sysfs test in ubuntu_stress_smoke test failed on B-AWS ARM64

Bug #1819447 reported by Po-Hsu Lin
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
ubuntu-kernel-tests
Invalid
Medium
Colin Ian King

Bug Description

kernel: 4.15.0-1034-aws Bionic

We have 3 ARM64 instances for AWS, 2 of them failed with this test (a1.2xlarge, a1.large):

sysfs STARTING
sysfs RETURNED 2
sysfs FAILED
stress-ng: debug: [14785] 2 processors online, 2 processors configured
stress-ng: info: [14785] dispatching hogs: 4 sysfs
stress-ng: debug: [14785] cache allocate: reducing cache level from L3 (too high) to L2
stress-ng: debug: [14785] cache allocate: default cache size: 2048K
stress-ng: debug: [14785] starting stressors
stress-ng: debug: [14786] stress-ng-sysfs: started [14786] (instance 0)
stress-ng: debug: [14785] 4 stressors spawned
stress-ng: debug: [14788] stress-ng-sysfs: started [14788] (instance 2)
stress-ng: debug: [14789] stress-ng-sysfs: started [14789] (instance 3)
stress-ng: debug: [14787] stress-ng-sysfs: started [14787] (instance 1)
stress-ng: error: [14787] stress-ng-sysfs: A SIGSEGV occurred while exercising /sys/bus/pci/slots/15/attention, aborting
stress-ng: debug: [14787] stress-ng-sysfs: exited [14787] (instance 1)
stress-ng: debug: [14786] stress-ng-sysfs: exited [14786] (instance 0)
stress-ng: debug: [14789] stress-ng-sysfs: exited [14789] (instance 3)
stress-ng: debug: [14785] process [14786] terminated
stress-ng: debug: [14788] stress-ng-sysfs: exited [14788] (instance 2)
stress-ng: error: [14785] process 14787 (stress-ng-sysfs) terminated with an error, exit status=1 (stress-ng core failure)

Po-Hsu Lin (cypressyew)
description: updated
Changed in ubuntu-kernel-tests:
assignee: nobody → Colin Ian King (colin-king)
importance: Undecided → Medium
status: New → In Progress
Revision history for this message
Colin Ian King (colin-king) wrote :

I've run this several times and I can't get this to reproduce. The failure is a segfault in stress-ng, and there stress test should report if there are any kernel messages; however no messages were emitted. So I'm really not sure why this is failing.

Tested using:

1. stand-alone stress-ng with this specific stressor on all CPUs
2. run just the kernel stress smoke test (several times)
3. run the dep8 tests stand-alone using:

git clone git://kernel.ubuntu.com/ubuntu/kernel-testing
kernel-testing/run-dep8-tests

Can't reproduce. Can we re-run these tests to see the failure rates?

Revision history for this message
Sean Feole (sfeole) wrote :

I have removed the auto-cleanup features in this test, and triggered a re-run, if it indeed fails again we can log in and have a look.

Revision history for this message
Sean Feole (sfeole) wrote :

I re-ran this test 4 times, all 4 times I was unable to reproduce the bug. All 4 times have successfully passed. Strange that this failed a series of two times, now completely problem free. I will take a look at the 2nd instance flavor type.

Revision history for this message
Colin Ian King (colin-king) wrote :

Still can't trip the issue.

Revision history for this message
Po-Hsu Lin (cypressyew) wrote :

Since we're all unable to reproduce this, I would say let close this bug and keep an eye on this test in the future.

Changed in ubuntu-kernel-tests:
status: In Progress → Invalid
Revision history for this message
Po-Hsu Lin (cypressyew) wrote :

Saw this with a1.large on D-AWS:
        sysfs FAILED
        stress-ng: debug: [16588] 2 processors online, 2 processors configured
        stress-ng: info: [16588] dispatching hogs: 4 sysfs
        stress-ng: debug: [16588] cache allocate: reducing cache level from L3 (too high) to L2
        stress-ng: debug: [16588] cache allocate: default cache size: 2048K
        stress-ng: debug: [16588] starting stressors
        stress-ng: debug: [16589] stress-ng-sysfs: started [16589] (instance 0)
        stress-ng: debug: [16588] 4 stressors spawned
        stress-ng: debug: [16590] stress-ng-sysfs: started [16590] (instance 1)
        stress-ng: debug: [16592] stress-ng-sysfs: started [16592] (instance 3)
        stress-ng: debug: [16591] stress-ng-sysfs: started [16591] (instance 2)
        stress-ng: error: [16590] stress-ng-sysfs: A SIGSEGV occurred while exercising /sys/fs/cgroup/devices/init.scope/notify_on_release, aborting
        stress-ng: debug: [16590] stress-ng-sysfs: exited [16590] (instance 1)
        stress-ng: error: [16589] stress-ng-sysfs: A SIGSEGV occurred while exercising /sys/devices/platform/ACPI0007:08/uevent, aborting
        stress-ng: debug: [16589] stress-ng-sysfs: exited [16589] (instance 0)
        stress-ng: error: [16588] process 16589 (stress-ng-sysfs) terminated with an error, exit status=1 (stress-ng core failure)
        stress-ng: debug: [16588] process [16589] terminated
        stress-ng: error: [16588] process 16590 (stress-ng-sysfs) terminated with an error, exit status=1 (stress-ng core failure)
        stress-ng: debug: [16588] process [16590] terminated
        stress-ng: debug: [16592] stress-ng-sysfs: exited [16592] (instance 3)
        stress-ng: debug: [16591] stress-ng-sysfs: exited [16591] (instance 2)
        stress-ng: debug: [16588] process [16591] terminated
        stress-ng: debug: [16588] process [16592] terminated
        stress-ng: info: [16588] unsuccessful run completed in 5.00s

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.