Public bug reported: A team within Microsoft is running the linux-azure Ubuntu kernel on a large AI cluster. They are hitting an issue which is resolved by the following commit:
https://github.com/torvalds/linux/commit/ebaf39e6032faf77218220707fc3fa22487784e0 The bug is that threads can get stuck in the kernel. This bug can happen when removing network namespaces. This is something docker swarm does anytime it removes a container. Commit ebaf39e6032f was added to the mainline kernel tree in v4.20-rc6. It was not cc’d to upstream stable, so only v4.20-rc6 and newer kernels will have it. A test kernel was built with this commit, which resolves the issue. ** Affects: linux-azure (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1830266 Title: [linux-azure] Please Include Mainline Commit ebaf39e6032f in the 16.04 and 18.04 linux-azure kernels To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-azure/+bug/1830266/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
