Two changes:
1. Move bin/web to bin/web_unicorn(removed PUMA switching logic)
Introduce new shadow script bin/web. Now we have 3 scripts working
together: bin/web, bin/web_puma, bin/web_unicorn
In bin/web, it checks ENV['USE_WEB_SERVER']:
- if value is 'puma' or value is not set, call bin/web_puma
- if value is 'unicorn', call bin/web_unicorn
- report error if other values given
2. Gitlab rails allow ENV setting to override timeout
- ENV['GITLAB_RAILS_RACK_TIMEOUT'] for service_timeout
- ENV['GITLAB_RAILS_WAIT_TIMEOUT'] for wait_timeout
This adds additional options that make it easier to use this script:
1. It adds `--dry-run` option that only displays Git commands that are
going to be executed.
2. It adds `--remote` option that makes it possible to override Git
remote name.
This allows us (and others) to test drive Puma without it affecting all
users. Puma can be enabled by setting the environment variable
"EXPERIMENTAL_PUMA" to a non empty value.
Was introduced in the time that GitLab still used NFS, which is not
required anymore in most cases. By removing this, the API it calls will
return empty responses. This interface has to be removed in the next
major release, expected to be 12.0.
1. Add support for `RAILS5=1|true` for the `bin/spinach` command.
2. Synchronize used spinach versions both for rails4 and rails5.
For rails5 it was accidently used spinach 0.10.1 instead of 0.8.10.
That brought some problems on running spinach tests.
Example of failure message:
NoMethodError: undefined method `line' for #<Spinach::Scenario:0x000000000c86ba80>
Did you mean? lines
/builds/gitlab-org/gitlab-ce/features/support/env.rb:52:in `before_scenario_run'
Moving the check out of the general requests, makes sure we don't have
any slowdown in the regular requests.
To keep the process performing this checks small, the check is still
performed inside a unicorn. But that is called from a process running
on the same server.
Because the checks are now done outside normal request, we can have a
simpler failure strategy:
The check is now performed in the background every
`circuitbreaker_check_interval`. Failures are logged in redis. The
failures are reset when the check succeeds. Per check we will try
`circuitbreaker_access_retries` times within
`circuitbreaker_storage_timeout` seconds.
When the number of failures exceeds
`circuitbreaker_failure_count_threshold`, we will block access to the
storage.
After `failure_reset_time` of no checks, we will clear the stored
failures. This could happen when the process that performs the checks
is not running.
Added the type attribute to a CHANGELOG entry. When you create a new
entry the software asks for the category of the change and sets the
associated type in the file.
Dumping too many jobs in the same queue (e.g. the "default" queue) is a
dangerous setup. Jobs that take a long time to process can effectively
block any other work from being performed given there are enough of
these jobs.
Furthermore it becomes harder to monitor the jobs as a single queue
could contain jobs for different workers. In such a setup the only
reliable way of getting counts per job is to iterate over all jobs in a
queue, which is a rather time consuming process.
By using separate queues for various workers we have better control over
throughput, we can add weight to queues, and we can monitor queues
better. Some workers still use the same queue whenever their work is
related. For example, the various CI pipeline workers use the same
"pipeline" queue.
This commit includes a Rails migration that moves Sidekiq jobs from the
old queues to the new ones. This migration also takes care of doing the
inverse if ever needed. This does require downtime as otherwise new jobs
could be scheduled in the old queues after this migration completes.
This commit also includes an RSpec test that blacklists the use of the
"default" queue and ensures cron workers use the "cronjob" queue.
Fixesgitlab-org/gitlab-ce#23370
When running Unicorn or Sidekiq in the foreground this change removes
an intermediate /bin/sh process. This makes process supervision in the
GitLab Development Kit more reliable.
This change does not affect Omnibus-GitLab (because there we do not
use these launch scripts). Installations from source do use the launch
scripts but for the standard GitLab init script this change will not
make a difference.
Custom installations using Upstart or Systemd may be affected however,
because under certain configurations these systems count exactly how
many forks happen during process startup, and we are reducing that
number by one here.
Using db:reset ensures existing tables are first dropped. This in turn
ensures that we can drop tables regardless of any foreign key
constraints. While CE currently doesn't have any foreign keys EE defines
the following relation:
remote_mirrors.project_id -> projects.id
MySQL will complain whenever you try to drop the "projects" table first
even when using "DROP TABLE ... CASCADE".
The old invocation only worked by accident because we have a '&'
somewhere in the init script for expediency. When ran from a terminal,
the mail_room daemon process ended up in the session of the terminal.
This commit adds a small wrapper that tries to do the textbook
daemonization steps (double fork, setsid etc.) while also taking
care that the pidfile is written before the 'start' process exits.
Avoid the background_jobs script killing every process with 'sidekiq' in
its argument string (e.g. 'rake gitlab:sidekiq start') by also catching
the number (2) in 'sidekiq 2.14'.
Aleksei Lipniagov
Thong Kuah
Heinrich Lee Yu
Qingyu Zhao
Nick Thomas
Robert Speicher
Alessio Caiazza
Pavel Shutsin
Yorick Peterse
Marin Jankovski
Tiago Botelho
Grzegorz Bizon
James Lopez
Jasper Maes
Jan Provaznik
Robert Speicher
Andrew Newdigate
James Lopez
William George
Zeger-Jan van de Weg
Marcel Amirault
Michael Kozono
Stan Hu
Jacob Vosmaer (GitLab)
Rémy Coutable
blackst0ne
Lin Jen-Shin
Sean McGivern
Bob Van Landuyt
Jacopo
Takuya Noguchi
Mike Greiling
Adam Niedzielski
Kamil Trzcinski
Jacob Vosmaer
Valery Sizov
Dmitriy Zaporozhets
Douwe Maan
Jeroen van Baarsen
Cyril Rohr
Pavel Novitskiy
ling.su
Marin Jankovski