FazBrowse GitHub Viewer | Trending |
URL:
| Home
Tools: [Download Repo ZIP]   [Original HTTPS Page]

Added possibility to set maximum size of output files by pbuehler · Pull Request #10192 · AliceO2Group/AliceO2 · GitHub

/ AliceO2 Public

Added possibility to set maximum size of output files - #10192

Closed
pbuehler wants to merge 2 commits into
AliceO2Group:devfrom
pbuehler:dev
ConversationCommits (2)ChecksFiles changed
Closed

Added possibility to set maximum size of output files#10192
pbuehler wants to merge 2 commits into
AliceO2Group:devfrom
pbuehler:dev

Conversation

pbuehler commented Nov 1, 2022

Copy link
Copy Markdown
Collaborator

Two items have been added to the aod-writer json:
maxfilesize (default value: -1): the approximate maximum size an output file shall reach in MBytes.
resdir (default value: "."): is the directory into which the output files are saved.

Before a DF is saved to disk, the size of all output files is checked to be below maxfilesize. If one of the files is larger than maxfilesize MBytes, then all output files are closed and new ones are created. The files are saved into directories resdir/xxx, where xxx is an increasing counter starting at 1. The test of the file size is disabled if maxfilesize is <= 0. In that case all DFs are saved into one set of files in directory resdir. Directory resdir as well as its subdirectories resdir/xxx are created if they do not exist already.

pbuehler requested a review from a team as a code owner November 1, 2022 10:37

pbuehler commented Nov 1, 2022

Copy link
Copy Markdown
Collaborator Author

Hi @ddobrigk, @jgrosseo
This is the PR which allows to set a maximum size of the output AODs. Please have a look and if possible test it in a realistic setting. I only had the chance to test it locally on a small scale.

alibuild commented Nov 1, 2022

Copy link
Copy Markdown
Collaborator

Error while checking build/O2/fullCI for 3ac2c3b at 2022-11-01 16:43:

## sw/BUILD/o2checkcode-latest/log
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/Common/test/testGPUsortCUDA.cu:22:10: error: 'boost/test/unit_test.hpp' file not found [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUTracking/TRDTracking/GPUTRDTracker.cxx:37:10: error: 'omp.h' file not found [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUTracking/Base/GPUReconstruction.cxx:37:10: error: 'omp.h' file not found [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUTracking/Base/GPUReconstructionCPU.cxx:45:10: error: 'omp.h' file not found [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUTracking/display/GPUDisplay.cxx:36:10: error: 'omp.h' file not found [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUTracking/Base/cuda/GPUReconstructionCUDAGenRTC.cu:16:10: error: 'omp.h' file not found [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/cuda/../Shared/Utils.h:26:10: error: 'boost/program_options.hpp' file not found [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/cuda/../Shared/Utils.h:26:10: error: 'boost/program_options.hpp' file not found [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/../Shared/Utils.h:146:34: error: use of undeclared identifier 'int4'; did you mean 'int'? [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:64:8: error: unknown type name '__host__' [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:64:27: error: expected ';' after top level declarator [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:64:33: error: overloaded 'operator+=' must have at least one parameter of class or enumeration type [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:64:44: error: unknown type name 'int4'; did you mean 'int'? [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:64:53: error: unknown type name 'int4'; did you mean 'int'? [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:66:4: error: member reference base type 'int' is not a structure or union [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:66:11: error: member reference base type 'int' is not a structure or union [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:67:4: error: member reference base type 'int' is not a structure or union [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:67:11: error: member reference base type 'int' is not a structure or union [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:68:4: error: member reference base type 'int' is not a structure or union [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:68:11: error: member reference base type 'int' is not a structure or union [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:69:4: error: member reference base type 'int' is not a structure or union [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:69:11: error: member reference base type 'int' is not a structure or union [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:85:1: error: unknown type name '__global__' [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:90:19: error: use of undeclared identifier 'blockIdx' [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:90:32: error: use of undeclared identifier 'blockDim' [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:90:45: error: use of undeclared identifier 'threadIdx' [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/Kernels.hip.cxx:90:78: error: use of undeclared identifier 'blockDim' [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/../Shared/Utils.h:146:34: error: use of undeclared identifier 'int4'; did you mean 'int'? [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/GPU/GPUbenchmark/hip/benchmark.hip.cxx:199:35: error: use of undeclared identifier 'int4'; did you mean 'int'? [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/Framework/Logger/include/Framework/Logger.h:14:10: error: 'fairlogger/Logger.h' file not found [clang-diagnostic-error]
/usr/local/cuda-11.7/include/cuda/std/detail/libcxx/include/type_traits:520:12: error: CUDA device code does not support variadic functions [clang-diagnostic-error]
/usr/local/cuda-11.7/include/cuda/std/detail/libcxx/include/type_traits:1059:69: error: CUDA device code does not support variadic functions [clang-diagnostic-error]
/usr/local/cuda-11.7/include/cuda/std/detail/libcxx/include/type_traits:1175:5: error: CUDA device code does not support variadic functions [clang-diagnostic-error]
/usr/local/cuda-11.7/include/cuda/std/detail/libcxx/include/type_traits:1852:16: error: CUDA device code does not support variadic functions [clang-diagnostic-error]
/usr/local/cuda-11.7/include/cuda/std/detail/libcxx/include/type_traits:2276:18: error: CUDA device code does not support variadic functions [clang-diagnostic-error]
/usr/local/cuda-11.7/include/cuda/std/detail/libcxx/include/type_traits:4336:16: error: CUDA device code does not support variadic functions [clang-diagnostic-error]
/usr/local/cuda-11.7/include/sm_20_atomic_functions.h:89:39: error: redefinition of 'atomicAdd' [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/Framework/Logger/include/Framework/Logger.h:14:10: error: 'fairlogger/Logger.h' file not found [clang-diagnostic-error]
/usr/local/cuda-11.7/include/cuda/std/detail/libcxx/include/type_traits:520:12: error: CUDA device code does not support variadic functions [clang-diagnostic-error]
/usr/local/cuda-11.7/include/cuda/std/detail/libcxx/include/type_traits:1059:69: error: CUDA device code does not support variadic functions [clang-diagnostic-error]
/usr/local/cuda-11.7/include/cuda/std/detail/libcxx/include/type_traits:1175:5: error: CUDA device code does not support variadic functions [clang-diagnostic-error]
/usr/local/cuda-11.7/include/cuda/std/detail/libcxx/include/type_traits:1852:16: error: CUDA device code does not support variadic functions [clang-diagnostic-error]
/usr/local/cuda-11.7/include/cuda/std/detail/libcxx/include/type_traits:2276:18: error: CUDA device code does not support variadic functions [clang-diagnostic-error]
/usr/local/cuda-11.7/include/cuda/std/detail/libcxx/include/type_traits:4336:16: error: CUDA device code does not support variadic functions [clang-diagnostic-error]
/usr/local/cuda-11.7/include/sm_20_atomic_functions.h:89:39: error: redefinition of 'atomicAdd' [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/Detectors/EMCAL/calibration/include/EMCALCalibration/EMCALCalibExtractor.h:36:10: error: 'omp.h' file not found [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/Detectors/EMCAL/calibration/include/EMCALCalibration/EMCALCalibExtractor.h:36:10: error: 'omp.h' file not found [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/Detectors/EMCAL/calibration/include/EMCALCalibration/EMCALCalibExtractor.h:36:10: error: 'omp.h' file not found [clang-diagnostic-error]
/sw/SOURCES/O2/10192-slc8_x86-64/0/Detectors/TOF/calibration/src/TOFChannelCalibrator.cxx:23:10: error: 'omp.h' file not found [clang-diagnostic-error]

Full log here.

jgrosseo reviewed Nov 2, 2022
View reviewed changes

jgrosseo left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Choose a reason Spam Abuse Off Topic Outdated Duplicate Resolved Low Quality

Hi Paul,
Thanks for this PR. I wonder if it is best to check the file size directly on the file system or if we could also keep a count of the size which is incremented on each tree write. I do it this way in the o2-aod-merger in O2Physics. What do you think?

pbuehler commented Nov 2, 2022

Copy link
Copy Markdown
Collaborator Author

Dear Jan,
To manage the output directories I need to use something like filesystem. So it seemed natural to me to use the filesystem methods to check the file size. But indeed, keeping track of the size could also be done with counters. Are you worried about the performance of file_size method?

jgrosseo commented Nov 2, 2022

Copy link
Copy Markdown
Collaborator

It was about having less file system calls but I think you are right that it is not a big overhead. So I am fine with this. Waiting also for @ktf to comment.

ktf commented Nov 2, 2022

Copy link
Copy Markdown
Member

As discussed privately, the only problem I see with this approach is if we do not write to disk, but to e.g. some XRootd endpoint directly. That said, I would merge this and then we improve, if needed.

pbuehler closed this by deleting the head repository Nov 4, 2022

jgrosseo commented Nov 8, 2022

Copy link
Copy Markdown
Collaborator

@pbuehler @ktf Did we forget to merge this?

ktf commented Nov 8, 2022

Copy link
Copy Markdown
Member

Apparently so. I assume @pbuehler rebased dev for some other development. Anyways, recovered in:

#10232

This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters. Learn more about bidirectional Unicode characters
Show hidden characters
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

jgrosseo jgrosseo left review comments

Assignees

No one assigned

Labels

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

4 participants

Back | FazBrowse Home | New Git URL