Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Collect addional info in mixin #206

Open
wants to merge 6 commits into
base: main
Choose a base branch
from
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
55 changes: 54 additions & 1 deletion eessi/testsuite/eessi_mixin.py
Original file line number Diff line number Diff line change
@@ -1,11 +1,13 @@
from reframe.core.builtins import parameter, run_after, variable
from reframe.core.exceptions import ReframeFatalError
from reframe.core.exceptions import ReframeFatalError, SanityError
from reframe.core.pipeline import RegressionMixin
from reframe.utility.sanity import make_performance_function
import reframe.utility.sanity as sn

from eessi.testsuite import hooks
from eessi.testsuite.constants import DEVICE_TYPES, SCALES, COMPUTE_UNIT, TAGS
from eessi.testsuite.utils import log
from eessi.testsuite import __version__ as testsuite_version


# Hooks from the Mixin class seem to be executed _before_ those of the child class
Expand Down Expand Up @@ -42,6 +44,14 @@ class EESSI_Mixin(RegressionMixin):
bench_name = None
bench_name_ci = None

# Create ReFrame variables for logging runtime environment information
cvmfs_repo_name = variable(str, value='None')
cvmfs_software_subdir = variable(str, value='None')
full_modulepath = variable(str, value='None')

# Make sure the version of the EESSI test suite gets logged in the ReFrame report
eessi_testsuite_version = variable(str, value=testsuite_version)

# Note that the error for an empty parameter is a bit unclear for ReFrame 4.6.2, but that will hopefully improve
# see https://github.com/reframe-hpc/reframe/issues/3254
# If that improves: uncomment the following to force the user to set module_name
Expand Down Expand Up @@ -165,3 +175,46 @@ def assign_tasks_per_compute_unit(self):
def request_mem(self):
"""Call hook to request the required amount of memory per node"""
hooks.req_memory_per_node(self, app_mem_req=self.required_mem_per_node())

@run_after('setup')
def log_runtime_info(self):
"""Log additional runtime information: which CVMFS repo was used (or if it was testing local software),
path to the modulefile, EESSI software subdir, EESSI testsuite version"""
self.postrun_cmds.append('echo "EESSI_CVMFS_REPO: $EESSI_CVMFS_REPO"')
self.postrun_cmds.append('echo "EESSI_SOFTWARE_SUBDIR: $EESSI_SOFTWARE_SUBDIR"')
if self.module_name:
# Get full modulepath, stripping leading spaces with sed and trailing colon by string substitution
get_full_modpath = f'modpath=$(module show {self.module_name} 2>&1 | '
get_full_modpath += 'grep ".lua" | '
get_full_modpath += 'sed "s/^[[:space:]]*//") && echo "FULL_MODULEPATH: ${modpath%:*}"'
Comment on lines +186 to +189
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

this can be simplified using the --location option:

Suggested change
# Get full modulepath, stripping leading spaces with sed and trailing colon by string substitution
get_full_modpath = f'modpath=$(module show {self.module_name} 2>&1 | '
get_full_modpath += 'grep ".lua" | '
get_full_modpath += 'sed "s/^[[:space:]]*//") && echo "FULL_MODULEPATH: ${modpath%:*}"'
# Get full modulepath
get_full_modpath = f'echo "FULL_MODULEPATH: $(module --location show {self.module_name})"'

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Lol, I was looking for something like this option, and couldn't find it. How did I overlook this option, while it's right there in the --help...

self.postrun_cmds.append(get_full_modpath)

@run_after('run')
def extract_runtime_info_from_log(self):
"""Extracts the printed runtime info from the job log and logs it as reframe variables"""
# If EESSI_CVMFS_REPO environment variable was set, extract it and store it in self.cvmfs_repo_name
# Try block is needed to surpress sanity error if there is no match
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
# Try block is needed to surpress sanity error if there is no match
# Try block is needed to suppress sanity error if there is no match

try:
repo_name = sn.extractsingle(r'EESSI_CVMFS_REPO: /cvmfs/(?P<repo>.*)$', f'{self.stagedir}/{self.stdout}',
'repo', str)
if repo_name:
self.cvmfs_repo_name = f'{repo_name}'
except SanityError:
pass
Comment on lines +197 to +203
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

i'd like to make a distinction between 'None', indicating a failing sanity check, and an empty string, indicating that the EESSI CVMFS repo was not used:

Suggested change
try:
repo_name = sn.extractsingle(r'EESSI_CVMFS_REPO: /cvmfs/(?P<repo>.*)$', f'{self.stagedir}/{self.stdout}',
'repo', str)
if repo_name:
self.cvmfs_repo_name = f'{repo_name}'
except SanityError:
pass
try:
self.cvmfs_repo_name = sn.extractsingle(r'EESSI_CVMFS_REPO: /cvmfs/(?P<repo>.*)$', f'{self.stagedir}/{self.stdout}',
'repo', str)
except SanityError:
pass

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm... confused. If EESSI_CVMFS_REPO is empty, it will not match the pattern EESSI_CVMFS_REPO: /cvmfs/(?P<repo>.*)$ either. I.e. both an unset EESSI_CVMFS_REPO and an empty EESSI_CVMFS_REPO will lead to a SanityError. If you don't want that, we should change the regex. But: I'd prefer for it to be the repo name (and not the full /cvmfs/software.eessi.io path) if it is set.

Why is it useful to distinguish in the logged data between an unset and empty EESSI_CVMFS_REPO? Both imply the use of a local software stack anyway...


try:
software_subdir = sn.extractsingle(r'EESSI_SOFTWARE_SUBDIR: (?P<subdir>.*)$',
f'{self.stagedir}/{self.stdout}', 'subdir', str)
if software_subdir:
self.cvmfs_software_subdir = f'{software_subdir}'
except SanityError:
pass
Comment on lines +205 to +211
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

same here


try:
module_path = sn.extractsingle(r'FULL_MODULEPATH: (?P<modpath>.*)$', f'{self.stagedir}/{self.stdout}',
'modpath', str)
print(f"Full modulepath: {module_path}")
if module_path:
self.full_modulepath = f'{module_path}'
except SanityError:
pass
Comment on lines +213 to +220
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

same here

Loading