valepakh opened a new pull request, #6470: URL: https://github.com/apache/ignite-3/pull/6470
https://issues.apache.org/jira/browse/IGNITE-26274 When multiple jobs are submitted at the same time, and they require on demand deploy process (that is, they need a unit and it's not deployed yet on target node), there could be a race happening between updating and getting deployment unit status from the metastorage. When deployment unit is acquired, we check it's status on a node, and when it's `UPLOADING` or doesn't exist, we start downloading it. Downloading is guarded by the tracker, which prevents concurrent downloads. At the end of the download the tracker updates the unit status to `DEPLOYED`. At this moment another thread that got the status just before is was updated sees the status still in the `UPLOADING` state and starts downloading. Download tracker allows it since previous download was just completed. The solution is to check the unit status again inside the download tracker. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: notifications-unsubscr...@ignite.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
