This repository has been archived by the owner. It is now read-only.
Releases: hubblestack/pulsar
Releases · hubblestack/pulsar
Version 2016.10.3
General
- Switch to threading instead of multiprocessing for returner calls. (If you're worried about the GIL, don't. These operations are I/O bound, not CPU bound, and so don't run into the limitations of the GIL)
Linux
- Add more excludes to /opt for cloudpassage data
Version 2016.10.2
- Add
__version__
attribute so that nebula can query the version of this module - Add new exclude for
/opt/imply
- Add troubleshooting information for max inotify watches
Version 2016.10.1
Windows
- Rename
win_notify
towin_pulsar
. (However,__virtualname__
resolves just topulsar
, same as Linux.) - Add support for returners in
win_pulsar
, including batch support
Linux
- Spin up returners in separate processes, to make sure they don't hang the salt-minion (NOTE: This is not working on win_pulsar yet)
- Rename
refresh_frequency
torefresh_interval
and set to 300 seconds by default - Only refresh file watches every
refresh_interval
to keep CPU use low. Note thatauto_add: True
should be used, or else events may be missed. - Add more verbose logging
- Remove
/var
, add/opt
with some excludes
Version 2016.9.4
- Reorganize paths to allow for spm builds
Version 2016.9.3
- Pull fileserver operations from the beacon to a scheduled job to work around an edge case bug in saltstack where fileserver operations in beacons could cause fileserver operations in scheduled jobs to hang.
Version 2016.9.2
Please upgrade to v2016.9.3, this version has a critical bug
- Add
disable_during_state_run: True
to the sample pillar data - Fix a couple of issues related to incompatible salt versions
- Fix the default returner in the provided hubblestack_pulsar_config.yaml
- Add more excludes to hubblestack_pulsar_config.yaml
Version 2016.9.1
Please upgrade to v2016.9.3, this version has a critical bug
- Fix stacktraces caused on some old versions of pyinotify related to trace logging
- Cache the config, only refreshing from the master every 60 intervals (seconds) by default. This interval is configurable
- Add verbose config option to control beacon logging
Version 2016.9.0
Please upgrade to v2016.9.3, this version has a critical bug
Windows:
- Add windows inotify beacon (doesn't support returners yet)
Linux
- Set
batch: True
in the default watch data - Add many exclusions to the default watch data
- Pull the watch data out of pillar and into the fileserver (backwards-incompatible change! You'll need to update your pillar, see the new pillar.example)
- Add regex support for excludes
Version 2016.7.0
Initial stable Pulsar release