- #644 - Updating requirements.txt with compatible dependencies
- #629 - Add summaries of major_unit and site_type
- #626 - Add summary of bundles in use
- #625 - Add Node Count by Node Type
Resolves:
- #601 - Add options to site_type and pantheon_size
Resolves:
- #603 - Typo in data structure for pantheon_size field
Resolves:
- #601 - Add fields for site_type and pantheon_size
Resolves:
- #597 - Remove functions to add packages into default profiles
Resolves:
- #588 - Remove opcache reset functions
Resolves:
- #586 - When importing an instance from another environment, include the settings array
Resolves:
- #581 - Existing symlink is deleted when a new 'current: false' code asset is deployed
Resolves:
- #507 - Deploy bundles into Express profile
- #579 - Current symlinks are not being removed when code item is patched
- #584 - Use the 'is_current' field as a lock to prevent deleting code assets
Resolves:
- #577 - get_eve() function has an off by one error
Resolves:
- #574 - Increase pager limit
This release reduces the default number of results returned per page from 200 to 25.
Resolves:
- #566 - Error on verify statistics
- #565 - Error on removing orphan statistics task
As part of deployment:
- Add
sn_key
to settings.php
Resolves:
- #559 - Need to add ServiceNow API key to settings.php
As part of deployment:
- Remove
LOCAL_WEB_ROOT
,LOCAL_CODE_ROOT
, andLOCAL_INSTANCE_ROOT
toconfig_servers.py
Resolves:
- #533 - Remove
type
field from site schema and rest of atlas code - #512 - Collapse variables back down
- #520 - Get rid of all references to load balancer or f5
- #530 - Add a value to site object and settings.php for Google Tag Manager
- #541 - Examine backup totals per instance
- #407 - Figure out how to handle max_results
- #515 - {"$ne"} query is returning items that match the string that we are trying to exclude
Resolves:
- #147 - Site sync - Started, can import instance to same sid if is not in use, or to new sid if it is in use.
- #511 - Restore instance does not restore statistics item
- #516 - When taking down an instance, remove webroot symlink
- #523 - Increase deployment speed
- #529 - Cannot deploy homepage files
Resolves:
- #508 - Backup files are not deleted with Atlas backup item
As part of deployment:
- Delete unused database fields if desired:
verification
dict;migration
, andactivation
in thedate
dict. - Check base URLs in config files.
- Add
LOCAL_WEB_ROOT
,LOCAL_CODE_ROOT
, andLOCAL_INSTANCE_ROOT
toconfig_servers.py
Resolves:
- #504 - Remove migration code
- #377 - Change how we deploy code
- #506 - Switch to using rsync for updating instance code
Resolves:
- #494 - Add support to enable memcache via atlas
- #493 - When deploying drupal structure to a new server, NFS mount is wiped
- #502 - Increase time allowed for taking backups
Resolves:
- Security updates for python packages
- #489 - Move drush commands to 'ops' server
This release supports the new infrastructure.
Resolves:
- #248 - Change the way we call commands
- #303 - Pipe Atlas logs to Logstash
- #328 - Add support for pools
- #366 - Figure out how to handle commands endpoint with new drush command syntax
- #370 - Switch to new homepage location
- #372 - Support new infrastructure
- #373 - Support pooling
- #393 - Round robin commands
- #414 - Add stat elements for beans_by_type.full_html and nodes_by_type.full_html
- #422 - Support running an array of commands
- #428 - Add protected paths
- #431 - Support static assets.
- #433 - Only update the f5 when working with legacy records
- #434 - Remove pools and use only type
- #446 - Move f5 work to ansible
- #448 - Self restore code
- #449 - Self restore instances
- #450 - Purge code and instances
- #451 - Function to update routing for verified instances
- #452 - Infrastructure Migration
- #454 - Separate drush commands and other commands
- #455 - Add multiple Atlas URLs to the settings file
- #458 - Import backups from current infra
- #461 - Verify that instances can be fully deleted.
Improves backup functionality, adds a logging url to the settings file
Resolves:
- #463 - Update stats for CuBoulder/express#2291
Resolves:
- #456 - Eve returns a different query result when compared to MongoDB
This release adds basic backup and restore functionality.
Resolves:
- #90 - Create resource for backups (aka Make backups work)
- #453 - Remove nightly instance churn.
Need to:
- Add
BACKUP_TMP_PATH
andMEDIA_STORAGE_PATH
toconfig_local.py
Resolves:
- #441 - Add a cookie domain to settings file
- #90 - Create resource for backups
Resolves:
- #437 - Remove commands posting to Slack
- #408 - Get timestamp into atlas.log
- #438 - Updating the current and label meta data for a code asset should not trigger a code-re-add process
This release adds granular deployment options for code assets. The following required fields (with the noted defaults) have been added to code items:
"deploy": {
"registry_rebuild": False,
"cache_clear": True,
"update_database": True
}
When a code item is added to an instance, the above drush commands will be run as indicated by the code asset. If multiple items are added, the commands will be run after all code is changed.
As part of deployment:
- Add a list of users to exclude from emails in
config_local.py
- Ensure
sudo -u [webserver_user] drush
can run without a password.
Resolves:
- #90 - Create resource for backups
- #95 - Periodically check for unused code
- #304 - As a Service Manager, need to be able to deploy profile updates without updb
- #22 - Add field for 'requires updb'
- #311 - Add field for 'requires registry rebuild'
- #312 - Add field for 'requires cache clear'
- #330 - During launch, drupal caches are cleared a lot
- #412 - Return something like
Command [xx:xx:xx] 2 of 16 - Success
in Slack - #10 - Re-add packages to sites when changing meta data of a code item.
- #338 - Exclude the '_id' of the item we are updating as a query parameter
- #401 - Remove test accounts from bundle adding emails
- #220 - Started to review settings.php
Resolves:
- #406 - Stats are getting deleted for available instances.
- #404 - Fixed homepage cron command.
Resolves:
- #403 - Bug with code labels in email messages.
Fixed issue that took down instance on Prod and prevented instances from restoring.
This release restructures Atlas and allows the Eve portion to be run from the command line.
Resolves:
- #371 - Refactor code to run as a proper python application
- #329 - Remove code dependencies Code
- #378 - Switch from APC to OPCache clearing
- #331 - Standardize logging structure
- #397 - Homepage sitemap.xml includes P1
- #383 - Pull stats from logs, not HTTP input
- #368 - Bug in slack post Bug
- #391 - Add
node_revision_total
- #386 - Update DB settings
- #200 - Delete GSA code Instance
- #106 - Check for meta uniqueness in code items
Resolves:
- #392 - When an instance gets taken down, delete it's statistic or update the status
- #394 - Add days_since_last_login as int like days_since_last_edit
Resolves:
- Add fields to support CSE
- Express/#1536 - Use HTTPS everywhere
Need to
- Make sure Atlas user can run
sudo -u [webserver_user] drush
without a password. - Update Celery queues to include a
cron_queue
.
Resolves:
- #53 - Handle no code state better
- #352 - Separate cron properly
- #354 - Atlas package emails get confusing with several bundles
- #356 - Change to run drush as apache
Hotfix:
- #381 - Add --uri to drush commands
Resolves:
- #305 - Support new cron command syntax
- #345 - Remove special handling for courses cron
Need to
- Add variable (
database_password
) toconfig_local.py
. Seeconfig_local.py.example
. - Add variable (
database.port
) toconfig_servers.py
. Seeconfig_servers.py.example
. - Update python packages.
Resolves:
- #282 - Provision often fails, but the instance ends up 'available'
- #317 - Switch to mysql-python connecter
- #326 - Make SO and CE user lists nullable
Resolves:
- #325 - Remove poola-custom
- #315 - Settings files - Conditional for DEV
- #319 - Adding query endpoint
Resolves:
- #313 - Fix clear APC command
- #309 - Switch to content editor
Resolves:
- #306 - Add
drush cc drush
to registry rebuilds
Resolves:
- #272 - Add 'locked' status
- Unbind LDAP
- #282 - Continue to try a squash this bug
Resolves:
- #266 - Allow More Headers In Requests
- #271 - Add user count to statistics
- #270 - Add status to statistics
- #273 - Clean up temp files
- #268 - Statistic objects for stuck 'pending' instances are not cleaning up
- #265 - Verify statistics objects are getting updated
- #251 - Stuck pending instances not being removed
- #269 - Capture all bundles as typed or other
- #282 - Provision often fails, but the instance ends up 'available'. Settings files or DB creation fails. Switched to Fabric templating. Wrapped fabric commands in exception catching.
- #433 - Only update the f5 when working with legacy records
Resolves:
- Instances are being marked as 'available' without databases.
Resolves:
- #245 - Available instances are not replaced each night
- #262 - Automatically clean up stats items that have deleted sites.
- #259 - Need to accept variable_livechat_license_number
- #256 - Code clone need labels
- #251 - Math for age of a pending instance is wrong
Resolves:
- #245 - Available instances are not replaced each night
Support new f5 command syntax. See #204 and #196.
Need to add variable (load_balancer_config_group
) to config_servers.py
. See config_servers.py.example
.
Resolves:
- Removed Celery email variables as that functionality is dprecated in 4.x
- #244 Remove homepage files from Atlas
- #234 Update Eve to 0.7.x
Hot fix to resolve bug in sending email.
This release requires updating Python packages and config_local.py
(see config_local.py.example
).
Resolves:
- #224 - POST cron time to elasticsearch Improvement
- #135 - Remove code to import from Inventory
- #223 - Do not post to Slack on cron success.
- #226 - Correct time calc
- #160 - Atlas should handle the email/take down functionality Improvement
- #190 - Atlas should email the user when a bundle is added Improvement
- #218 - Fix pending site removal Bug
- #215 - Update packages
- #208 - Do not run 'add-packages' when there are no packages to add Bug