This release requires Python >=3.6 and JupyterHub >=1.5.1.
- allow for req_keepvars_extra to be configured #295 (@mark-tomich, @minrk)
- Remove
which jupyterhub-singleuser
command fromSlurmSpawner.batch_script
#265 (@t20100, @consideRatio)
- TST: don't assume test user is OS user #301 (@minrk)
- Add python 3.12 for tests #299 (@Ph0tonic, @consideRatio)
- maint: req py36+ and jh 1.5.1+, fix tests, add RELEASE.md, add pre-commit hooks, add dependabot #273 (@consideRatio, @mbmilligan, @ryanlovett, @yuvipanda, @mahendrapaipuri)
- Upgrade singleuser.py to JupyterHub 4 #267 (@mahendrapaipuri, @minrk, @consideRatio)
- Remove reading/setting HubAuth SSL attributes in singeuser #259 (@cmd-ntrf, @consideRatio)
- Fix Slurm test used regular expression #256 (@t20100, @consideRatio)
- Quell async warning, and POST with body for jupyterhub 3.0 #247 (@ryanlovett, @mbmilligan, @rcthomas, @minrk, @jbeal-work, @mawigh, @cmd-ntrf, @jaescartin1)
- Improve submit_batch_script logging #219 (@cmd-ntrf, @consideRatio, @mbmilligan)
- Add temporary info about a temporary bug with JupyterHub 3+ #290 (@krokicki, @consideRatio)
- Modernize test matrix #252 (@mbmilligan)
The following people contributed discussions, new ideas, code and documentation contributions, and review. See our definition of contributors.
(GitHub contributors page for this release)
@basnijholt (activity) | @cmd-ntrf (activity) | @consideRatio (activity) | @jaescartin1 (activity) | @jbeal-work (activity) | @krokicki (activity) | @mahendrapaipuri (activity) | @mark-tomich (activity) | @mawigh (activity) | @mbmilligan (activity) | @minrk (activity) | @opoplawski (activity) | @Ph0tonic (activity) | @rcthomas (activity) | @ryanlovett (activity) | @t20100 (activity) | @yuvipanda (activity)
Changed
- PR #237: Replace use of scripts with entry_points
- PR #208 #238 #239 #240 #241: updates to CI - bumping versions and aligning with Jupyterhub standards
- PR #220: remove code supporting Jupyterhub earlier than 0.9
Fixed
- PR #229: LSF jobs with multiple slots display each hostname ':' separated
Added (user)
- PR #170: SlurmSpawner: add
req_gres
to specify-go-res
. - PR #137: GridEngineSpawner: spawner will now add the following system environment values to the spawner environment, in accordance with the Univa Admin Guide:
SGE_CELL
,SGE_EXECD
,SGE_ROOT
,SGE_CLUSTER_NAME
,SGE_QMASTER_PORT
,SGE_EXECD_PORT
,PATH
Added (developer)
- PR #187: support for unknown job state
Changed
- PR #177: Fail on first error in batch script by setting
set -e
to script templates. - PR #165: SlurmSpawner: Update template to use
--chdir
instead of--workdir
. Users of Slurm older than 17.11 may need to revert this locally. - PR #189: remove bashism from default script template
- PR #195: fix exception handling in run_command
- PR #198: change from Travis to gh-actions for testing
- PR #196: documentation
- PR #199: update setup.py
- PR #189: batchspawner/batchspawner: Don't use
-o pipefail
in /bin/sh scripts - PR #180: travis: Attempt to fix CI
- PR #177: Fail hard on first error in batch script
- PR #170: add 'gres' option to SlurmSpawner
- PR #165: Update batchspawner.py to use --chdir instead of --workdir
- PR #137: Grab environment variables needed for grid engine
This release requires minimum JupyterHub 0.9 and Python 3.5.
Added (user)
- Add support for JupyterHub named servers. #167
- Add Jinja2 templating as an option for all scripts and commands. If '{{' or
{%
is used anywhere in the string, it is used as a jinja2 template. - Add new option exec_prefix, which defaults to
sudo -E -u {username}
. This replaces explicitsudo
in every batch command - changes in local commands may be needed. - New option:
req_keepvars_extra
, which allows keeping extra variables in addition to what is defined by JupyterHub itself (addition of variables to keep instead of replacement). #99 - Add
req_prologue
andreq_epilogue
options to scripts which are inserted before/after the main jupyterhub-singleuser command, which allow for generic setup/cleanup without overriding the entire script. #96 - SlurmSpawner: add the
req_reservation
option. #91 - Add basic support for JupyterHub progress updates, but this is not used much yet. #86
Added (developer)
- Add many more tests.
- Add a new page
SPAWNERS.md
which information on specific spawners. Begin trying to collect a list of spawner-specific contacts. #97 - Rename
current_ip
andcurrent_port
commands toip
andport
. No user impact. #139 - Update to Python 3.5
async
/await
syntax to support JupyterHub progress updates. #90
Changed
- PR #58 and #141 changes logic of port selection, so that it is selected after the singleuser server starts. This means that the port number has to be conveyed back to JupyterHub. This requires the following changes:
jupyterhub_config.py
must explicitely importbatchspawner
- Add a new option
batchspawner_singleuser_cmd
which is used as a wrapper in the single-user servers, which conveys the remote port back to JupyterHub. This is now an integral part of the spawn process. - If you have installed with
pip install -e
, you will have to re-install so that the new scriptbatchspawner-singleuser
is added to$PATH
.
- Update minimum requirements to JupyterHub 0.9 and Python 3.5. #143
- Update Slurm batch script. Now, the single-user notebook is run in a job step, with a wrapper of
srun
. This may need to be removed usingreq_srun=''
if you don't want environment variables limited. - Pass the environment dictionary to the queue and cancel commands as well. This is mostly user environment, but may be useful to these commands as well in some cases. #108, #111 If these environment variables were used for authentication as an admin, be aware that there are pre-existing security issues because they may be passed to the user via the batch submit command, see #82.
Fixed
- Improve debugging on failed submission by raising errors including error messages from the commands. #106
- Many other non-user or developer visible changes. #107 #106 #100
- In Travis CI, blacklist jsonschema=3.0.0a1 because it breaks tests
Removed
- Fix regression: single-user server binding address is overwritten by previous session server address, resulting in failure to start. Issue #76
This release is compatible with JupyterHub 0.5.0 through 0.8.1/0.9dev.
- SlurmSpawner: Remove
--uid
for (at least) Slurm 17.11 compatibility. If you usesudo
, this should not be necessary, but because this is security related you should check that user management is as you expect. If your configuration does not usesudo
then you may need to add the--uid
option in a custombatch_script
. - add base options
req_ngpus
req_partition
req_account
andreq_options
- Fix up logging
- Merge
user_options
with the template substitution vars instead of having it as a separate key - Update ip/port handling for JupyterHub 0.8
- Add
LICENSE
(BSD3) andCONTRIBUTING.md
- Add
LsfSpawner
for IBM LFS - Add
MultiSlurmSpawner
- Add
MoabSpawner
- Add
condorSpawner
- Add
GridEngineSpawner
- SlurmSpawner: add
req_qos
option - WrapSpawner and ProfilesSpawner, which provide mechanisms for runtime configuration of spawners, have been split out and moved to the
wrapspawner
package - Enable CI testing via Travis-CI
- initial release containing
TorqueSpawner
andSlurmSpawner