-
Notifications
You must be signed in to change notification settings - Fork 0
Release Notes
ltrestka edited this page Mar 14, 2024
·
1 revision
include_toc: true show_comments: false toc: true
- Within a campaign, any stage can be configured to wait for approval before being submitted.
- Documentation has been greatly enhanced to provide greater details. For example stages have been enlarged with a more complete description.
- Each webpage is now linked to a wiki help page. The link is shown as a question mark at the top of the page.
- Bug #22243: kill submission, etc. get worker jobs, but not dagman, also should set submission to "Removed"
- Bug #22234: Transition to Located on single file projects
- Bug #22162: cloning sample files needs to fix experiment= and actually fix the name
- Bug #22073: Job submitted, failed to run successfully but POMS so no data file generated, but POMS marked them located
- Bug #22064: editing a login/setup entry for analysis campaign requires an experiment hostname
- Bug #22047: upload_file page without choosing file
- Bug #22012: Got errors while killing a running job
- Feature #22110: create a pause/hold feature while submitting stages in the same campaign
- Feature #22090: poms_client API for Campaign Stage Submissions
- Feature #22026: show_campaigns and list_campaigns
- Feature #22021: Special use of UPLOADS env var for Analysis users
- Feature #22015: Override-able parameters in the gui editor: Database Change
- Feature #21955: poms client submit a job by campaign
- Feature #21953: poms_client add function upload ini should return the campaign and all stage ids/names
- Feature #21952: poms_client make_poms_call
- Feature #22177: poms_client APIs get job_type_snapshot_obj and login_setup_snap_obj by their IDs
- Feature #22026: show_campaigns and list_campaigns
- Feature #21993: create a campaign from ini
- Feature #22176: POMS Client API for campaign names and stage names giving campaignid and.or stage_id
- Feature #22213: extend poms_client campaign_stage_submissions function to take as argumrents 'tdays', 'tmin' and 'tmax'
- Feature #22035: poms_client: remove functions needed
- Task #21967: poms_client: clean up the user interface to reflect the change of campaign and campaign stage definition
- Idea #22046: How to handle dataset for intermediate stages
- Task #21555: Add link to Campaign Documentation on POMS home page
- Task #21477: Add one info page where multiple ? are present
- Task #21046: Document Handling custom placeholders in param overrides
- Task #19708: Documentation: Recovery Campaigns
- Task #19707: Documentation: Launch Script
- Added support for Analysis users. We are testing this in production before distributing across experiments.
- Bug #20619: Need a check on campaign stage edits
- Bug #20865: GUI Campaign editor "Double click to open" box remains too long.
- Bug #20904: Tags form fails creating new tags.
- Bug #20978: Need time range limit in poms_submission_agent.py...
- Bug #21045: Campaign stage form - two links point to same plot.
- Bug #21341: INC000000993451 - Users crashing when changing campaigns - but logged into different experiment.
- Bug #21380: campaign_task_files - the previous/next day links are not working
- Bug #21401: Need to add knowledge of experiment in campaign_time_bars .
- Bug #21402: When using form from Compose campaign stage, no error reported if stage has no login setup or job type.
- Bug #21413: datasets generated for dependent job launches should be more restrictive.
- Bug #21605: Kill / hold / release jobs doesn't hold right on DAGs (RITM0761732)
- Bug #21707: Campaign editor: cs_split_type does not save changes
- Bug #21775: Deleting tagged campaign fails
- Feature #19128: Revise user and experiment forms to follow user entry procedures.
- Feature #20213: Multiparameter workflow
- Feature #20217: Tool to interactively setup the same environement as a job launch
- Feature #20234: Report failure of a POMS campaign submission in the offline dashboard of experiments.
- Feature #20244: Maintain campaign_stage_n_active dataset
- Feature #20590: Workflow editor, internal structure
- Feature #20763: Make status (i.e. for SubmissionHistory) a vocab table with ids ordered(New, LaunchFailed, Idle, Running, Held, Failed, Completed, Located)
- Feature #21103: Submissions: Code update
- Feature #20771: Add button for custom recovery launches somewhere...
- Feature #20852: add time frame for stage plots
- Feature #20881: set X509_USER_PROXY auomatically for production jobs
- Feature #20944: Make a tag selection searchable.
- Feature #20993: Add a launch campaign button from Campaign page
- Feature #21109: Expose POMS "hidden" data to users
- Feature #21207: Need recovery launch info in .ini file dump, workflow editor, etc.
- Feature #21208: Need recovery launch info in .ini file dump
- Feature #21209: need recovery launch info in campaign editor on job type
- Feature #21210: Add recovery launch bits to ini file uploader.
- Feature #21282: Campaign Editor: Cloning Stages
- Feature #21411: Delete a job in New status
- Feature #21420: Update poms_client to pass POMS4_XXX parameters to jobsub
- Feature #21445: Request RITM0739824: login_setup and job_type are not highlighted in aqua for the campaign stage.
- Feature #21502: Analysis user support
- Feature #21503: File upload facility for analysis users
- Feature #21504: Sandbox directory creator
- Feature #21505: Modify launch code for analysis sandboxes
- Feature #21506: Amend login/setup for Analysis jobs
- Feature #21593: restrict login/setup launch_host and launch_user for analysis users
- Feature #21507: Add a file / proxy upload script to poms_client
- Feature #21508: Modify launch code to hold launches for analysis users without a proxy
- Feature #21613: Replace list of launch logs with submission list in campaign info page
- Feature #21701: About Campaign Editor.
- Feature #21702: About Campaign Editor.
- Task #19706: Documentation: Launch Template
- Task #20598: Split type plugins
- Task #20995: Add check on campaign stage edits
- Task #20996: Add check on campaign stage edits
- Task #21102: Submissions: Database changes
- Task #20855: Search Box - Campaigns and stages forms
- Task #20887: Obtain coordinators from FERRY.
- Task #20902: Move active flag from stages to campaigns
- Task #20941: Store nodes positions in campaign editor.
- Task #21187: Add campaign name and stage name to form produced by list_launch_file.
- Task #21372: Testing campaigns show up in uboone production reports
- Task #21476: type fields in campaigns, campaign_stage
- Task #21436: Documentation: Experiment version number
- Task #21486: Revise to the submission time bars page
- Task #21509: Convert Coordinator to Superuser
- Task #21510: Convert the submission time bars to tables
- Task #21523: Make Campaign field readonly when editing an existing stage.
- Task #21526: Rename function to compose job type.
- Task #21697: About Uploaded Files option.
- Task #21698: About Uploaded Files page.
- Task #21699: About Launch button.
- Added new split type for draining datasets.
- Added button to force submissions to move to located.
- Bug #21336 wrapup_tasks using wrong file pattern,,,
- Bug #21296 Provide user understandable data for a Campaign Stage Launch 404 Error.
- Bug #21167 Corrected dependent/recovery jobs to store and show correct user.
- Bug #21166 Fixed recovery bugs caused by invalid dimension query.
- Feature #21318 Button to force submissions to move to Located
- Feature #20243 Added byexistingruns split type.
- Backward compatibility documentation provided "here":https://cdcvs.fnal.gov/redmine/projects/prod_mgmt_db/wiki/Release_v400_Backward_Compatibility_-_What_You_need_to_know
- Campaign Editor has had minor improvements for easier usage.
- Campaign page active/inactive setting now works.
- Tagging campaigns is working again.
- Bug #20916 Spaces in stage/job type/login setup names
- Bug #20917 Multiline launch scripts in login/setup forms
- Bug #20935 Fix code that handles Tags
- Bug #20936 Campaign Editor: prevent linking the campaign defaults to the stage
- Feature #20938 Campaign Editor improvements
- Feature #20939 Documentation: talk about 'backward compatibility' for new release
- Bug #19699 Unregistered user logging in causes stack dump.
- Bug #19745 Aliases fail when shibboleth session has not be established.
- Documentation Revised - Written for new users
- Fully integrated with Fifemon
- Experiment's Analysis users can login and view production
- .ini files can be exported, edited and uploaded
- NEW GUI Campaign Editor - View/edit entire campaign.
- Feature #20654 Previous Users should be able to view anything in their experiment
- Feature #20673 Upgrade POMS to python 3.6
- Task #19187 Tutorials for experiment's production teams.
- Task #19276 Give Landscape the info of exp for the specific CAMPAIGN_ID, Campaign NAME, user running campaign
- Task #19694 Documentation: Overview Documentation
- Task #19696 Documentation: Production Processing High Level View
- Task #19701 Documentation: Requirements for Creating a Campaign
- Task #19698 Documentation: Adding a New Experiment or User
- Task #19702 Documentation: Creating a Campaign - Details
- Task #19710 Documentation: Campaign Actions
- Task #19770 Documentation: Cloning
- Task #19758 Create Campaign Stage Dependency Ordering
- Task #19263 Webpage in POMS to build the workflow
- Task #19761 Test Workflow Editor
- Task #19837 Fix up prototype workflow editor
- Task #19931 Automatically add users to poms_announce
- Task #20037 Table/field ids renaming
- Task #20041 Conversion of POMS to Fifemon
- Task #20042 Conversion: Database Changes
- Task #20043 Conversion: Eliminate scraping applications
- Task #20045 Conversion: Convert reports/plots to use elasticsearch
- Task #20101 Rename the tables and fields and update all source code to match functional names.
- Task #20225 Update the Show Campaigns (tags) form to use a table.
- Task #20226 Revise the main page.
- Bug #19445 SAM project link not working in POMS - Opened SNOW ticket INC000000935297
- Bug #19624 Whitespace confuses split_types code
- Bug #19922 bulk update is locking to heavily...
- Bug #19852 Stack dumps when deleting a job type
- Bug #19704 Jump to Campaign from the Campaign Stage Info only works for Active campings.
- Bug #19647 The link to the SNOW downtime page appears different from the other two under external links in POMS menu (blue characters and not centered).
- Bug #19624 Whitespace confuses split_types code
- Feature #19504 Dependencies should accept SAM metadata query bits, not just file pattern.
- Feature #19434 Use the Exp Software Version to tell POMS which version of the experiment's code would be loaded in the job.
- Feature #19418 Add link to the SNOW outage calendar under external links menu
- Feature #19324 add completed jobs index
- Feature #19199 Add "Test mode" to the Campaign stage.
- Feature #17852 Get downtimes from SNOW downtime calendar
- Task #19256 Fix .ssh/config for the round-robin alias
- Bug #18627: submission time bars prev/next doesn't work for tag= instead of campaign_id=.
- Bug #18833: Restore the fuzzy search for users.
- Bug #18918: "Tag/Untag Campaigns" action not working (as expected?)..
- Bug #18945: addusers.py does not update POMS when VOMS data for existing experimenter changes.
- Bug #19029: addusers.py seems to switch to inactive users with production role.
- Bug #19055: People with role production in VOMS have analysis role in POMS.
- Bug #19139: Failed jobs classified as located in dbrailsf_gen_poms_test campaign stage.
- Bug #19142: Kill submission campaign not working.
- Bug #19192: Test button not working on just-added launch templates.
- Bug #19338: declared_files_agent is starving...Bug #19368: release job is killing instead..
- Feature #16747: Show dependency, if any, in the campaign info page.
- Feature #18074: Last-modified info on campaigns, etc.
- Feature #18123: Need a way to reset, step back, split dataset iterators...
- Feature #18830: Role pull down menu for User Authorization and add security check.
- Feature #19297: Add "Config File Templates" in the POMS navigation menu.
- Feature #19024: addusers.py development.
- Feature #19120: DNS round-robin alias dune-computing.fnal.gov as a job launch node in POMS.
- Task #18871: Integration database.
- Task #19298: Add configuration file templates to the "Config File Templates" sub-menu.
- Task #18911: Provide a launch config template to interface POMS to the experiments' configurations, job submission and executable.
- Task #18988: Download/upload whole campaign/workflow to/from file.
- Task #19088: Add coordinators to each of the experiments according to the partial list in the description (Will be update as the info come).
- Task #19198: Change "Submit" button in all Configure Work with "Save"
- Bug #18176: Don't fail if session_experiment is not set.
- Bug #18194: Hold reason not being propagated to database.
- Bug #18300: Deal with files with unknown status in project summary/datasets for "delivered" file stats.
- Bug #18361: Button to reset split dataset sequence in a campaign.
- Bug #18388: Held jobs that are removed should be marked as Removed, not Completed.
- Feature #12195: Test button for launch template, campaign definition.
- Feature #17741: Add "Tag/UnTag Campaign".
- Feature #17779: Add "campaign_type" field to "campaigns" table &model.
- Feature #17894: More datasets splitting types.
- Feature #17895: Make links on campaign_info page links to last active week of campaign, not current week.
- Feature #17931: Add "HOLD/RELEASE Jobs button".
- Feature #17934: Add "HOLD/RELEASE Launches button(s)".
- Feature #18365: Add a byrun split type.
- Feature #18731: Track who launches/submits jobs including the cron launcher.
- Feature #17859: Define roles into POMS.
- Task #18772: Change "root" from role to flag.
- Bug #17638: Deadlock errors in bulk_update, with Nova's 30k jobs queued.
- Bug #17641: Memory overrun from jobsub_q_scraper -* uwsgi on pomsgpvm02.
- Bug #17684: Crontab code is generating wrong path to the launcher again.
- Bug #17685: Bulk_update jobs doesn't collapse task_ids in task_updates.
- Feature #17599: Additional histograms of job information -- run-time, transfer-time, etc.
- Feature #17349: job reporters should split updates into multiple queues by task_id (odd/even, modulo 3, etc).
- Feature #17324: Have in the main campaign list webpage only basic information.
- Feature #17323: New campaign should invalidate the cache.
- Feature #17188: Select bunch of campaings and mark them inactive capability.
- Feature #17187: Filter to order campaign list.
- Feature #17026: Color code numbers in job tables.
- Feature #17025: Add more file totals to show_campaigns -- we have pending files, but should have total in datasets, delivered, etc. rollups.
- Feature #16869: Interface to SNOW request and incident ticket.
- Feature #16782: Interface to ECL.
- Feature #16709: Link to some Kibana stats for each campaign.
- RITM0582786 from Alex. Run time and file transfer time histograms like job efficiency histogram.
- RITM0583138 from Alex. Implement the capability to select a bunch of campaigns from the list displayed and mark them inactive.
- Bug #17532: Unknown error when clicking Config work/compose Job type, the create a new job type.
- Bug #17525: Inactive campaigns (No activity in the latest x 7? days) shown as active.
- Bug #17524: Unknown error when clicking Campaign Data/campaign Stages to delete.
- Bug #17522: unknown error when clicking on configure work/Compose launch template to duplicating with the existing name.
- Bug #17506: limit launch template hosts to have experiment name in them.
- Bug #17327: bulk_update_jobs is cpu_bound and not keeping up.
- Bug #17284: Spreadsheet report page choked on large Campaigns.
- Bug #17264: project descriptions not going in Marc Mengel.
- Bug #17259: Delete function not working in the web application.
- Bug #17258: Strip spaces from fields before storing - causing apparent duplication issues.
- Bug #17204: Have the full string "Launch Template" readable from the POMS campaign info.
- Bug #17200: New campaigns don't appear in the list.
- Bug #15181: Campaign Layer Sheet - - cannot click through.
- Bug #17028. Proxy Error while loading POMS pages is appearing again. NOvA is experiencing this problem and opened a ticket (RITM0575905).
- Bug #16989. Rework show_campaigns.
- Bug #17024. Fix headers in show_campaigns.
- Bug #17027. Unsual Completed but not Located job counts for NOvA.
- Bug #17085. Use N/A to indicate when information in a certain table cell is not provided.
- Bug #17086. Use N/A instead of -1 in poms/campaign_sheet pending column.
- Feature #14287 More detailed overall campaign-at-a-glance status page
- Request #15560Alex H. suggests to have a 1-week time window in the default page of active campaigns and mark those with 1 week of inactivity as inactive with the default time window equal the duration of the campaign.
- Bug #11963. Accessing development POMS application on fermicloud045, occasionally I see this error message:
- The proxy server received an invalid response from an upstream server.
- The proxy server could not handle the request GET /poms/campaign_sheet.
- Bug #15042: 'IntegrityError' is not defined
- Bug #15043: The link on the number for Total Completed jobs on poms/show_campaigns page returns an empty set.
- Bug #15545: Non exit code showing. Not reporting an exit code via ifdh log, or the joblog_scraper agent being down/out of memory when the job was running. Now we may get the exitcode, etc. from the joblog parser
- Bug #15547: In cut versions, we print a "fatal" git error trying to get the version from git.
- Bug #15646 fts_scanner error on startup
- Bug #16403 Changed tag to version
- Bug #16405 Info in a time window are not correct.
- Non-new features in this version.
- Feature #16067: Use "streaming" mode in queries to reduce memory usage
- Feature #15058. We need several software test-stands:
- Feature #15144: Add mock webservice for testing agents
- Feature #15208: add tests to test/testUtilsPOMS.py
- Feature #15317: Add more tests to test/test_JobsPOMS.py
- Feature #15342: Add tests to test/test_AccessPOMS.py
- Feature #15343: add more tests to test/test_CalendarPOMS.py
- Feature #15344: Add more tests to test/test_CampaignsPOMS.py
- Feature #15345: Add mor tests to test/test_FilesPOMS.py
- Feature #15346: Add more tests to test/test_TagsPOMS.py
- Feature #15347: test_TriagePOMS.py Closed Marc Mengel
- Feature #15348: Add test to stand up a cherrypy instace and fetch one each
- Feature #15624: Add username column to experimenters table to fix VOMS import
- Bug #15751. The Tag is not working. This was discovered by Margarita when she did the test of this module.
- Non-new features in this version. This is a bug release.
- Feature #15945: Change "Campaign Layers" to "Campaign Stages" in page templates
- Feature #15946: Hide unused files-per-job fields
- Feature #15948: Change name to Campaign Layers
- Feature #15949: Splitted menu into: "Campaign info" and "Configure Work"
- Feature #15950: Change the order of the steps into "Configure Work" menu in to make the options more intuitive.
- Bug #15866: Moving task updates out of update_job_common, and handle them differently in bulk_update_job.
- Bug #15867: Updated some code to make poms installable via setup.py; which included a local poms
- Feature #15947: In main menu, get the composition/edit links under a different heading from info links
- Feature #14270: Allow users to run arbitrary executables/scripts and workloads through the system via arbitrary launch scripts/executables
- Feature #14271: Support the model of sending data to jobs
- Feature #14272: Supports data externally pre-staged manually at sites (i.e. different site local caches)
- Feature #14274: Have appropriate permissions and privileges implemented to authorize each operation
- Feature #14248: Now we are using Elasticsearch to fetch job data.
- Feature #15044: Allow editing of param_overrides on campaign recoveries through UI
- Bug #15007: If we get a repeated error (all retries fail) on a SAM, POMS Stop querying.
- Bug #15349: The job_log scraper had its reporting threads crashing and stopped actually reporting things.
- Bug #15332: Some jobs have two job_histories records for one job.
- Bug #15501: Campaign page breaks if accessed w/o authentication.
- Bug #15499: In output_pending_jobs POMS only lists files that match the output format for the job type. Furthermore, update_job only flags files as output files (and not log files, etc.) if they match the output format
- Bug #15447: Database deadlock error
- Update fife_launch in fife_utils to integrate poms_client calls #14471:
- Statistics pages use campaign history to pick software version avoiding change the version in the whole dataset. #12948
- The snapshot id field is filled when the task is created. #13888
- Active for experiement showed experiments you were not in. Currently, it check if the user is actually a member. #14463
- The incorrect Job_Type and Launch Template's names for all NOvA's campaigns were corrected using one for NOvA #14464
- Incorrect info for Launch Host, Launch Account and launch Setup for NOvA campaigns #14465
- Change the default analysis VO in the campaign definition to production VO #14554
- Assure that when the task go to located the jobs also go to located #14476
- Ability to get reports about how effective/efficient are the campaigns #13699
- Add to the Campaign (Layer) two fields: completion_type and completion_threshold. Feature #12751:
- Feature #14752: Add job log parser to get final job info
- The runtime,memory, and disk were included in the Campaign Def/Campaign info #12966
- Poms should keep a list of job launches requested while job launches are "held" #14659
- POMS hold job launches when services are "red" on dashboard. Feature #14660:
- Poms keeps a list of job launches requested while job launches are "held" Feature #14659:
- Add a script to extract VO Role=Production user info. #14747
- POMS is integrated with SAM putting information into SAM db about campaign and task id, then the info is fetched faster than using project descriptions Feature #14504
- New screen for a campaign "why is my job not running?". Integrate it with fifemon #12965
- The HTML templates were renamed to line up with the URL's, so they are easier to find. #12933
- Improve in the labeling of the job status, now the table describe the complete jobs with the total number, number of jobs located and number of jobs non-located.. #14477
- A complete refurbish of the POMS code in order to split in modules by functionality, having independent test units. #12934
- The Campaign Info page was updated for new fields added to Campaign(Layers), LaunchTemplates, etc. #14531
- Better split types were added allowing the user to divide the dataset for the campaign #13707 (In testing)
- The code for pending campaigns made a query that didn't work if you did the whole campaign. Furthermore, the generation of these dimension strings was factored out to common routines #13889
- When the jobs dropped out of the queue condor_q no longer suffices. Now it is stored into the elasticsearch logs. #14376
- Task completion/launching dependencies/recoveries not happening Bug #14866
- The uwsgi threads in development do not seem to be obviously leaking memory anymore either. Bug #14886
- Now it has a sub-item for each experiment in persistent DCache, and each goes to degraded at 80% and Failed at 90%. If enough areas fill up, we hold job launches. Bug #15008
- Include pending file counts on Active Campaign summary #13734
- Including some small features requested by NOvA as: change in campaign info stat time frame, button interface to mark campaign as inactive, specify output patterns in poms. #13849
- Solved the problem that efficiency histogram convolved low-efficiency jobs with jobs with unreported data, giving the wrong metric. #14245
- Compose CampaignLayer/JobType/LaunchTemplate have now a clone button, so you can clone an existing one and fix it up. #14251
- Update Terms Campaign-Layer/Submission/JobType versus Campaign/Task/CampaignDefinition to make it more clear for experimenters.
- Clean up compose CampaignLayer/JobType/LaunchTemplate in order to make the interface visually better for the users. #14249
- Reporting agents retrying on 404 errors forever in development #14246
- A bug that prevent submissions go into Located status #14242
- The code that was doing condor_q was adapted to be consistent with the changes made in jobsub #14247
- The problem of incorrect dataset definition for dependent job submissions was corrected. #14243
- Correction of the bug that was showing zero pending files for all campaigns #14250
- After the upgrade of the single sing-on the login code in fifemon status reader was no longer working #14332
- Add series of recovery types to campaign definitions; and to say one campaign depends on another. #12730
- Add CampaignDependency items that attach this campaign to other campaigns, along with a file-pattern-match for what output files from the previous campaign we want (default "%" for match-any). #12892
- Self dependencies run the next batch as soon as it finish. #13278
- Method in poms_service.py that looks for Campaigns that have had no tasks in the last week, and turns off their active flags. #12890
- Add CampaignRecovery entries attached to the CampaignDefinition, in order by "recovery_order" field, choosing recovery types from the RecoveryType table when you add entries. #12891
- After a job is marked as held then the system will kill the job. #12975
- The job launch command timeout in order it doesn't run forever. #12758
- The failed_jobs_by_whatever method loses tdays when you pick fields. The problem was solved. #12931
- Some pages are missing date picker/forward-back links #12932
- Launch template edit was failing with double quotes in setup block #13590
- Adding files-by-submission page showing requested,delivered, etc. #12193
- Each display page should handle time ranges #11386
- Campaign should launch jobs for campaigns that depend on this one. #12721
- Services status dashboard. #12620
- File-based Task status page #12193
- task bars and job bars need jobsubjobid, not internal task/job_id. #11382
- monitor how full dcache persisitent partitions are via webpage #12727
- joblog scraper was missing input files sometimes #12290
- Getting the correct percent efficiency for jobs
- build a status page with file-based stats as requested by OPOS
- Fix timezone shift in status pages #11379
- job reporter code did not retry properly on connection errors #12722
- Getting constraint errors when updating jobs frequently #12723
- Support older idfh better in joblog_scraper #12922