cli may
have no access to it in a jailed shell. See http://drupal.org/node/1428638 for details.
Drush was attempting to connect to:
Drupal version : 7.28
Site URI : http://default
Database driver : mysql
Database username : username_foo
Database name : database_foo
PHP configuration :
PHP OS : Linux
Drush version : 7.0-dev
Drush temp directory : /tmp
Drush configuration :
Drush alias files :
Drupal root : /home/foo/www/foo
Site path : sites/default
Everything I can find (and indeed the link in the error message) say the solution is to change the host value in the settings.php file from localhost to 127.0.0.1. However, this has not been the solution for us.
The frontend site has no problems connecting to the database, and drush itself seemingly does in many cases too.
drush sql-connect will generate a string that you can use to connect to MySQL.
drush sql-cli will successfully connect to MySQL
The settings.php file is definitely in a folder called default (path: www/foo/sites/default) and I get the exact same error when specifying --root and --uri options.
Drush was originally a ~5.x release, and it had the same issues. We updated to the ~7.x to try to eliminate the error. The host value in settings.php has been localhost as well as 127.0.0.1 with equally poor results.
We've verified that MySQL is available via socket from the PHP CLI. Drush is up to date and a fresh install. The Drupal 7 site is a brand new fresh install.
I'm at a loss. Why would this work with the D6 sites, but not the D7? Any suggestions?
Is your settings.php file in a folder called 'default'? If not, you might need to tell Drush where to find it by using --uri=mysite.com or -l mysite.com.
There are several ways to specify which Drupal site Drush will target. The most basic option is fairly verbose; run:
$ drush --root=/path/to/drupal --uri=http://example.com status
You can do the same thing with a slightly different syntax:
$ drush /path/to/drupal#example.com status
You can also specify the Drupal site implicitly, by setting the cwd to the folder that contains the settings.php file for your site:
cd /path/to/drupal/sites/default # or /path/to/drupal/sites/mysite.com, as appropriate
$ drush status
In all of the cases above, if settings.php is in a folder called "default", then you do not need to specify the --uri component; you may, for example, cd /path/to/drupal followed by drush status, and the correct Drupal site will be found. If settings.php is not in a folder named 'default', then you will need to either specify --uri, or cd to the folder that contains the settings.php file.
Source
According to this message:
pm-update needs the following modules installed/enabled to run: update
Drush requires Update module to be enabled, so the following command should fix the problem:
drush -y en update

Updating PHP in fedora Instance on AMAZON EC2

I have a fedora Instance runnig in an Amazon EC2. The problem is, I have a php already installed, but I want to update (PS: Somebody already gave to me the instance with the php installed). When I try the command: yum update php an error message appears.
Setting up Update Process
Setting up repositories
Cannot find a valid baseurl for repo: updates-released
Error: Cannot find a valid baseurl for repo: updates-released
I tried to uncomment the baseurl line in the yum repositorie files, but a new error appears.
Setting up Update Process
Setting up repositories
http://download.fedora.redhat.com/pub/fedora/linux/core/updates/4/i386/repodata/repomd.xml: [Errno 4] IOError: <urlopen error (-2, 'Name or service not known')>
Trying other mirror.
Cannot open/read repomd.xml file for repository: updates-released
failure: repodata/repomd.xml from updates-released: [Errno 256] No more mirrors to try.
Error: failure: repodata/repomd.xml from updates-released: [Errno 256] No more mirrors to try.
The error that you are getting sounds like it's a connection or DNS related issue -- see if you can ping 'download.fedora.redhat.com,' then see if you can use wget to attempt to get the file.
Also, your /etc/yum.repos.d/ directory may contain a repository file with an invalid repository in it -- I'm not sure if they're still publishing updates for Fedora 4 (it was released around 2004ish). Your best bet is to make sure that the version in /etc/fedora-release matches with the version in that repository configuration file.

Categories