Description of problem: When scheduled lots of tasks in async, we occasionally get into "unresponsiveness" issue. To avoid/workaround this, we can not do async things and just wait for the tasks and do things one by one. How reproducible: nondeterministic Steps to Reproduce: 1. import multiple larger Red Hat repos Actual results: machine stops responding Expected results: ...
Since this issue was entered in Red Hat Bugzilla, the release flag has been set to ? to ensure that it is properly evaluated for this release.
By specifying a max. number of parallel tasks, we may reduce the sat6 load. By setting it to '1', the tasks would be run one by one.
Yeah, that was my thought as well - lets us put it under user-control, too.
First take at this... hammer-cli-import.git: 80dde679c2eec586b2114df34c727a966c1c2b9f
Extend code so it is accessible... hammer-cli-import.git: ca1889f8d3fbbf54b6889dbf19fc795a1774ea8b
Need this for GA - running as-asynch-as-possible tends to cause unpredictable behavior in Sat6, customers will need to be able to throttle
DEVELOPER VERIFIED: 1) start with a Sat6 that has its orgs and manifests, which has not sync'd any Red Hat content yet, and a channels.csv that has multiple Red Hat channels. 2) hammer import repository-enable --synchronize --wait 3) In the UI, go to <sat6>/katello/sync_management and Expand All 4) Note that *all* the repositories have sync-tasks started Next: 1) start with a Sat6 that has its orgs and manifests, which has not sync'd any Red Hat content yet, and a channels.csv that has multiple Red Hat channels. 2) hammer import repository-enable --synchronize --wait --no-async 3) In the UI, go to <sat6>/katello/sync_management and Expand All 4) Note that at any given moment, only one repository is synchronizing, and that a new one starts when the first finishes. Verified on rubygem-hammer_cli_import-0.10.4-1.1.el6sat.noarch
This was delivered with Satellite 6.0 which was released on 10 September 2014.