Can't easily force rebuild of AfeClient or TkoClient
Reported by
jrbarnette@chromium.org,
Feb 22 2018
|
||
Issue descriptionThe "site_utils/deploy_server_local.py" includes a `--force_update` option that can be used to force rebuilding the GWT client code for either the TKO or the AFE, even if they don't appear to need it. That's a good thing. Unfortunately, there seems to be no easy way to force that option to be passed during automated deploy. Mercifully, the "site_utils/deploy_server.py" does include a "-x" option that can be used for the purpose. However, that command isn't easily invoked for the purpose: If you reach the point where you need to force all shards and the master, and every AFE server to update, you must manually gather up the names of all the affected servers, and construct the command line for yourself. All of that is too much work: The automated_deploy script should know how to arrange for --force_update to be passed to every server from its own command line.
,
Feb 22 2018
The automated_deploy.py is used for regular push_to_prod, it equals to update_prod_branch + deploy_server.py. It will update the prod branch, then trigger the deploy_server.py. The deploy_local_server.py then decides whether to run certain updated based whether there are changes to the corresponding pre-defined repo in this pushed batch. So, automated_deploy.py is just a wrapper of deploy_server.py and this script is *only* used for regular push to prod. All other occasional servers updates without pushing prod branch should use deploy_server.py.
,
Feb 22 2018
|
||
►
Sign in to add a comment |
||
Comment 1 by nxia@chromium.org
, Feb 22 2018