Milestone information
- Project:
- LAVA Scheduler (deprecated)
- Series:
- linaro-11.11
- Version:
- 2011.10
- Released:
- Registrant:
- Paul Larson
- Release registered:
- Active:
- No. Drivers cannot target bugs and blueprints to this milestone.
Activities
- Assigned to you:
- No blueprints or bugs assigned to you.
- Assignees:
- 2 Michael Hudson-Doyle, 2 Paul Larson
- Blueprints:
- 1 Implemented
- Bugs:
- 3 Fix Released
Download files for this release
Release notes
* The jobs listed in the LAVA scheduler now include the name or the description of each job
Changelog
2011-10-20 Paul Larson <email address hidden>
release 0.5.3 final
2011-10-20 Paul Larson <email address hidden>
Make errors when submitting jobs easier to understand
2011-10-20 Michael Hudson-Doyle <email address hidden>
use except ... as ... e: syntax
2011-10-20 Michael Hudson-Doyle <email address hidden>
use simplejson & its more specific error classes
2011-10-19 Michael Hudson-Doyle <email address hidden>
raise a helpful error when neither 'target' nor 'device_type' found in job data.
2011-10-19 Michael Hudson-Doyle <email address hidden>
raise, rather than return, the fault
2011-10-19 Michael Hudson-Doyle <email address hidden>
this should be more useful
2011-10-19 Michael Hudson-Doyle <email address hidden>
try a bit harder to fix bug 861149 in all cases
2011-10-19 Michael Hudson-Doyle <email address hidden>
hack harder
2011-10-19 Michael Hudson-Doyle <email address hidden>
log when a conflict happens
2011-10-19 Michael Hudson-Doyle <email address hidden>
another approach
2011-10-19 Michael Hudson-Doyle <email address hidden>
merge trunk
2011-10-18 Michael Hudson-Doyle <email address hidden>
work around django bug that meant start_time was usually wrong on job page
2011-10-18 Michael Hudson-Doyle <email address hidden>
in getJobForBoard, use commit_on_success rather than manual transaction stuff
see https:/
2011-10-14 Michael Hudson-Doyle <email address hidden>
bump version to 0.5.3dev
2011-10-14 Michael Hudson-Doyle <email address hidden>
sprinkle a few select_relateds around to massively reduce query count (felt most on alljobs page for now)
2011-10-10 Paul Larson <email address hidden>
Add job descriptions in the scheduler UI
2011-10-10 Paul Larson <email address hidden>
add default empty string for job description
2011-10-05 Paul Larson <email address hidden>
Add job description to job.html, and make description default to null
2011-09-30 Paul Larson <email address hidden>
Add job description to the scheduler UI
2011-10-04 Paul Larson <email address hidden>
Make other references to the test device in the scheduler link to the details page for that device
2011-10-04 Paul Larson <email address hidden>
remove a couple of overly defensive defaults
2011-09-30 Paul Larson <email address hidden>
Oops, forgot to check for the case where actual_device might not exist.
Also improved it a bit:
if the actual_device exists, we link to that
else if the requested_device exists, we link to that, text is in italics
else if the requested_
2011-09-30 Paul Larson <email address hidden>
Add a link to the device details when we mention the device name
2011-09-23 Michael Hudson-Doyle <email address hidden>
bump version to 0.5.2
2011-08-30 Michael-Doyle Hudson <email address hidden>
lava-scheduler
2011-08-30 Michael-Doyle Hudson <email address hidden>
better default sorts and sizes for the scheduler tables
2011-08-25 Michael-Doyle Hudson <email address hidden>
bump version
2011-08-25 Michael-Doyle Hudson <email address hidden>
add a jsonp-compatible view of a job
2011-08-24 Michael-Doyle Hudson <email address hidden>
use get_object_or_404 to avoid a 500
2011-08-24 Michael-Doyle Hudson <email address hidden>
a touch of pep 8
2011-08-23 Michael-Doyle Hudson <email address hidden>
twiddle content type depending on presence of jsonp callback
2011-08-23 Michael-Doyle Hudson <email address hidden>
add JSONP supporting json representation of a job
2011-08-24 Michael-Doyle Hudson <email address hidden>
* change fake-dispatcher to do the fake test result outputty thing as the very
last thing -- this exposes a race between jobOobData completing and the
process exiting.
* fix the race by using the same lock the Job instance uses to serialized db
operations.
2011-08-22 Michael-Doyle Hudson <email address hidden>
bump version to a more confident sounding 0.5.0
2011-08-22 Michael-Doyle Hudson <email address hidden>
run lava-scheduler-
lava-scheduler
upstart when the lava-scheduler itself is restarted
2011-08-22 Michael-Doyle Hudson <email address hidden>
continue to update the output of a job while it is CANCELING
2011-08-22 Michael-Doyle Hudson <email address hidden>
display an empty cell rather than None before a job has a device in job listings
2011-08-22 Michael-Doyle Hudson <email address hidden>
bump version to 0.5.0a1 (and this really _is_ an alpha!)
2011-08-22 Michael-Doyle Hudson <email address hidden>
* add a simple device view
* add buttons to allow privileged people to move devices in and
out of maintenance mode
2011-08-22 Michael-Doyle Hudson <email address hidden>
remove some xxxs
2011-08-22 Michael-Doyle Hudson <email address hidden>
wrap long line in template
2011-08-22 Michael-Doyle Hudson <email address hidden>
remove unused css
2011-08-22 Michael-Doyle Hudson <email address hidden>
don't blindly mark the device idle in jobCompleted
2011-08-22 Michael-Doyle Hudson <email address hidden>
template tweaks
2011-08-22 Michael-Doyle Hudson <email address hidden>
button for going back online again
2011-08-22 Michael-Doyle Hudson <email address hidden>
add button to take a device offline
2011-08-22 Michael-Doyle Hudson <email address hidden>
merge trunk
2011-08-22 Michael-Doyle Hudson <email address hidden>
add a simple device view
2011-08-22 Michael-Doyle Hudson <email address hidden>
scaffolding for device view
2011-08-22 Michael-Doyle Hudson <email address hidden>
add going offline state
2011-08-22 Michael-Doyle Hudson <email address hidden>
forbid GETs to the job_cancel view
2011-08-19 Michael-Doyle Hudson <email address hidden>
Job cancellation
* the process running the dispatcher checks with the job source every 10
seconds to see if the dispatcher should be killed (i.e. if the status is
no longer RUNNING)
* add model apis to check if a user can cancel a job and to cancel a job
* add RPC to cancel a job
* add a button in the ui to cancel a job
2011-08-19 Michael-Doyle Hudson <email address hidden>
squelch warning
2011-08-19 Michael-Doyle Hudson <email address hidden>
brain fade
2011-08-19 Michael-Doyle Hudson <email address hidden>
add api for job cancellation
2011-08-19 Michael-Doyle Hudson <email address hidden>
return a 403 when appropriate
2011-08-19 Michael-Doyle Hudson <email address hidden>
add cancellation api and button
2011-08-19 Michael-Doyle Hudson <email address hidden>
call jobCheckForCanc
2011-08-19 Michael-Doyle Hudson <email address hidden>
add jobCheckForCanc
2011-08-19 Michael-Doyle Hudson <email address hidden>
add canceling state
2011-08-18 Michael-Doyle Hudson <email address hidden>
The scheduler now runs a subprocess which runs the dispatcher and reports
completion status to the db; this means that the scheduler can be killed and
restarted while jobs are still running.
2011-08-18 Michael-Doyle Hudson <email address hidden>
fix tests
2011-08-18 Michael-Doyle Hudson <email address hidden>
oops, wait for jobComplete to finish before exiting
2011-08-18 Michael-Doyle Hudson <email address hidden>
this appears to work(!)
2011-08-18 Michael-Doyle Hudson <email address hidden>
move calling jobCompleted into the Job class
2011-08-18 Michael-Doyle Hudson <email address hidden>
possibly working version of monitor
2011-08-18 Michael-Doyle Hudson <email address hidden>
oops
2011-08-18 Michael-Doyle Hudson <email address hidden>
split out getting the log file for a job into its own jobsource method
2011-08-18 Michael-Doyle Hudson <email address hidden>
add the skeleton of the monitor script
2011-08-18 Michael-Doyle Hudson <email address hidden>
reduce code duplication a bit
2011-08-18 Michael-Doyle Hudson <email address hidden>
If the connection to the DB fails or drops for the scheduler daemon, close the
connection so that we attempt to reopen it on the next access.
Also improve the errbacks that handle errors to log the traceback as well.
2011-08-18 Michael-Doyle Hudson <email address hidden>
merge trunk
2011-08-18 Michael-Doyle Hudson <email address hidden>
tidy slightly
2011-08-18 Michael-Doyle Hudson <email address hidden>
ah, the important detail
2011-08-17 Michael-Doyle Hudson <email address hidden>
try to force reconnection when database connection drops.
does not work, but hopefully that is just a detail
2011-08-17 Michael-Doyle Hudson <email address hidden>
make sure the deferred returned by _updateBoards always fires
2011-08-18 Michael-Doyle Hudson <email address hidden>
bump version to 0.4.4
2011-08-18 Michael-Doyle Hudson <email address hidden>
add setup.cfg
2011-08-18 Michael-Doyle Hudson <email address hidden>
record and display the url of the bundle submitted by a job
2011-08-17 Michael-Doyle Hudson <email address hidden>
result link is too long to include as text
2011-08-17 Michael-Doyle Hudson <email address hidden>
oops
2011-08-17 Michael-Doyle Hudson <email address hidden>
lava-scheduler now takes a --dispatcher option
2011-08-17 Michael-Doyle Hudson <email address hidden>
update template some more
2011-08-17 Michael-Doyle Hudson <email address hidden>
update template
2011-08-17 Michael-Doyle Hudson <email address hidden>
change jobOobData to match model & dispatcher changes
2011-08-17 Michael-Doyle Hudson <email address hidden>
rename bundle_sha1 to results_link (including editing the migration: if you deployed this branch, you just lost)
2011-08-16 Michael-Doyle Hudson <email address hidden>
pass --oob-fd 3 to lava-dispatch
2011-08-16 Michael-Doyle Hudson <email address hidden>
improve link
2011-08-05 Michael-Doyle Hudson <email address hidden>
link to the results from the job page if they are there
2011-08-05 Michael-Doyle Hudson <email address hidden>
call jobOobData on the jobsource when appropriate
2011-08-05 Michael-Doyle Hudson <email address hidden>
do something more structure with the output on fd 3
2011-08-05 Michael-Doyle Hudson <email address hidden>
pass the job source around as far as the DispatcherProce
2011-08-05 Michael-Doyle Hudson <email address hidden>
do something a little bit special with output on fd 3 from the dispatcher
2011-08-05 Michael-Doyle Hudson <email address hidden>
fix fake-dispatcher path computation
2011-08-05 Michael-Doyle Hudson <email address hidden>
add migration
2011-08-05 Michael-Doyle Hudson <email address hidden>
merge south pactching up from trunk
2011-08-04 Michael-Doyle Hudson <email address hidden>
add a method to store the bundle sha1 on a job
2011-08-05 Michael-Doyle Hudson <email address hidden>
(belatedly) add empty migration after merging of branches that both contained migrations
2011-08-04 Michael-Doyle Hudson <email address hidden>
fix display of device in job listings
2011-08-01 Michael-Doyle Hudson <email address hidden>
bump version to 0.4.3
2011-08-01 Michael-Doyle Hudson <email address hidden>
use http://
2011-08-01 Michael-Doyle Hudson <email address hidden>
pass the environment along (in the bizarre way you get Twisted to do this)
to the dispatcher
2011-07-28 Michael-Doyle Hudson <email address hidden>
lava-scheduler now reads a log file to configure logging
2011-07-27 Michael-Doyle Hudson <email address hidden>
add a comment
2011-07-27 Michael-Doyle Hudson <email address hidden>
allow the config file to set the logging destination
2011-07-27 Michael-Doyle Hudson <email address hidden>
oops
2011-07-27 Michael-Doyle Hudson <email address hidden>
change default log level to info
2011-07-27 Michael-Doyle Hudson <email address hidden>
tweak log levels a bit
2011-07-27 Michael-Doyle Hudson <email address hidden>
move the setting of DJANGO_
2011-07-27 Michael-Doyle Hudson <email address hidden>
support reading the log level from a config file
2011-07-27 Michael-Doyle Hudson <email address hidden>
make executable a stub, as it should be
2011-07-27 Zygmunt Krynicki <email address hidden>
Bump version to 0.4.2a1
2011-07-27 Zygmunt Krynicki <email address hidden>
Upgrade dependencies a little
2011-07-27 Michael-Doyle Hudson <email address hidden>
Sigh, another problem in device type targeting: the target wasn't being put
into the json data before being given to the dispatcher.
2011-07-27 Michael-Doyle Hudson <email address hidden>
fix some tests broken by semantic merge conflict
2011-07-27 Michael-Doyle Hudson <email address hidden>
use the right name for the dispatcher script
2011-07-27 Michael-Doyle Hudson <email address hidden>
fix two bugs in the new job view:
1) a newline was inserted on each poll even if there was no new data
2) the json definition could flow out of the tab
2011-07-27 Michael-Doyle Hudson <email address hidden>
use overflow:auto on the display of the json
2011-07-27 Michael-Doyle Hudson <email address hidden>
do not insert an empty <pre> block when there is no new log
output
2011-07-27 Michael-Doyle Hudson <email address hidden>
make fake dispatcher produce output less often than once a second
2011-07-27 Michael-Doyle Hudson <email address hidden>
bump version to 0.4.2a0
2011-07-26 Michael-Doyle Hudson <email address hidden>
bump version
2011-07-26 Michael-Doyle Hudson <email address hidden>
maintain numeric sort by id on alljobs
2011-07-26 Michael-Doyle Hudson <email address hidden>
some extremely lame option parsing in lava-scheduler
2011-07-26 Michael-Doyle Hudson <email address hidden>
bump version
2011-07-26 Michael-Doyle Hudson <email address hidden>
Add the ability to view the logs of jobs, even as they run.
2011-07-26 Michael-Doyle Hudson <email address hidden>
remove debugging
2011-07-26 Michael-Doyle Hudson <email address hidden>
do not scroll to the bottom of the log output on the first load of log data
2011-07-26 Michael-Doyle Hudson <email address hidden>
bump up the largest chunk we send by default considerably
2011-07-26 Michael-Doyle Hudson <email address hidden>
use {% url %} to compute url to load log output from
2011-07-26 Michael-Doyle Hudson <email address hidden>
merge trunk
2011-07-26 Michael-Doyle Hudson <email address hidden>
suppress global ajax progress bar for log updates
2011-07-26 Michael-Doyle Hudson <email address hidden>
flakes
2011-07-26 Michael-Doyle Hudson <email address hidden>
no longer need LAVA_LOGS settings
2011-07-26 Michael-Doyle Hudson <email address hidden>
rename migration that adds TestJob.log_file
2011-07-26 Michael-Doyle Hudson <email address hidden>
merge device-
2011-07-26 Michael-Doyle Hudson <email address hidden>
make sure that devices targeted to a board don't run on other boards\!
2011-07-26 Michael-Doyle Hudson <email address hidden>
another off by one
2011-07-26 Michael-Doyle Hudson <email address hidden>
use filefield interface in views
2011-07-26 Michael-Doyle Hudson <email address hidden>
oops, nasty bug probably preventing type-targeted jobs from being marked completed
2011-07-26 Michael-Doyle Hudson <email address hidden>
ridiculous gyrations to get getJobForBoard to return a file open for writing
2011-07-26 Michael-Doyle Hudson <email address hidden>
use FileField instead of manually specified locations for the log files
2011-07-26 Michael-Doyle Hudson <email address hidden>
put the log data in the response body to avoid thinking about unicode
(it turns out there was a reason jenkins did it this way!)
2011-07-25 Michael-Doyle Hudson <email address hidden>
allow loading of skipped content
2011-07-25 Michael-Doyle Hudson <email address hidden>
don't drip feed the log to the client, but rather skip the beginning if there is too much
2011-07-25 Michael-Doyle Hudson <email address hidden>
make the fake dispatcher run for much longer
2011-07-25 Michael-Doyle Hudson <email address hidden>
just because the deferred returned by Job.run() no longer fires with
an *interesting* value does not mean it does not fire with *a* value!
2011-07-25 Michael-Doyle Hudson <email address hidden>
fix tests
2011-07-25 Michael-Doyle Hudson <email address hidden>
when inserting new content, scroll the log output down if
it was already at the bottom
2011-07-25 Michael-Doyle Hudson <email address hidden>
consolidate log file path calculation
2011-07-25 Michael-Doyle Hudson <email address hidden>
sigh
2011-07-25 Michael-Doyle Hudson <email address hidden>
fix off by one error that was accidentally beneficial to my sample data
scan a little further to try to find a newline
2011-07-25 Michael-Doyle Hudson <email address hidden>
stop loading the log when you switch away from the tab
do not reload from scratch when you select the tab again
2011-07-25 Michael-Doyle Hudson <email address hidden>
woo incremental loading works
2011-07-25 Michael-Doyle Hudson <email address hidden>
*only* load the log data lazily...
2011-07-25 Michael-Doyle Hudson <email address hidden>
lazy loading of log data
2011-07-25 Michael-Doyle Hudson <email address hidden>
checkpoint
2011-07-25 Michael-Doyle Hudson <email address hidden>
slightly crummy viewing of log file
2011-07-25 Michael-Doyle Hudson <email address hidden>
oops
2011-07-25 Michael-Doyle Hudson <email address hidden>
put the log files somewhere predictable, and flush after every bit of output from the subprocess
2011-07-26 Michael-Doyle Hudson <email address hidden>
Fixes two small but serious problems with targeting jobs to device_types:
1) A bug in job completed meant that such a job would never be marked
completed.
2) A job targeted to a specific device could run on another device of the same
type.
2011-07-26 Michael-Doyle Hudson <email address hidden>
clarify test name
2011-07-26 Michael-Doyle Hudson <email address hidden>
bah typos rule me today
2011-07-26 Michael-Doyle Hudson <email address hidden>
template tweaks
2011-07-26 Michael-Doyle Hudson <email address hidden>
fix TestJob.__unicode__
2011-07-26 Michael-Doyle Hudson <email address hidden>
whoops
2011-07-26 Michael-Doyle Hudson <email address hidden>
adapt tests & dbjobsource
2011-07-26 Michael-Doyle Hudson <email address hidden>
rejig model to cleanly separate the board a job ran on and the board it
was requested to run on
2011-07-26 Michael-Doyle Hudson <email address hidden>
remove old fields
2011-07-26 Michael-Doyle Hudson <email address hidden>
add fields to make distinction more explicit between requested and actual devices for a job
2011-07-26 Michael-Doyle Hudson <email address hidden>
allow device_type to be non-null
2011-07-26 Michael-Doyle Hudson <email address hidden>
test illustrating problem
2011-07-26 Michael-Doyle Hudson <email address hidden>
fix bug where device_type targeted jobs would never be marked completed
2011-07-25 Michael-Doyle Hudson <email address hidden>
a simple job view
2011-07-25 Michael-Doyle Hudson <email address hidden>
add a comment indicating where this is going
2011-07-25 Michael-Doyle Hudson <email address hidden>
show the definition too
2011-07-25 Michael-Doyle Hudson <email address hidden>
fiddle with whitespace
2011-07-25 Michael-Doyle Hudson <email address hidden>
slightly naff css to make better use of the page space
2011-07-25 Michael-Doyle Hudson <email address hidden>
ditto for alljobs page
2011-07-25 Michael-Doyle Hudson <email address hidden>
still sort by id
2011-07-25 Michael-Doyle Hudson <email address hidden>
merge trunk
2011-07-25 Michael-Doyle Hudson <email address hidden>
include status in job view
2011-07-21 Michael-Doyle Hudson <email address hidden>
a bit of content to the job view
2011-07-21 Michael-Doyle Hudson <email address hidden>
scaffolding for a job view
2011-07-25 Michael-Doyle Hudson <email address hidden>
remove twisted/plugins directory that was left in by mistake
2011-07-21 Paul Larson <email address hidden>
Version 0.3
2011-07-21 Michael-Doyle Hudson <email address hidden>
Dispatch jobs that are targeted at a device type not just a device. (Jobs that
are targeted directly at a board will be processed before jobs that are
targeted at a device type).
Also, fix the attempt to process older jobs first to actually work.
2011-07-21 Michael-Doyle Hudson <email address hidden>
the test was wrong. phew
2011-07-21 Michael-Doyle Hudson <email address hidden>
well, this *should* work?
2011-07-21 Michael-Doyle Hudson <email address hidden>
django is not storm django is not storm
2011-07-21 Michael-Doyle Hudson <email address hidden>
tests wrt ordering. which don't work, for no reason i can figure out.
2011-07-21 Michael-Doyle Hudson <email address hidden>
merge trunk
2011-07-20 Michael-Doyle Hudson <email address hidden>
add (failing) test for selecting a job based on device type
2011-07-20 Michael-Doyle Hudson <email address hidden>
mention the django ticket i filed
2011-07-20 Michael-Doyle Hudson <email address hidden>
add a test, fix a slightly ridiculous problem, slightly whingy comment about said problem
2011-07-20 Michael-Doyle Hudson <email address hidden>
fix tests (sorry)
2011-07-20 Michael-Doyle Hudson <email address hidden>
a very simple ui for the scheduler
2011-07-20 Michael-Doyle Hudson <email address hidden>
prettify all jobs
2011-07-20 Michael-Doyle Hudson <email address hidden>
fix indent
2011-07-20 Michael-Doyle Hudson <email address hidden>
some template factoring
2011-07-20 Michael-Doyle Hudson <email address hidden>
prettify (mostly by cargo culting from the dashboard)
2011-07-20 Michael-Doyle Hudson <email address hidden>
css + fix mistake in urls.py
2011-07-20 Michael-Doyle Hudson <email address hidden>
start to dataTable-ize
2011-07-20 Michael-Doyle Hudson <email address hidden>
merge trunk
2011-07-12 Michael-Doyle Hudson <email address hidden>
add an alljobs page
2011-07-12 Michael-Doyle Hudson <email address hidden>
only show pending and running jobs on the front page
2011-07-12 Michael-Doyle Hudson <email address hidden>
improve TestJob.__unicode__
2011-07-08 Michael-Doyle Hudson <email address hidden>
merge trunk
2011-06-27 Michael-Doyle Hudson <email address hidden>
remove print
2011-06-24 Michael-Doyle Hudson <email address hidden>
be explicit about the id
2011-06-24 Michael-Doyle Hudson <email address hidden>
remove stub content
2011-06-24 Michael-Doyle Hudson <email address hidden>
merge trunk picking up some changes from daemon-v1 on the way
2011-06-22 Michael-Doyle Hudson <email address hidden>
equally naive ui for jobs
2011-06-22 Michael-Doyle Hudson <email address hidden>
naive ui for devices
2011-07-12 Zygmunt Krynicki <email address hidden>
Bump version to 0.3b1
2011-07-12 Zygmunt Krynicki <email address hidden>
Make /lava-server/
2011-07-08 Zygmunt Krynicki <email address hidden>
Update to match lava-server changes
2011-07-08 Zygmunt Krynicki <email address hidden>
PEP8 setup.py
2011-07-08 Zygmunt Krynicki <email address hidden>
Fix typo in setup.py
2011-07-08 Zygmunt Krynicki <email address hidden>
Bump version to 0.2, version numbers are cheap
2011-07-08 Zygmunt Krynicki <email address hidden>
Bump version to 0.1.1 final
2011-07-07 Michael-Doyle Hudson <email address hidden>
bah!
2011-07-07 Michael-Doyle Hudson <email address hidden>
bump version a bit
2011-07-07 Michael-Doyle Hudson <email address hidden>
add south to deps
2011-07-07 Michael-Doyle Hudson <email address hidden>
initial migration
2011-07-07 Michael-Doyle Hudson <email address hidden>
install the scheduler script
2011-07-07 Michael-Doyle Hudson <email address hidden>
work in a venv
2011-07-07 Michael-Doyle Hudson <email address hidden>
add bang (oops)
2011-07-07 Michael-Doyle Hudson <email address hidden>
give up on twistd
2011-07-07 Michael-Doyle Hudson <email address hidden>
another approach
2011-07-07 Michael-Doyle Hudson <email address hidden>
put twisted back in install_requires
2011-07-07 Michael-Doyle Hudson <email address hidden>
install the plugin
2011-07-07 Michael-Doyle Hudson <email address hidden>
convert to a twistd plugin
2011-07-07 Michael-Doyle Hudson <email address hidden>
remove twisted from install_requires for now
2011-07-07 Michael-Doyle Hudson <email address hidden>
fix the test
2011-07-07 Michael-Doyle Hudson <email address hidden>
use ubertestcase
2011-07-07 Michael-Doyle Hudson <email address hidden>
depend on twisted
2011-07-07 Michael-Doyle Hudson <email address hidden>
maybe work on 2.6 now
2011-06-29 Michael-Doyle Hudson <email address hidden>
a database-backed jobsource for the scheduler daemon
2011-06-24 Michael-Doyle Hudson <email address hidden>
merge trunk
2011-06-23 Michael-Doyle Hudson <email address hidden>
tests for jobComplete
2011-06-23 Michael-Doyle Hudson <email address hidden>
finish getJobForBoard tests for now
2011-06-23 Michael-Doyle Hudson <email address hidden>
another test
2011-06-23 Michael-Doyle Hudson <email address hidden>
some test helper refactoring
2011-06-23 Michael-Doyle Hudson <email address hidden>
a test for getJobForBoard
2011-06-23 Michael-Doyle Hudson <email address hidden>
start to test DatabaseJobSource
2011-06-23 Michael-Doyle Hudson <email address hidden>
dithering over naming
2011-06-23 Michael-Doyle Hudson <email address hidden>
avoid using defer_to_thread decorator, so i can write tests that do not use threads
2011-06-23 Michael-Doyle Hudson <email address hidden>
merge trunk (no text changes)
2011-06-23 Michael-Doyle Hudson <email address hidden>
record start and stop time
2011-06-22 Michael-Doyle Hudson <email address hidden>
add a comment
2011-06-22 Michael-Doyle Hudson <email address hidden>
finally an approach to concurrency safety i can like
2011-06-21 Michael-Doyle Hudson <email address hidden>
merge improvements from daemon-v0
2011-06-17 Michael-Doyle Hudson <email address hidden>
oops, properly mark the device as idle when a job finishes
2011-06-17 Michael-Doyle Hudson <email address hidden>
appears to work now!!
2011-06-17 Michael-Doyle Hudson <email address hidden>
sundry fixes, and the first of what will surely be many errbacks.
2011-06-17 Michael-Doyle Hudson <email address hidden>
(maybe) implement getJobForBoard
2011-06-17 Michael-Doyle Hudson <email address hidden>
add admin interface for scheduler models; one fix
2011-06-17 Michael-Doyle Hudson <email address hidden>
start a db-backed jobsource
2011-06-23 Paul Larson <email address hidden>
Add copying for AGPL
2011-06-23 Zygmunt Krynicki <email address hidden>
Use versiontools in setup.py
2011-06-23 Zygmunt Krynicki <email address hidden>
Add MANIFEST.in
2011-06-23 Michael-Doyle Hudson <email address hidden>
first cut at a scheduler daemon: this one just uses the filesystem as its datastore
2011-06-21 Michael-Doyle Hudson <email address hidden>
update some parameter names
2011-06-21 Michael-Doyle Hudson <email address hidden>
some more tests, and fix some bugs they found
2011-06-21 Michael-Doyle Hudson <email address hidden>
add more infrastructure; add test
2011-06-21 Michael-Doyle Hudson <email address hidden>
one more test, plus a double
2011-06-21 Michael-Doyle Hudson <email address hidden>
two more tests
2011-06-21 Michael-Doyle Hudson <email address hidden>
initial test, some more docs
2011-06-20 Michael-Doyle Hudson <email address hidden>
an over-detailed docstring
2011-06-20 Michael-Doyle Hudson <email address hidden>
remove the log file
2011-06-17 Michael-Doyle Hudson <email address hidden>
*cough*
2011-06-17 Michael-Doyle Hudson <email address hidden>
i missed adding the callback that requested another job when one finished on a board
2011-06-17 Michael-Doyle Hudson <email address hidden>
partition code some more; add more logging
2011-06-17 Michael-Doyle Hudson <email address hidden>
start to factor code into separate files
2011-06-16 Michael-Doyle Hudson <email address hidden>
this is not needed any more
2011-06-16 Michael-Doyle Hudson <email address hidden>
remove initial implementation
2011-06-16 Michael-Doyle Hudson <email address hidden>
start looking for jobs on newly found boards
2011-06-16 Michael-Doyle Hudson <email address hidden>
more logging, more working
2011-06-16 Michael-Doyle Hudson <email address hidden>
closer to working
2011-06-16 Michael-Doyle Hudson <email address hidden>
it is sad that i can almost type this sort of thing
in my sleep
2011-06-15 Michael-Doyle Hudson <email address hidden>
*cough*
2011-06-15 Michael-Doyle Hudson <email address hidden>
don't stomp over completed job files
2011-06-15 Michael-Doyle Hudson <email address hidden>
a slightly more interesting pretend dispatcher
2011-06-15 Michael-Doyle Hudson <email address hidden>
save output, do not dispatch more than one job at once
2011-06-15 Michael-Doyle Hudson <email address hidden>
this does stuff
2011-06-15 Michael-Doyle Hudson <email address hidden>
a start?
2011-06-14 Michael-Doyle Hudson <email address hidden>
add a submit_job api frontend to job creation
2011-06-14 Michael-Doyle Hudson <email address hidden>
check that login succeeded in TestTransport
2011-06-14 Michael-Doyle Hudson <email address hidden>
remove whiny comments
2011-06-13 Michael-Doyle Hudson <email address hidden>
more comments
2011-06-13 Michael-Doyle Hudson <email address hidden>
add a comment
2011-06-13 Michael-Doyle Hudson <email address hidden>
merge trunk, which includes a small change to create-TestJob
2011-06-13 Michael-Doyle Hudson <email address hidden>
all tests pass at last
2011-06-13 Michael-Doyle Hudson <email address hidden>
some sort of progress
2011-06-13 Michael-Doyle Hudson <email address hidden>
some formatting improvements
2011-06-13 Michael-Doyle Hudson <email address hidden>
"merge" create-TestJob, although it is nearly a null merge
2011-06-13 Michael-Doyle Hudson <email address hidden>
improvements
2011-06-10 Michael-Doyle Hudson <email address hidden>
very simple xmlrpc-level test
2011-06-13 Michael-Doyle Hudson <email address hidden>
Add a method for creating TestJobs, and enough of the model required to do
that.
2011-06-13 Michael-Doyle Hudson <email address hidden>
change to a TextField for TestJob.definition
2011-06-13 Michael-Doyle Hudson <email address hidden>
make the DeviceType.name primary
2011-06-13 Michael-Doyle Hudson <email address hidden>
comment out stuff that is not tested yet and update docstrings
2011-06-13 Michael-Doyle Hudson <email address hidden>
repair test
2011-06-10 Michael-Doyle Hudson <email address hidden>
comment out some stuff that is not tested yet
2011-06-10 Michael-Doyle Hudson <email address hidden>
one more test
2011-06-10 Michael-Doyle Hudson <email address hidden>
some more tests and code to pass them
2011-06-10 Michael-Doyle Hudson <email address hidden>
make test pass
2011-06-10 Michael-Doyle Hudson <email address hidden>
add some models and one test (which fails)
2011-06-10 Michael-Doyle Hudson <email address hidden>
expose the api trivially
2011-06-07 Michael-Doyle Hudson <email address hidden>
expose a trivial api call
2011-05-25 Michael-Doyle Hudson <email address hidden>
lava-server extenson skeleton
2011-05-25 Michael-Doyle Hudson <email address hidden>
django-admin startapp lava_scheduler_app
1 blueprint and 3 bugs targeted
Blueprint | Priority | Assignee | Delivery | |
---|---|---|---|---|
Include the job description in the scheduler UI | Include the job description in the scheduler UI | 2 Low | Paul Larson | 11 Implemented |
Bug report | Importance | Assignee | Status | |||
---|---|---|---|---|---|---|
861149 | #861149 | Submitted time is later than start time | 4 Medium | Michael Hudson-Doyle | 10 Fix Released | |
861988 | #861988 | json errors in submit job should be shown to the user | 4 Medium | Michael Hudson-Doyle | 10 Fix Released | |
863757 | #863757 | Device name should link to device details page | 5 Low | Paul Larson | 10 Fix Released |