def test_kdump_no_specify_cpu(self):
'''
bz: 1654962
polarion_id: RHEL7-58669
'''
for cmd in ['sudo kdumpctl showmem','cat /proc/cmdline','systemctl is-active kdump']:
utils_lib.run_cmd(self, cmd, expect_ret=0)
output = utils_lib.run_cmd(self, 'lscpu', expect_ret=0)
product_id = utils_lib.get_product_id(self)
if utils_lib.is_arch(self, 'aarch64') and not utils_lib.is_metal(self) and float(product_id) < 8.6:
self.skipTest("Cancel as bug 1654962 in arm guest earlier than 8.6 2082405" )
utils_lib.run_cmd(self,
r'sudo rm -rf /var/crash/*',
expect_ret=0,
msg='clean /var/crash firstly')
utils_lib.run_cmd(self, r'sudo sync', expect_ret=0)
self.log.info("Before system crash")
utils_lib.run_cmd(self,
r'find /var/crash',
expect_ret=0,
msg='list /var/crash')
utils_lib.run_cmd(self, "sudo bash -c \"echo c > /proc/sysrq-trigger\"", msg='trigger crash')
utils_lib.init_connection(self, timeout=self.ssh_timeout)
self.log.info("After system crash")
utils_lib.run_cmd(self,
r'find /var/crash',
expect_ret=0,
msg='list /var/crash after crash')
cmd = r'sudo cat /var/crash/*/vmcore-dmesg.txt|tail -50'
utils_lib.run_cmd(self, cmd, expect_ret=0, expect_kw='write_sysrq_trigger')
def test_check_microcode_load(self):
'''
bz: 1607899
des: Don't attempt to perform early microcode update on virtualized guests
This case checks it from dmesg output.
'''
cpu_info = utils_lib.run_cmd(self,
'lscpu',
expect_ret=0,
cancel_not_kw="aarch64",
msg="Only run in x86 platform")
utils_lib.run_cmd(self,
"rpm -qa|grep microcode",
msg='get microcode pkg version')
utils_lib.run_cmd(self,
"rpm -qa|grep linux-firmware",
msg='get linux-firmware pkg version')
cmd = 'dmesg|grep microcode'
if utils_lib.is_metal(self):
utils_lib.run_cmd(self,
cmd,
expect_kw="microcode updated early",
msg='microcode should load in bare metal')
else:
utils_lib.run_cmd(self,
cmd,
expect_not_kw='microcode updated early',
msg='microcode should not load in VMs')
def setUp(self):
utils_lib.init_case(self)
if self.params['remote_node'] is None:
self.skipTest("Only support to run in server-client mode!")
if utils_lib.is_metal(self) or utils_lib.is_ahv(self):
self.ssh_timeout = 1200
self.SSH.interval = 60
else:
self.ssh_timeout = 180
self.log.info('set ssh connection timeout to {}'.format(self.ssh_timeout))
if 'kdump' in self.id():
timeout = 120
interval = 5
time_start = int(time.time())
while True:
cmd = 'sudo systemctl is-active kdump'
ret = utils_lib.run_cmd(self, cmd,ret_status=True, msg='check kdump is active')
if ret == 0: break
time_end = int(time.time())
if time_end - time_start > timeout:
self.log.info('timeout ended: {}'.format(timeout))
break
self.log.info('retry after {}s'.format(interval))
time.sleep(interval)
def test_kdump_no_specify_cpu(self):
'''
bz: 1654962
polarion_id: RHEL7-58669
'''
cmd = 'systemctl is-active kdump'
utils_lib.run_cmd(self, cmd, expect_ret=0, msg='check kdump service')
output = utils_lib.run_cmd(self, 'lscpu', expect_ret=0)
self.is_metal = utils_lib.is_metal(self)
if utils_lib.is_arch(self, 'aarch64') and not utils_lib.is_metal(self):
self.cancel("Cancel as bug 1654962 in arm guest which \
no plan to fix it in the near future!")
utils_lib.run_cmd(self,
r'sudo rm -rf /var/crash/*',
expect_ret=0,
msg='clean /var/crash firstly')
utils_lib.run_cmd(self, r'sudo sync', expect_ret=0)
self.log.info("Before system crash")
utils_lib.run_cmd(self,
r'find /var/crash',
expect_ret=0,
msg='list /var/crash')
utils_lib.run_cmd(self,
'sudo echo c > /proc/sysrq-trigger',
msg='trigger crash')
if self.is_metal:
self.log.info("Wait 180s")
time.sleep(180)
else:
self.log.info("Wait 30s")
time.sleep(30)
utils_lib.init_connection(self)
self.log.info("After system crash")
utils_lib.run_cmd(self,
r'find /var/crash',
expect_ret=0,
msg='list /var/crash after crash')
cmd = r'sudo cat /var/crash/*/vmcore-dmesg.txt|tail -50'
utils_lib.run_cmd(self,
cmd,
expect_ret=0,
expect_kw='write_sysrq_trigger')
def test_check_virtwhat(self):
'''
bz: 1782435
polarion_id: RHEL7-103857
test virt-what, not use systemd-detect-virt
'''
utils_lib.is_cmd_exist(self, cmd='virt-what')
utils_lib.run_cmd(self,
"rpm -q virt-what",
expect_ret=0,
msg='get virt-what version')
virt_what_output = utils_lib.run_cmd(self,
r"sudo virt-what",
expect_ret=0)
lscpu_output = utils_lib.run_cmd(self, 'lscpu', expect_ret=0)
if 'Xen' in lscpu_output:
self.log.info("Found it is a xen system!")
if 'full' in lscpu_output:
self.assertIn('xen-hvm', virt_what_output)
else:
self.assertIn('xen-domU', virt_what_output)
elif 'KVM' in lscpu_output and not utils_lib.is_metal(self):
self.log.info("Found it is a kvm system!")
self.assertIn('kvm', virt_what_output)
elif 'VMware' in lscpu_output:
self.log.info("Found it is a vmware system!")
self.assertIn('vmware', virt_what_output)
elif 'Microsoft' in lscpu_output:
self.log.info("Found it is a Hyper-V system!")
self.assertIn('hyperv', virt_what_output)
elif utils_lib.is_metal(self) and utils_lib.is_aws(self):
self.log.info("Found it is a bare metal system!")
self.assertEqual('aws', virt_what_output.strip('\n'))
elif utils_lib.is_metal(self):
self.log.info("Found it is a bare metal system!")
elif utils_lib.is_aws(self) and utils_lib.is_arch(self,
arch='aarch64'):
self.assertEqual('aws', virt_what_output.strip('\n'))
else:
self.skipTest("Unknow hypervisor")
def setUp(self):
utils_lib.init_case(self)
if not self.vm:
self.skipTest("no vm found!")
if self.params['remote_node'] is None:
self.skipTest("Only support to run in server-client mode!")
if utils_lib.is_metal(self):
self.ssh_timeout = 1200
self.SSH.interval = 60
else:
self.ssh_timeout = 180
self.log.info('set ssh connection timeout to {}'.format(
self.ssh_timeout))
def test_check_tuned_adm_active(self):
'''
case_name:
test_check_tuned_adm_active
case_priority:
1
component:
kernel
bugzilla_id:
1893063
polarion_id:
n/a
maintainer:
[email protected]
description:
Check tuned-adm loads default "virtual-guest" in vm and does not load virtual-guest in metal instance
key_steps:
1. # tuned-adm active
expected_result:
Should not load virtual-guest in bare metal.
Should load virtual-guest in vm by default.
'''
utils_lib.is_cmd_exist(self, cmd='tuned-adm', cancel_case=True)
if 'inactive' in utils_lib.run_cmd(self,
'sudo systemctl is-active tuned'):
utils_lib.run_cmd(self,
'sudo systemctl enable --now tuned',
msg='enable tuned service')
if utils_lib.is_metal(self):
utils_lib.run_cmd(
self,
'tuned-adm active',
expect_not_kw='virtual-guest',
msg='Should not load virtual-guest in bare metal')
else:
utils_lib.run_cmd(self,
'tuned-adm active',
expect_kw='virtual-guest',
msg='Should load virtual-guest in vm by default')
def test_ltp_cpuhotplug(self):
'''
bz: 1464095,1908550
polarion_id: RHEL7-98752
'''
# ltp will considering fail if more than 1 cpus cannot be offline
# in bare metal instance
# but in large metal instances, it is expected. So do not do it in
# bare metal instances
utils_lib.run_cmd(self,
'lscpu',
expect_ret=0,
cancel_not_kw="Xen",
msg="Not run in xen instance as bug 1641510 which is very low priority")
if utils_lib.is_aws(self) and utils_lib.is_metal(self):
self.skipTest('Cancel test as aws bare metal needs 1+ cpus working \
at least which ltp not handle')
else:
self._ltp_run(case_name="cpuhotplug")
def test_check_cpu_vulnerabilities(self):
'''
check if cpu has unexpected Vulnerable
'''
utils_lib.run_cmd(self,
"rpm -qa|grep microcode",
msg='Get microcode version')
utils_lib.run_cmd(self,
"rpm -qa|grep linux-firmware",
msg='get linux-firmware pkg version')
check_cmd = r"sudo grep . /sys/devices/system/cpu/vulnerabilities/* | \
sed 's/:/^/' | column -t -s^"
utils_lib.run_cmd(self, check_cmd, expect_ret=0)
output = utils_lib.run_cmd(self, 'uname -r', expect_ret=0)
if utils_lib.is_metal(self):
self.log.info(
"Bare metal instance should not have any vulnerable (microload loaded)."
)
cmd = r"sudo grep . /sys/devices/system/cpu/vulnerabilities/* | \
sed 's/:/^/' | column -t -s^"
elif 'el7' in output:
self.log.info(
"Skip spec_store_bypass,Retpoline and mds in RHEL7 vms")
check_cmd = r"sudo grep . /sys/devices/system/cpu/vulnerabilities/* | \
grep -v spec_store_bypass|grep -v 'tsx_async_abort'|grep -v 'Vulnerable: Retpoline'|\
grep -v mds| sed 's/:/^/' | column -t -s^"
else:
self.log.info(
"Skip spec_store_bypass and mds,itlb_multihit in vms")
check_cmd = r"sudo grep . /sys/devices/system/cpu/vulnerabilities/*|\
grep -v spec_store_bypass|grep -v 'tsx_async_abort'|grep -v mds|grep -v \
itlb_multihit|sed 's/:/^/' | column -t -s^"
utils_lib.run_cmd(self,
check_cmd,
expect_ret=0,
expect_not_kw='Vulnerable')
def test_ltp_cpuhotplug(self):
"""
case_tag:
kernel
case_name:
test_ltp_cpuhotplug
component:
xen
bugzilla_id:
1464095,1908550
is_customer_case:
False
maintainer:
[email protected]
description:
check if cpu can restore online after offline it
key_steps:
1.check machine is not xen or aws bare metal machine
2.run cpuhotplug ltp test
expect_result:
test should be PASS and no crash seen
debug_want:
dmesg
"""
'''
bz: 1464095,1908550
polarion_id: RHEL7-98752
'''
# ltp will considering fail if more than 1 cpus cannot be offline
# in bare metal instance
# but in large metal instances, it is expected. So do not do it in
# bare metal instances
utils_lib.run_cmd(self,
'lscpu',
expect_ret=0,
cancel_not_kw="Xen",
msg="Not run in xen instance as bug 1641510 which is very low priority")
if utils_lib.is_aws(self) and utils_lib.is_metal(self):
self.skipTest('Cancel test as aws bare metal needs 1+ cpus working \
at least which ltp not handle')
else:
self._ltp_run(case_name="cpuhotplug")
def test_virsh_pci_reattach(self):
'''
case_name:
test_virsh_pci_reattach
case_priority:
1
component:
kernel
bugzilla_id:
1700254
polarion_id:
n/a
maintainer:
[email protected]
description:
Test no exception when system does pci detach and attach operation in virsh.
virsh can detach host pci device and attach it to guest.
key_steps:
1. #virsh nodedev-detach $pci
2. #virsh nodedev-reattach $pci
expected_result:
No panic/crash happen.
eg. # virsh nodedev-detach pci_0000_2b_00_0
Device pci_0000_2b_00_0 detached
# virsh nodedev-reattach pci_0000_2b_00_0
Device pci_0000_2b_00_0 re-attached
'''
utils_lib.is_metal(self, action="cancel")
cmd = "sudo yum install -y libvirt"
utils_lib.run_cmd(self, cmd, msg="install libvirt pkg")
cmd = "sudo systemctl restart libvirtd"
utils_lib.run_cmd(self, cmd, cancel_ret='0', msg="restart libvirtd")
utils_lib.is_cmd_exist(self, cmd='virsh')
if utils_lib.is_arch(self, arch='x86_64'):
utils_lib.run_cmd(self,
'cat /proc/cmdline',
cancel_kw='intel_iommu=on',
msg='Check boot line')
utils_lib.run_cmd(self, 'sudo lspci', msg="get pci list")
tmp_pci = None
cmd = "lspci|grep 'Non-Volatile memory'|wc -l"
out = utils_lib.run_cmd(self, cmd)
if int(out) > 0:
cmd = 'sudo find /sys/devices -name *nvme*n1p1*'
tmp_root = utils_lib.run_cmd(self, cmd, msg="get boot nvme pci")
boot_pci = tmp_root.split('/')[-2]
cmd = 'sudo find /sys/devices -name *nvme*|grep -v %s|\
grep -i pci|grep n1' % boot_pci
ret = utils_lib.run_cmd(self,
cmd,
msg="get test pci",
ret_status=True)
if int(ret) == 0:
tmp_pci = utils_lib.run_cmd(self, cmd, msg="get test pci")
if len(tmp_pci) > 0:
tmp_pci = tmp_pci.split('/')[-4]
else:
tmp_pci = None
else:
tmp_pci = None
if tmp_pci is None:
cmd = 'sudo find /sys/devices -name *ttyS0*|grep [0-9]:[0-9]'
tmp_root = utils_lib.run_cmd(self,
cmd,
msg="try to get ttyS0 pci device")
if len(tmp_root) == 0:
tmp_pci = None
else:
serial_pci = tmp_root.split('/')[-3]
tmp_pci = serial_pci
if tmp_pci is None:
cmd = 'sudo find /sys/devices -name *vga*|grep [0-9]:[0-9]'
tmp_root = utils_lib.run_cmd(self,
cmd,
msg="try to get vga pci device")
if len(tmp_root) == 0:
tmp_pci = None
else:
vga_pci = tmp_root.split('/')[-2]
tmp_pci = vga_pci
if tmp_pci is not None:
self.log.info("Get pci device: {}".format(tmp_pci))
else:
self.skipTest("No pci device found to detach")
tmp_pci = tmp_pci.replace('.', '_')
tmp_pci = tmp_pci.replace(':', '_')
pci_dev_1 = utils_lib.run_cmd(
self,
'sudo virsh nodedev-list|grep %s |tail -1' % tmp_pci,
msg='pick up device to detach')
if pci_dev_1.endswith('1'):
pci_dev_0 = pci_dev_1.rstrip('1') + '0'
utils_lib.run_cmd(self,
'sudo virsh nodedev-detach %s' % pci_dev_0,
msg='detach pci device',
expect_ret=0)
utils_lib.run_cmd(self,
'sudo virsh nodedev-detach %s' % pci_dev_1,
msg='detach pci device',
expect_ret=0)
utils_lib.run_cmd(self,
'sudo virsh nodedev-reattach %s' % pci_dev_1,
msg='reattach pci device',
expect_ret=0)
if pci_dev_1.endswith('1'):
utils_lib.run_cmd(self,
'sudo virsh nodedev-reattach %s' % pci_dev_0,
msg='reattach pci device',
expect_ret=0)
utils_lib.check_log(self,
"error,warn,fail,trace,Trace",
log_cmd='dmesg -T',
cursor=self.dmesg_cursor)
def test_kdump_each_cpu(self):
"""
case_tag:
kdump
case_name:
test_kdump_each_cpu
case_file:
os_tests.tests.test_lifecycle.test_kdump_each_cpu
component:
kdump
bugzilla_id:
1396554
is_customer_case:
False
customer_case_id:
N/A
testplan:
N/A
maintainer:
[email protected]
description:
Test kdump on each cpu core
key_steps: |
1. Triger crash on each cpu core
2. Check if kdump is working and dump file will be generated
expect_result:
kdump is working and dump file will be generated
debug_want:
N/A
debug_want:
N/A
"""
utils_lib.run_cmd(self, 'lscpu', expect_ret=0)
product_id = utils_lib.get_product_id(self)
if utils_lib.is_arch(self, 'aarch64') and not utils_lib.is_metal(self) and float(product_id) < 8.6:
self.skipTest("Cancel as bug 1654962 in arm guest earlier than 8.6 2082405" )
cmd = "grep processor /proc/cpuinfo | wc -l"
cpu_counts = int(utils_lib.run_cmd(self, cmd, expect_ret=0,
msg = "Get cpu counts"))
for core_num in range(cpu_counts):
self.log.info("Trigger kdump on core %d" % core_num)
cmd = "systemctl is-active kdump || sudo systemctl start kdump"
utils_lib.run_cmd(self, cmd, expect_ret=0, msg="check kdump service status")
utils_lib.run_cmd(self,
"sudo rm -rf /var/crash/*",
expect_ret=0,
msg="clean /var/crash")
utils_lib.run_cmd(self, "sudo sync", expect_ret=0)
self.log.info("Before system crash")
res_before = utils_lib.run_cmd(self,
"find /var/crash",
expect_ret=0,
msg="list /var/crash before crash")
cmd = "sudo bash -c 'taskset -c %d echo c > /proc/sysrq-trigger'" % core_num
utils_lib.run_cmd(self, cmd, msg='trigger crash')
time.sleep(30)
self.params['remote_node'] = self.vm.floating_ip
utils_lib.init_connection(self, timeout=self.ssh_timeout)
self.log.info("After system crash")
res_after = utils_lib.run_cmd(self,
"find /var/crash",
expect_ret=0,
msg="list /var/crash after crash")
self.assertNotEqual(res_after, res_before,
"Test failed as no crash dump file found")
cmd = "sudo cat /var/crash/*/vmcore-dmesg.txt|tail -50"
utils_lib.run_cmd(self, cmd, expect_ret=0,
expect_kw="write_sysrq_trigger",
msg="Check if crash happened")