* monitor "schemas" queue to run refresh_schema
`refresh_schema` tasks won't run because "schemas" queue isn't consumed
with default settings.
and it cause leaking redis storage, a "schemas" list is growing with time.
this PR fix it, adds "schemas" queue to monitor by celery.
* use scheduled_worker for "schemas" queue
instead of "adhoc_worker"
Testing script on clean install of Ubuntu Desktop 16.04.1.
pip fails if not upgraded beforehand, whether "sudo" or "sudo -s" or "sudo -sH" is run.
After this modification it works perfectly from a clean install. Below are my steps if anyone should want to replicate or validate:
1. Clean install of Ubuntu 16.04.1 Desktop (on VM)
2. sudo apt-get -y install git
3. mkdir ~/git
4. cd ~/git
5. git clone https://github.com/getredash/redash
6. cd ~/git/redash/setup/ubuntu
7. chmod +x bootstrap.sh
8. nano bootstrap.sh
9. added "pip install --upgrade pip" to the script above the "pip install -U..." line.
10. Saved script.
11. sudo -H ./bootstrap.sh
Note that the "-H" is necessary in order to run the script successfully, otherwise you will run into pip ownership issues.
- Update version to 0.11.1.
- Move all apt commands to the beginning of the script and add the missing ones (lib-sasl).
- Fix the Redash Metadata data source creation command.
Because some dependencies were missing and we dont really need everything in the file...just pymongo
Also removed unneeded files, reverted circle.yml to as it is on redash, so it doesnt show in the PR diff
Also reset circle.yml to be as in the main redash repo so our changes are not included in the changeset
Check earlier PR https://github.com/getredash/redash/pull/1027 to follow conversation.
