New issue
Advanced search Search tips

Issue 905466 link

Starred by 3 users

Issue metadata

Status: Assigned
Owner:
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 2
Type: Bug


Show other hotlists

Hotlists containing this issue:
LXD


Sign in to add a comment

VM startup files sometimes with 'lxd waitready returned non-zero'

Project Member Reported by jkardatzke@chromium.org, Nov 14

Issue description

Currently, this is the only reason we are getting failures in the tests for vm.CrostiniStartTime. There's many examples readily available by looking at the test results in stainless. It also happens every now and then in the other VM tests.
 
Relevant part of the failure:
2018-12-01T21:06:40.117669+00:00 ERR VM(5)[13312]:  lxd[120]: lvl=eror msg="Failed to start the daemon: failed to open cluster database: failed to open database: unable to set WAL mode: disk I/O error: disk I/O error (2570)" t=2018-12-01T21:06:39+0000#012
What LXD version are you seeing this with?

We've fixed quite a number of issues with dqlite in 3.0.3 based on us occasionally hitting those in our own CI.
We're seeing it on 3.0.2. I'll go ahead and upgrade us to 3.0.3 and see if this persists.
We see a different message on 3.0.3 but more or less the same effect:

lxd[121]: t=2018-12-05T01:04:01+0000 lvl=eror msg="Failed to start the daemon: failed to open cluster database: failed to open database: unable to switch off syncs: locking protocol (15)"

This looks like https://github.com/lxc/lxd/issues/5248

The test that triggers this is doing startup and shutdown speed measurements, so perhaps our shutdown path isn't giving LXD enough time to stop? https://chromium.googlesource.com/chromiumos/platform/tast-tests/+/master/src/chromiumos/tast/local/bundles/cros/vm/crostini_start_time.go#102

Sign in to add a comment