NordicHPC/sonar issues and pull requests

#131 - Fix #126 - report total memory and per-job rss

Pull Request - State: closed - Opened by lars-t-hansen about 1 year ago - 2 comments

#130 - Fix #119 - use a file that is known to be present for a test

Pull Request - State: closed - Opened by lars-t-hansen about 1 year ago - 1 comment

#129 - Simplify interface?

Issue - State: closed - Opened by lars-t-hansen about 1 year ago - 5 comments
Labels: enhancement

#128 - Add cluster identification

Issue - State: closed - Opened by lars-t-hansen about 1 year ago - 2 comments
Labels: enhancement

#127 - Add the possibility of JSON output

Issue - State: closed - Opened by lars-t-hansen about 1 year ago - 1 comment
Labels: enhancement

#126 - More memory data

Issue - State: closed - Opened by lars-t-hansen about 1 year ago - 10 comments
Labels: enhancement

#125 - Link/compare to remora

Issue - State: open - Opened by bast about 1 year ago
Labels: documentation

#124 - Fix #123 - implement a lockfile system

Pull Request - State: closed - Opened by lars-t-hansen about 1 year ago - 7 comments

#123 - Guard against multiple concurrent invocations

Issue - State: closed - Opened by lars-t-hansen about 1 year ago - 4 comments
Labels: enhancement

#122 - Containers, Java and things of that ilk

Issue - State: open - Opened by lars-t-hansen over 1 year ago - 2 comments
Labels: question

#121 - Can we read /proc/PID/cmdline to get more information?

Issue - State: open - Opened by lars-t-hansen over 1 year ago - 1 comment
Labels: question, Later

#120 - Fix #100 - obtain timestamp early

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#119 - test command::test_safe_command fails when running locally but seems to pass on GitHub

Issue - State: closed - Opened by bast over 1 year ago - 1 comment
Labels: bug

#118 - Cosmetics

Pull Request - State: closed - Opened by bast over 1 year ago

#117 - For #94 - virtualize procfs, add tests

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#116 - Fix #111 - skip dead jobs, add defunct tag

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#115 - Fix #113 - avoid NaN values in procfs output

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago

#114 - Bump rustix from 0.37.20 to 0.37.25

Pull Request - State: closed - Opened by dependabot[bot] over 1 year ago - 1 comment
Labels: dependencies

#113 - procfs can create data with infinite values

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 3 comments
Labels: bug

#112 - Fix #107: Always print at least one record

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 2 comments

#111 - procfs must be more clever about jobs in the X and Z states

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 5 comments
Labels: bug

#110 - Fix #72: Surface GPU failures in the log records

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#109 - Make timeout 5s, and share it. Fix #92.

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago

#108 - Procfs tweak: ignore benign and expected errors

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#107 - Always write something

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 6 comments
Labels: enhancement, question

#106 - Data volume, disk and network contention

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 12 comments
Labels: question, performance

#105 - System configuration data

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 10 comments
Labels: question

#104 - more robust date query; closes #103

Pull Request - State: closed - Opened by bast over 1 year ago

#103 - Running `date` three times to compute the output file name is risky

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment
Labels: bug

#102 - Bump the version number for other things than file format changes

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 2 comments
Labels: enhancement

#101 - `sonar help` should print version and build date

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 6 comments
Labels: enhancement

#100 - Maintain monotonicity of timestamps even in the face of some failures

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment
Labels: enhancement

#99 - Fix #86

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#98 - Fix #97

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#97 - Sonar is slow without --batchless because of how we get slurm job IDs

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 6 comments
Labels: bug, performance

#96 - Improve error reporting with more detail

Issue - State: closed - Opened by lars-t-hansen over 1 year ago
Labels: enhancement

#95 - Fix #93

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago

#94 - Add `ps` test cases

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment
Labels: Tests

#93 - Sonar rolls up cpu_percentage for gpu_percentage

Issue - State: closed - Opened by lars-t-hansen over 1 year ago
Labels: bug

#92 - Intermittent failures on busy systems - 2s timeout may be too low

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 2 comments
Labels: enhancement

#91 - Add switch to exclude command names

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#90 - Print the pid field when it is meaningful

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 3 comments

#89 - Unlikely peak CPU usage for jobs with --batchless and subjobs

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 6 comments
Labels: bug

#88 - Use exec, not shell, to run subprograms

Issue - State: closed - Opened by lars-t-hansen over 1 year ago
Labels: enhancement, Later, performance

#87 - Read data directly from GPU APIs

Issue - State: open - Opened by lars-t-hansen over 1 year ago - 15 comments
Labels: enhancement, performance, important, Stability

#86 - Read data directly from /proc

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 5 comments
Labels: enhancement, performance

#85 - Exclude commands

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#84 - Add an option to filter by command name

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 2 comments
Labels: enhancement

#83 - Print the pid field when it is meaningful

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#82 - Improve README.md

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#81 - Handle unknown gpus. Fix #75

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 3 comments

#80 - Even better filtering. Fix #73

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 2 comments

#79 - Revamp ps. Fixes #78

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 9 comments

#78 - job_id is possibly a unique key for job information, but "it's complicated"

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 6 comments
Labels: enhancement

#77 - Fixes #76

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#76 - pid is a unique key for results from the subprocesses

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#75 - For `gpus` implement "unknown"

Issue - State: closed - Opened by bast over 1 year ago - 1 comment

#74 - Define, implement, and document defaults for values where it makes sense

Issue - State: closed - Opened by bast over 1 year ago - 2 comments

#73 - Job filtering should be post-accumulation, and must be more subtle

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 5 comments
Labels: bug

#72 - Log GPU status

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment
Labels: enhancement

#71 - Better error reporting. Fixes #52

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#70 - Add cputimes. Fixes #68

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 3 comments

#69 - Clean up documentation for GPU cards

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#68 - Print raw cpu time

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 11 comments
Labels: enhancement

#67 - Log communications volume

Issue - State: open - Opened by lars-t-hansen over 1 year ago
Labels: enhancement, important

#66 - make output fields self-identifying; closes #41

Pull Request - State: closed - Opened by bast over 1 year ago - 7 comments

#65 - update timestamp function; closes #34, closes #64

Pull Request - State: closed - Opened by bast over 1 year ago

#64 - Do not report sub-second precision

Issue - State: closed - Opened by bast over 1 year ago

#63 - update list of authors, add CITATION.cff; closes #57

Pull Request - State: closed - Opened by bast over 1 year ago

#62 - log_cmderror is only used in one file, moving definition there

Pull Request - State: closed - Opened by bast over 1 year ago

#61 - Print the error

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 5 comments

#60 - Print the error when a command fails

Issue - State: closed - Opened by lars-t-hansen over 1 year ago

#59 - Submit 51 53 58

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#58 - Synthesize jobs on batchless systems

Issue - State: closed - Opened by lars-t-hansen over 1 year ago

#57 - Update author information

Issue - State: closed - Opened by bast over 1 year ago - 1 comment

#56 - Batchless2

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#55 - Handle large inputs

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#54 - Propagate subprocess error

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 2 comments

#53 - Propagate errors from the command submodule

Issue - State: closed - Opened by lars-t-hansen over 1 year ago

#52 - Log fatal errors

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 6 comments
Labels: enhancement

#51 - Large output is a reality, and killing the subprocess is not a workaround

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 2 comments

#50 - Batchless job manager

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 2 comments

#49 - Amd gpu support

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 4 comments

#48 - new command 'Read' which fetches data from specified filed

Pull Request - State: closed - Opened by benteb over 1 year ago - 3 comments

#47 - Cleanup

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago

#46 - Refactor jobs

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 2 comments

#45 - Refactor ps

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#44 - Add a pre-check whether host/node is GPU-aware before running any other GPU commands

Issue - State: closed - Opened by bast over 1 year ago - 4 comments
Labels: enhancement

#43 - Ideas for "sonar analyze"

Issue - State: closed - Opened by bast over 1 year ago - 10 comments

#42 - Generalize the job ID

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 2 comments

#41 - Make output fields self-identifying?

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 11 comments

#40 - nvidia monitoring

Pull Request - State: closed - Opened by lars-t-hansen over 1 year ago - 6 comments

#39 - Rust version requirement

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 4 comments

#38 - Processes with "," in name break the CSV output

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#37 - Add GPU usage statistics

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 8 comments

#36 - Process names containing spaces are truncated

Issue - State: closed - Opened by lars-t-hansen over 1 year ago - 1 comment

#35 - Record also Slurm job ID

Issue - State: closed - Opened by bast almost 2 years ago

#34 - Record time with offset for local timezone

Issue - State: closed - Opened by bast almost 2 years ago - 1 comment

#33 - Still experiencing a timeout on a cluster

Issue - State: closed - Opened by bast about 2 years ago - 2 comments

#32 - Notes about the Rust version

Issue - State: closed - Opened by bast about 2 years ago

GitHub / NordicHPC/sonar issues and pull requests