(Not sure if this is the right component, feel free to move as appropriate.) I'm seeing the Gecko Decision Task for my Try pushes fail frequently, with an error of the form: abort: repository /home/worker/checkouts/gecko: timed out waiting for lock held by abd5439b1213:7 Example: https://treeherder.mozilla.org/#/jobs?repo=try&revision=886618c4e475d310e9e5e5177ee1a9899e3f3dd1 This has happened for me at least 6 times in the last few hours; I've had to resubmit the majority of my Try pushes, sometimes multiple times, to get them to actually schedule builds.

This bug effectively results in cache poisoning on the worker. The worker's cache needs to be nuked or the worker needs to be killed. Making noise in #taskcluster about a bad worker is how this gets resolved until bug 1297153 is fixed and deployed.