* Cleanup app state on initialization error
When returning an initialization error, the app state was not cleaned
up. This is especially visible during tests, as `appCount` is not
decremented, and makes the new app initialization fail.
* Test the `platform server` command
As the `platform server` command only exits when interrupted by
a signal, it is not possible to test it as the other cobra
commands. Instead we directly test the actual command function.
The internal command handler is slighly refactored to take
a channel in argument, and registers it as the signal handler.
Nothing very different—except than controlling this channel
from the outside allows the test to send the system signal
itself, thus preventing the server to run forever.
Only a handful of critical errors are present in the codebase.
They all occur during server startup (in `app.StartServer()`).
Currently, when one of these critical error occurs, it is simpled
mentionned in the logs – then the error is discarded, and the app
attempts to continue the execution (and probably fails pretty quickly in
a weird way).
Rather than continuing operations in an unknow state, these errors should
trigger a clean exit.
This commit rewrites critical startup errors to be correctly
propagated, logged, and then terminate the command execution.
Additionnaly, it makes the server return a proper error code to the
shell.
When starting the server using `platform server`, errors occuring during
startup are not reported in the console. The command exit with a 0 exit
code (i.e. "success"), although the server failed to launch.
With this change, when an error occurs while initializing the app (like
a missing or invalid configuration file):
- the error is printed to the console;
- the command exit with a "-1" exit code.
This allow shell scripts to properly detect the startup failure, and to
react to it.
Example of error displayed:
```
$ platform server
Error: LoadConfig: Error decoding config file=config.json, err=While parsing config: invalid character ':' after top-level value,
``
* don't use global app for api / api4 tests
* put sleep back. we're gonna have to do some goroutine wrangling
* fix oauth test config assumptions
* jobs package, i'm comin' for you next
* app test fix
* try increasing sleep a little
* tie back-end plugins together
* fix comment typo
* add tests and a bit of polish
* tests and polish
* add test, don't let backend executable paths escape the plugin directory
* Prepare file upload infrastructure for Data Retention.
This commit prepares the file upload infrastructure for the data
retention feature that is under construction. Changes are:
* Move file management code to utils to allow access to it from jobs.
* From now on, store all file uploads in a top level folder which is the
date of the day on which they were uploaded.
This commit is based on Harrison Healey's branch, but updated to work
with the latest master.
* Use NewAppError
* Fixes Elasticsearch to have consistent capitalisation everywhere
across the code and UI (except the config file unfortunately).
* Adds basic unit tests for Elastichsearch.
* Adds a Elasticsearch docker image to the Makefile to enable testing
the Elasticsearch feature.
This commit redesigns the jobserver to be based around an architecture
of "workers", which carry out jobs of a particular type, and "jobs"
which are a unit of work carried by a particular worker. It also
introduces "schedulers" which are responsible for scheduling jobs of a
particular type automatically (jobs can also be scheduled manually when
apropriate).
Workers may be run many times, either in instances of the platform
binary, or the standalone jobserver binary. In any mattermost cluster,
only one instance of platform OR jobserver must run the schedulers. At
the moment this is controlled by a config variable, but in future will
be controlled through the cluster leader election process.
* PLT-6471 Properly panic when translations can't be loaded
* Print usage messages when errors occur during CLI initialization
* Reverted behaviour of FindDir and added second return value to it
* Fixed merge conflict
* Added initial job server
* Added job server to be ran as part of platform
* Added test job to the enterprise repo
* Fixed job server not loading license
* Renamed job package to jobs
* Fixed TE not being buildable
* Added JobStatus table to database
* Changed fields used by JobStatus
* Added APIs to query job status
* Added config change listener to server
* Added option to run job server from Makefile
* Added ability to enable/disable jobs from config
* Commented out placeholder for search indexing job
* Fixed govet
* Removed debug messages and fixed job api init message
* PLT-6472: Basic Elastic Search implementation.
This currently supports indexing of posts at create/update/delete time.
It does not support batch indexing or reindexing, and does not support
any entities other than posts yet. The purpose is to more-or-less
replicate the existing full-text search feature but with some of the
immediate benefits of using elastic search.
* Alter settings for AWS compatability.
* Remove unneeded i18n strings.
