With set -e added to the script, this test will fail at this line:
image=`uvt-simplestreams-libvirt query | grep -P $DISTRO.*$ARCHITECTURE`
This is because the grep will return 1 if there is no match and thus
script terminates.
Fixes: 545e719ed34a5d ("UBUNTU: SAUCE: ubuntu_kvm_smoke_test: terminates early when the test has failed")
Reported-by: Sean Feole <email address hidden>
Signed-off-by: Po-Hsu Lin <email address hidden>
UBUNTU: SAUCE: ubuntu_kvm_smoke_test: terminates early when the test has failed
While debugging lp:1929925, Khalid found that the report is a bit
confusing as the script will still try to proceed even with KVM
creation failed at the very beginning. This is what you will see:
+ uvt-kvm create bjf-test release=groovy arch=arm64
uvt-kvm: error: libvirt: unsupported configuration: Emulator
'/usr/bin/qemu-system-aarch64' does not support virt type 'kvm'
+ uvt-kvm wait bjf-test --insecure --ssh-private-key-file
/home/ubuntu/.ssh/id_rsa
uvt-kvm: error: libvirt: Domain not found: no domain with matching
name 'bjf-test'
++ uvt-kvm ip bjf-test
uvt-kvm: error: libvirt: Domain not found: no domain with matching
name 'bjf-test'
+ ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null
-o LogLevel=quiet -i /home/ubuntu/.ssh/id_rsa ubuntu@ uname -a
++ cut '-d' -f 2
+++ uvt-kvm ip bjf-test
uvt-kvm: error: libvirt: Domain not found: no domain with matching
name 'bjf-test'
++ ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null
-o LogLevel=quiet -i /home/ubuntu/.ssh/id_rsa ubuntu@ lsb_release -c
+ RESULT=
+ '[' '' '!=' groovy ']'
+ echo '** ERROR: The codename returned from lsb_release -c was (' ')
and we were expecting '\''groovy'\'''
+ exit 1
+ cleanup
+ uvt-kvm destroy bjf-test
uvt-kvm: error: domain 'bjf-test' not found.
Use set -e to bail out early when anything has failed.
Also redirect the destroy command output to /dev/null to make it less
noisy.
Revert "UBUNTU: SAUCE: ubuntu_kernel_selftests: bump clang to clang-13 for Linux 5.13"
clang-13 is still in -prposed for Impish. As we're not enabling
proposed if ppa information passed with the test request, we won't
be able to get this package for the test. Revert this patch for
the moment until we can get this resolved.
This reverts commit ee65c5d6fbc400e22b943c92e44285109ea08239.
UBUNTU: SAUCE: ubuntu_kernel_selftests: remove workaround for ftrace/kprobe_args_user.tc
This kprobe_args_user.tc in ftrace was disabled due to lp:1869227 it
will be marked as unresolved if not supported and marked as a failure.
With patch b730d668138cb3 ("ftrace/selftest: make unresolved cases
cause failure if --fail-unresolved set") landed (lp:1877958), an
unresolved case will not be treated as a failure unless we requested.
Therefore we can remove the workaround here.
Also note that with patch 1e11b7dbef17bb ("selftests/ftrace: Return
unsupported for the unconfigured features") this test will now be
marked correctly as unsupported.
Signed-off-by: Po-Hsu Lin <email address hidden>
Acked-by: Thadeu Lima de Souza Cascardo <email address hidden>
Signed-off-by: Po-Hsu Lin <email address hidden>
UBUNTU: SAUCE: ubuntu_kernel_selftests: bump clang to clang-13 for Linux 5.13
Linux 5.13 selftests/bpf require clang-13 (see also
selftests/bpf/README.rst::Kernel function call test and Clang version)
to fix a FTBFS:
...
libbpf: elf: skipping unrecognized data section(5) .rodata.str1.1
libbpf: elf: skipping unrecognized data section(5) .rodata.str1.1
libbpf: failed to find BTF for extern 'tcp_cong_avoid_ai' [27] section: -2
Error: failed to open BPF object file: No such file or directory
...
Signed-off-by: Paolo Pisati <email address hidden>
Acked-by: Po-Hsu Lin <email address hidden>
Acked-by: Thadeu Lima de Souza Cascardo <email address hidden>
Signed-off-by: Po-Hsu Lin <email address hidden>
UBUNTU: SAUCE: ubuntu_aufs_smoke_test: disable for impish
Signed-off-by: Paolo Pisati <email address hidden>
Acked-by: Colin Ian King <email address hidden>
Acked-by: Po-Hsu Lin <email address hidden>
Signed-off-by: Po-Hsu Lin <email address hidden>
Add a 5000 bogo-ops maximum limit on each stressor to speed up
the whole test. We now run each stressor for 5 seconds or 5000
bogo-ops and terminate when either max limit is met. Thus faster
systems can complete stress tests faster if they can complete
enough soak bogo-op cycles before the 5 second threshold.
On a typical test server this reduces run time by ~30%
Signed-off-by: Colin Ian King <email address hidden>