lxd op.Wait() can hang forever on 3.0.3 |
|||
Issue descriptionI've seen this failure in a couple tests recently where the SSH forwarded port for the webserver doesn't work. Here's the message in the log: 2018-12-04T23:27:11.453080-08:00 ERR vm_cicerone[13070]: GetContainerUsername RPC failed: Deadline Exceeded 2018-12-04T23:27:11.453163-08:00 ERR vm_cicerone[13070]: Failed to get container penguin username for SSH forwarding: Deadline Exceeded Which comes from these logs: https://stainless.corp.google.com/browse/chromeos-autotest-results/263661469-chromeos-test/
,
Dec 7
Sorry...3 current failures...forgot about the hana/elm problem too. :)
,
Dec 7
Sorry haven't had a chance to look closely yet, still dealing with issue 894299 and the follow-on issue 911333.
,
Dec 7
No problem, those other ones are far more important. :)
,
Dec 12
Okay I ended up with a repro for the underlying issue while I was working on refactoring the StartLXDContainer RPC. Anything that calls op.Wait() in tremplin with lxd 3.0.3 can hang forever, even when the operation has finished according to lxc monitor. I haven't had time to dig deeper yet, but we'll revert 3.0.3 for now and fix it once the rootfs shifting stuff is done with.
,
Dec 12
,
Dec 13
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/8469e98c6ac89bda5a85e74842f485e4847fb48d commit 8469e98c6ac89bda5a85e74842f485e4847fb48d Author: Stephen Barber <smbarber@chromium.org> Date: Thu Dec 13 04:10:04 2018 Revert "app-emulation/lxd: add back ec384 patch" This reverts commit 8609a91aabd60880c6db3383f81e4b4dceff78d1. Reason for revert: We are failing tests with LXD 3.0.3 BUG=chromium:912189,chromium:910806 TEST=emerge-tatl lxd Change-Id: I1ebcda7f82dd7fa55fdd843681d53e305e4ddbc4 Reviewed-on: https://chromium-review.googlesource.com/1373924 Commit-Ready: Stephen Barber <smbarber@chromium.org> Tested-by: Stephen Barber <smbarber@chromium.org> Reviewed-by: Chirantan Ekbote <chirantan@chromium.org> [rename] https://crrev.com/8469e98c6ac89bda5a85e74842f485e4847fb48d/app-emulation/lxd/lxd-3.0.3-r3.ebuild [modify] https://crrev.com/8469e98c6ac89bda5a85e74842f485e4847fb48d/app-emulation/lxd/lxd-3.0.3.ebuild [delete] https://crrev.com/41d9bfb0500e8a0437f0d8d673d5dbf77efe6b40/app-emulation/lxd/files/lxd-3.0.3-cert-ec384.patch
,
Dec 13
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/3693113b01774e6bf2dc26bfefe511f045df25f6 commit 3693113b01774e6bf2dc26bfefe511f045df25f6 Author: Stephen Barber <smbarber@chromium.org> Date: Thu Dec 13 04:10:05 2018 Revert "app-emulation/lxd: uprev to 3.0.3" This reverts commit 9591c468995a86c316b8ce7eaf9a4516a53ae88b. Reason for revert: We are failing tests with LXD 3.0.3 BUG=chromium:912189,chromium:910806 TEST=emerge-tatl lxd Change-Id: I43421193015d7d690cbf3d53188875b94293e6e5 Reviewed-on: https://chromium-review.googlesource.com/1373925 Commit-Ready: Stephen Barber <smbarber@chromium.org> Tested-by: Stephen Barber <smbarber@chromium.org> Reviewed-by: Chirantan Ekbote <chirantan@chromium.org> [add] https://crrev.com/3693113b01774e6bf2dc26bfefe511f045df25f6/app-emulation/lxd/files/lxd-3.0.2-dqlite-timeout.patch [add] https://crrev.com/3693113b01774e6bf2dc26bfefe511f045df25f6/app-emulation/lxd/lxd-3.0.2-r5.ebuild [delete] https://crrev.com/8469e98c6ac89bda5a85e74842f485e4847fb48d/app-emulation/lxd/lxd-3.0.3-r3.ebuild [rename] https://crrev.com/3693113b01774e6bf2dc26bfefe511f045df25f6/app-emulation/lxd/lxd-3.0.2.ebuild [modify] https://crrev.com/3693113b01774e6bf2dc26bfefe511f045df25f6/app-emulation/lxd/Manifest [add] https://crrev.com/3693113b01774e6bf2dc26bfefe511f045df25f6/app-emulation/lxd/files/lxd-3.0.2-cert-ec384.patch [add] https://crrev.com/3693113b01774e6bf2dc26bfefe511f045df25f6/app-emulation/lxd/files/lxd-3.0.2-pass-all-usb.patch |
|||
►
Sign in to add a comment |
|||
Comment 1 by jkardatzke@chromium.org
, Dec 7