The ubuntu cloud image is usually of 2GB size and sometimes, due to
network delays, it might take more than the default 600s time to
download it. So, increase it to 1200.
When wget times out, the downloaded file will be in corrupted state.
Lisa should raise exception in this case, instead of returning the
corrupted downloaded path.
* mdatp: Add noexec workaround
LISA has a bug where it won't consider mount permissions before placing
the working path. This example workaround demonstrates a possible fix.
* update Mount to record partition mount options
* MDATP/Mount: Update warning message for non-matching mountpoint info
Updating the warning messages to specify that this condition
is not fatal *for the mdatp test*.
If this code is hoisted outside of the test, it should be fatal.
It indicates that our Mount and DF tools are broken and we cannot
validate the mount options on partitions.
* MDATP/Mount workaround cleanup
- Address more code comments
- add comments to warn and success paths
- re-order if/elseif to put positive case first
- rename functions for clarity
* fix BSD mount options regex
* mdatp: move warn->info level logging
* mdatp: refactor mount/df function names
* mdatp: refactor remount to handle additinal options and BSD/Linux differences in remount commant
* Introduce mshv_install testsuite
comment out kdstub
fix typo
fix import
fix import
disable serial console
update node paths to use Path class
+self
Add logic to copy files into privileged dirs
convert to str
debug logging
return code
* address comments
Use Cp tool
add sha256sum for verification
before and after
fix exe
import Cp
test sleep
* readd kdstub, remove test instrumentation
bump timeout to 120s
use mshv_binpath to locate bins, copy kdstub to appropriate path in efi part
clean up names
log binpath for testing
* drop test_data dir, parse and validate mshv_binpath instead
lint flake8
lint 2
lint 3
lint 4
* Run automatic linter
* collect dmesg for diagnosing mshv driver failure
* Update imports
* flake
* fix up tools invocations to remove repeated code
* Remove Infiniband on_before_deploy
Availability sets are an optimization not a requirement and it is causing problems for Ultra Disk and Availability Feature.
* Infiniband: Require Availability Set
Infiniband tests with two nodes need to have an availability set for them to be able to connect over IB IP address.
ARM images wait for an early console which is only present in the cloud, not in lab environment. Remove this kernel parameter to allow for boot in lab.
This will add 3 parameters related to CLH block perf test
to be run with data disk and can configure the caching
strategy
Signed-off-by: Smit Gardhariya <sgardhariya@microsoft.com>
1. Load VDPA kernel modules for CBLMariner distro.
2. Create VDPA device with proper permission and owner
only for CBLMariner.
Signed-off-by: Smit Gardhariya <sgardhariya@microsoft.com>
Lili Deng
Kameron Carr
lubaihua33
Yi
feng-j678
Purna Pavan Chandra Aekkaladevi
kanchansenlaskar
Nuno Das Neves
Aditya Nagesh
Minghe Ren
Smit Gardhariya
mcgov
Cameron E Baird
mayankdaruka-msft
r-dailey
Siddhartha Mathiharan
Srikanth Myakam
panfengxue
Francisco Munoz
Srikanth Myakam