Releases: jhuckaby/Cronicle
Version 0.2.2
This version has a new flat UI design, for a more modern look. Now using Lato from Google Fonts.
Version 0.2.1
Minor release, just a bug fix:
- Fixed bug where "Run Again" button could sometimes freeze up UI if job completed immediately.
Version 0.2.0
This version includes Custom Data Tables and Custom HTML Content on the Job Details page, generated from JSON data emitted by your custom Plugin code.
Version 0.1.9
A few small misc fixes in this release:
- No longer auto-displaying logs over 10 MB on Job Details page, unless user asks for it (threshold is configurable).
- Fixed bug when sending e-mails about jobs which failed to launch (now tracking which events launched per minute tick).
- Unit test now runs on HTTP port 4012, and UDP port 4014, as to not interfere with other servers in a cluster which may be using the default ports (3012 / 3014).
Version 0.1.8
Cronicle will now use streams when managing its log files, as to reduce memory usage. Log files are no longer loaded into memory under any circumstances. However, please note that Node.js uses rather large internal stream buffers, up to 200 MB RAM or so. This is "normal".
Some misc. other bug fixes and tweaks:
- Bug fix: Upcoming event countdown time was based on client's local clock, not server clock.
- Bug fix: Prevent flickering on pagination click on certain pages.
- Text change in dialog: "Waiting for master server..."
- Cosmetic fix for table pagination white-space wrap.
- Cosmetic hack for "day" to "days" in server uptime display.
- New versions of pixl-request, pixl-server-web, and pixl-server-storage.
Version 0.1.7
Minor release, just fixed a race condition dealing with appending to log files.
On some OSes (CentOS 6.x for example), writing to log files with fs.appendFile()
and no callback results in out-of-order lines. Switched to fs.appendFileSync()
.
Version 0.1.6
Changes in this release:
- Fixed memory leaks with Node v0.12 (with new pixl-server 1.0.3)
- Fixed race condition where shell/detached processes may close STDIN before job can be written to it, resulting in an EPIPE error.
- Fixed rare situation in UI Home tab where jobs may be rendered before supporting data is available.
- Fixed issue where state data was getting polluted with user IP address / header data.
- Now including
process.argv
in startup debug message. - Now including
process.memoryUsage()
in GC debug messages. - More misc. memory leaks plugged (using
async.ensureAsync()
instead ofprocess.nextTick()
). - Removed
--color
flag in CLI debug mode by default (can be added by user on CLI)
Version 0.1.5
Changes in this release:
- Added new feature to "Run Now" which allows the user to customize the internal date/time for the job in isolation.
- Hold Shift which clicking Run Now to bring up the dialog.
- Fixed display bug with long error messages on Job Details page.
- Fixed bug in
choose_date_time()
which would not floor the seconds to 0. - Fixed bug where Chain Reaction mode could not be disabled after it was enabled on an event.
- Misc fixes in README.md.
- Fixed Markdown error in README.md.
- Better
.gitignore
file in README.md. - Removed some accidental commits (symlinks, 3rd party libs, part of build process).
Version 0.1.3
Changes in this version:
- New "Chain Reaction" feature for chaining events together.
- Plugins can use
chain_data
to pass arbitrary data between chained events. - Emails will now be sent to the
notify_fail
recipients if a job fails to launch due to server availability, etc. - New date/time picker dialog for the Time Machine (Event Clock).
- Time Machine / Event Clock now uses the event's timezone.
- Header and Floating clocks now include the timezone abbreviation.
- Shell Plugin now sends the job JSON blob into the shell process (in case someone decides to use it).
- Now sorting events, categories, plugins and server groups properly using
localeCompare()
. - Misc other fixes.
Version 0.1.2
Misc. fixes in this release:
- Now monitoring server resources every 10 seconds if local active jobs, 60 seconds if not.
- Added Algorithm IDs (for use in API).
- Renamed a few algo IDs (prefer_first and prefer_last).
- Now logging res monitor stuff at debug level 10 to prevent spamming log.
- Changed default job_startup_grace to 5 seconds.
- cp.exec for ps -ef now has a 5 sec timeout.