# VTK 9.1 tests failing on Alpine Linux
**URL:** https://discourse.vtk.org/t/vtk-9-1-tests-failing-on-alpine-linux/7250
**Category:** Support
**Created:** [November 22, 2021, 11:50pm UTC](https://discourse.vtk.org/t/vtk-9-1-tests-failing-on-alpine-linux/7250 "2021-11-22T23:50:21Z")
**Posts on this page:** 4
**Page:** 1
### Author:  [@boomanaiden154](https://discourse.vtk.org/u/boomanaiden154)
#### Post date: [November 22, 2021, 11:50pm UTC](https://discourse.vtk.org/t/vtk-9-1-tests-failing-on-alpine-linux/7250/1 "2021-11-22T23:50:21Z")
When compiling and running the tests for VTK 9.1 on Alpine Linux, there are about 20 tests that fail, and they are the following:
```auto
230 - VTK::IOImportCxx-OBJImport-MixedOrder1 (Failed)
232 - VTK::IOImportCxx-OBJImport-MTLwithoutTextureFile (Failed)
543 - VTK::RenderingVolumeCxx-TestGPURayCastMapperRectilinearGrid (Failed)
635 - VTK::FiltersSelectionCxx-TestLinearSelector3D (Failed)
674 - VTK::FiltersHyperTreeCxx-TestHyperTreeGridBinaryEllipseMaterial (Failed)
681 - VTK::FiltersHyperTreeCxx-TestHyperTreeGridTernary3DAxisClipBox (Failed)
695 - VTK::FiltersHyperTreeCxx-TestHyperTreeGridTernary3DDualContour (Failed)
702 - VTK::FiltersHyperTreeCxx-TestHyperTreeGridTernary3DPlaneCutterDual (Failed)
712 - VTK::FiltersHyperTreeCxx-TestHyperTreeGridToDualGrid (Failed)
786 - VTK::RenderingOpenGL2Cxx-TestCoincident (Failed)
915 - VTK::ChartsCoreCxx-TestChartBadPoints (Failed)
919 - VTK::ChartsCoreCxx-TestChartMatrix (Failed)
921 - VTK::ChartsCoreCxx-TestChartMatrix3 (Failed)
962 - VTK::ChartsCoreCxx-TestPlotMatrix (Failed)
1150 - VTK::RenderingCoreCxx-TestEdgeFlags (Failed)
1190 - VTK::RenderingCoreCxx-TestTextureRGBADepthPeeling (Failed)
1280 - VTK::FiltersGeometryCxx-TestLinearToQuadraticCellsFilter (Failed)
1310 - VTK::FiltersGeneralCxx-TestDensifyPolyData (Failed)
1347 - VTK::FiltersGeneralCxx-TestYoungsMaterialInterface (Failed)
```
Most of are failing seem t be failing on image comparisons with minor differences, such as `VTK::FiltersHyperTreeCxx-TestHyperTreeGridBinaryEllipseMaterial`.
Actual:
Difference:
I am building VTK with the following configuration:
```auto
cmake -B build \
-Wno-dev \
-DCMAKE_SKIP_INSTALL_RPATH=ON \
-DCMAKE_INSTALL_LIBDIR=lib \
-DBUILD_SHARED_LIBS=ON \
-DCMAKE_INSTALL_PREFIX=/usr \
-DBUILD_DOCUMENTATION=OFF \
-DVTK_USE_FFMPEG_ENCODER=ON \
-DModule_vtkIOPDAL=ON \
-DVTK_USE_LARGE_DATA=ON \
-DVTK_QT_VERSION="5" \
-DVTK_PYTHON_VERSION="3" \
-DVTK_CUSTOM_LIBRARY_SUFFIX="" \
-DVTK_INSTALL_INCLUDE_DIR=include/vtk \
-DCMAKE_BUILD_TYPE=None \
-DVTK_MODULE_USE_EXTERNAL_VTK_hdf5=ON \
-DVTK_MODULE_USE_EXTERNAL_VTK_netcdf=ON \
-DVTK_BUILD_TESTING=ON
```
I think the most likely issue here is perhaps some version conflict with the external dependencies, but I don’t have a ton of experience working with VTK. Here is the [test log](https://tpaste.us/xnOr) produced with `ctest -a --rerun-failed --output-on-failure`.
---
### Author:  [@ben.boeckel](https://discourse.vtk.org/u/ben.boeckel)
#### Post date: [November 23, 2021, 12:10am UTC](https://discourse.vtk.org/t/vtk-9-1-tests-failing-on-alpine-linux/7250/2 "2021-11-23T00:10:46Z")
This looks to be a known issue with some graphics drivers. I’ve seen this in our CI where vertical lines end up not rendering correctly. I haven’t had any luck tracking it down other than it was Linux-specific.
The issue is [here](https://gitlab.kitware.com/vtk/vtk/-/issues/18098), but the CDash links have expired since.
Cc: @ken-martin
---
### Author:  [@boomanaiden154](https://discourse.vtk.org/u/boomanaiden154)
#### Post date: [November 23, 2021, 12:35am UTC](https://discourse.vtk.org/t/vtk-9-1-tests-failing-on-alpine-linux/7250/3 "2021-11-23T00:35:46Z")
That would make sense. I’m also running this in a CI context, using non-accelerated rendering with `mesa-dri-gallium` and running `ctest` with `xvfb-run` so that X is available.
---
### Author:  [@ben.boeckel](https://discourse.vtk.org/u/ben.boeckel)
#### Post date: [November 24, 2021, 1:02am UTC](https://discourse.vtk.org/t/vtk-9-1-tests-failing-on-alpine-linux/7250/4 "2021-11-24T01:02:22Z")
What we do is just exclude tests that have these problems in our CI. I don’t think there’s been anyone that’s had time to investigate where the problem(s) actually lie (VTK, driver, or Mesa).