flux-framework/flux-core

Issues

• kvs-watch: retrieve content blobs in chunks rather than all at once

  #6456 opened a month ago by chu11
  1

• shell: error from input plugin when upstream broker dies could be improved

  #6490 opened 17 days ago by grondo
  0

• shell: turn fatal EPIPE error into warning

  #6487 opened 17 days ago by grondo
  2

• flux-job: job shell Broken pipe

  #6488 opened 18 days ago by garlick
  2

• macos: make check fails

  #6485 opened 18 days ago by garlick
  0

• integrate non-node resources into flux resource tools

  #6477 opened 19 days ago by garlick
  1

• ci: add a macos builder

  #6484 opened 18 days ago by garlick
  0

• resource.scheduling key should only be used on rank 0 if possible

  #6480 opened 18 days ago by grondo
  0

• idea: a base class for python-based jobtap plugins based on python-cffi

  #6483 opened 18 days ago by wihobbs
  1

• Flux Emulator Revival Questions

  #6466 opened 21 days ago by washwor1
  4

• kvs-watch: avoid re-fetching the same data repeatedly

  #6414 opened 18 days ago by garlick
  21

• flux hostlist: request idset with -x or --exclude

  #6464 opened 18 days ago by vsoch
  6

• idea: add `FLUX` environment variable that holds "closest enclosing jobid"

  #6474 opened 19 days ago by grondo
  0

• flux-resource: ERROR: error preparing response: No such file or directory

  #6470 opened 20 days ago by garlick
  0

• system instance "dry run" option

  #6469 opened 20 days ago by chu11
  1

• flux prolog not killed on cancel

  #6465 opened 20 days ago by vsoch
  5

• kvs commit: no such file or directory

  #6463 opened 24 days ago by vsoch
  28

• Core dump with flux archive create

  #6461 opened 24 days ago by vsoch
  17

• content: support RFC10 defined blob size limit

  #6457 opened a month ago by chu11
  0

• idea: kvs: support a native "EOF"

  #6455 opened a month ago by chu11
  4

• flux perilog-run is deprecated

  #6428 opened a month ago by garlick
  0

• sched-simple: add support for preemptible-after attribute

  #6446 opened a month ago by garlick
  0

• unsatisfiable request error

  #6437 opened a month ago by wadudmiah
  9

• Flux job uri not immediately available when job transitions to RUN

  #6441 opened a month ago by tgmachina
  3

• flux module stats job-ingest doesn't show subprocesses once they are in the trash

  #6432 opened a month ago by garlick
  0

• need a way to adminstratively hold a job

  #6439 opened a month ago by grondo
  0

• libflux: functions for creating out of tree watchers are brittle with respect to ABI

  #6433 opened a month ago by garlick
  0

• job validator hang at exit

  #6434 opened a month ago by grondo
  1

• systemd managed prolog can be slow to be canceled

  #6420 opened a month ago by grondo
  1

• flux jobs: show urgency=0 (HOLD) jobs as "held" in the INFO colum

  #6426 opened a month ago by garlick
  9

• zlistx comparators should return -1, 0, or 1

  #6398 opened 2 months ago by chu11
  2

• Request: job stats info based on timeframe

  #6387 opened 2 months ago by kkier
  7

• doc: there is no documentation for running jobs under systemd

  #6407 opened a month ago by garlick
  0

• TOML config updates can overwrite tables

  #6396 opened a month ago by grondo
  0

• consider using local time for eventlog formatter ISO 8601 format

  #6421 opened a month ago by grondo
  0

• throughput of jobs slows down over time - part deux

  #6419 opened 2 months ago by chu11
  5

• need a way to force a Flux subinstance to rediscover resources

  #6418 opened 2 months ago by grondo
  1

• job-exec: kill multi-user jobs with SIGUSR1

  #6406 opened 2 months ago by garlick
  1

• systemctl stop flux on rank 0 is fragile

  #6388 opened 2 months ago by garlick
  2

• job-manager: don't use flux imp kill to signal housekeeping, perilog

  #6409 opened 2 months ago by garlick
  0

• libsubprocess: duplicated stdio file descriptors are inherited by subprocesses

  #6415 opened 2 months ago by garlick
  0

• flux-shell: support stdin flow control to subprocesses

  #6413 opened 2 months ago by chu11
  0

• In the `queues` config table, (some?) policy settings do not appear to be overriding system defaults

  #6400 opened 2 months ago by kkier
  1

• broker/test: test_overlay.t: error binding to ipc://@test0: Address already in use

  #6404 opened 2 months ago by chu11
  9

• sdexec: systemd tries to kill guest tasks without sufficient permission

  #6399 opened 2 months ago by garlick
  2

• when the host:rank mapping in `bootstrap.hosts` differs across nodes, havoc ensues

  #6389 opened 2 months ago by garlick
  0

• flux-exec: do not operate on failed subprocesses

  #6395 opened 2 months ago by chu11
  1

• flux conda packages not built with systemd support

  #6394 opened 2 months ago by vsoch
  2

• Suggestion: reject configurations that don't match rank 0

  #6391 opened 2 months ago by kkier
  5

• Suggestion: improve error messaging with more context

  #6392 opened 2 months ago by kkier
  0