Status endpoint aggregator
Allow templating of the collection sharing web page
[keep-web] machine-readable file listings
Don't schedule jobs on nodes which are bigger than requested
Copying a project with a docker image doesn't copy the docker metadata
Support SLURM parameter passthrough from CWL jobs
Nginx config should speak JSON when returning its own response for unproxyable API requests
Deprecate / remove arv-web
Subprojects tab has wrong Description column
Update Pipeline Optimization wiki with CWL/Crunchv2
Activating an account (setting is_active from false to true) does not email the new user
[SDKs] [API] [keepstore] [keepproxy] Facilitate request tracing for arv-put
crunch-run handling non-group readable files in output directory
mandatory keep_web configuration missing from workbench install instructions
Collections/data access API
[SDKs] Improve wording of Keep error messages in Python SDK
[CWL] Run directly from git
[node manager] failing test
[Node manager] Fix flaky integration tests
CWL should check yml to make sure pdh's exist
Be able to create a new collection based on a subdirectory in workbench
crunchstat-summary gives invalid min_cores recommendation
Windows file system driver for Keep (arv-mount equiv)
Find arvados-cwl-runner version easier in a container request
[arv-mount] Do not crash if /sys/fs/fuse/connections is empty
On the collections page indicate expiring collections
[CWL] arvados-cwl-runner should support tagging output collection using properties
Migrate Collections.properties to JSONB
Timeout attempting to create job for component gatk-haplotypecaller-cram-gvcf: Net::ReadTimeout
Error fetching collection (HttpError 422) during call to arvados.util.listdir_recursive
[FUSE] high memory consumption (possible leak) in long-running arv-mount
[Workbench] Update job stats to be more useful to end user
[API] Remove async permissions code
[R SDK] Create a Bioconductor/R SDK
User accounts optionally have read-only access
[OPs] Jenkins test run ordering issue
[Workbench] Add a "Browse" tab that links to tree/master in repository#Show page
Convert Collection properties column type to JSONB
[Go] CLI programs should exit 0 after giving -help
[Workbench] incorrect text about log messages on empty collection
[Nodemanager] Improve status.json for monitoring
[Python SDK] Implement support for universal newline mode in Collections API
Support schema.org/Google structured markup for datasets
Standard convention for Nagios-compatible health check scripts kept alongside source for services
Import Docker images without requiring Docker on client
[Workbench] The sort order of items on /pipeline_instances and /container_requests page is incorrect
CWL Expression Tool Javascript issue when using str.includes
Using cwl, when you obtain files (file objects) from a directory listing they lose some fields when compare to files listed in cwl doc
[Workbench] Provenance graph consolidates scatter steps
[Workbench] Logs containers with undefined exit code
[Crunch2] crunch-dispatch-slurm hits scancel too much
[SDKs] Exceptions raised by Python SDK should include the "error token" if provided in the API response
[SDKs] publish python3-arvados-python-client packages
[Workbench] Replace schizophrenic generic "groups" page with "roles" page that only creates/lists groups with group_class="role".
[keepstore] Support X-Keep-Desired-Tier header
[keepstore] Add "tier" field to volume config
[node manager] not assigning weights
Example CWL code for cwl-runner doesn't work
[Workbench] Ensure that running processes are visible on the dashboard
[Workbench] A cancelled container request page should show the "Re-run" button
arv-put detects symlinks to common files and directories to avoid duplicated uploads
[API] [Wrokbench] set a default config.log_level
[API] [Workbench] config option serve_static_assets to serve_static_files
a-c-r should log the command used to run the pipeline
[crunch-run] Fix excessive memory use
[API] Delete old job/container logs in main server process, instead of rake tasks invoked from cron jobs
[CWL] Allow "keep:" references on command line.
[API] Limit number of lock/unlock cycles for a given container
[Workbench] Improve user page group memberships management [to be groomed]
[CWL] [Workbench] Support for copying/moving pipelines
occasional API server badness: 'incomplete response received from application'
[API Server] Support for pulling Docker images and uploading them to Keep
[Crunch2] Create workflow object for CWL runs which are initiated from the command line
Support display & reseting of expiration dates for collections
[FUSE] Integration test using real programs with complex write patterns.
[API] Webhook callbacks for server-to-server communication
[FUSE] Writable fuse creating conflicts with itself
workbench needs to handle API server timeouts better
[API] Excessive memory and time consumed by API and dispatch when container fields (like "mounts") are large
[API] clean up stale/conflicting dns data from deleted node records
Implement disk usage reporting for the administrator
User activity page takes too long to load
"arv collection list" command parsing allows invalid lines
HTTP 503 status for planned downtime
bcbio NA12878 validation runs: job re-use failure with non-existant collection
bcbio NA12878 validation runs: failed steps due to being unable to setup container
[API] container_requests#update alternately responds 422 or 404 for no apparent reason
Support federated search across a set of Arvados clusters
Federated user identity which works across a network of Arvados clusters
Support rendering of CWL description fields with embedded markup
[SDK] avoid interfering with socket open/close - use pycurl sockoptfunction instead of opensocketfunction
[SDK] Resolve symbolic names to uuid/pdh (cwl + other tools)
[Workbench] UUIDs & PDHs should be hyperlinked
[SDKs] support text-mode open() in Python 3
[SDK] tool for copying CWL file + input collections
[arv-put] Lumpy / slow upload
[Workbench] [UNC] Copying collections also copies metadata
[arv-put] bad progress indicator
[CWL] Fails with $schemas referencing remote resources.
[API] full text search on pipeline_instances table is extremely slow (wrong index?)
[API] "arv group contents" produces a 404 error (seems to POST instead of GET?)
[Tests] Make workbench integration tests finish faster (at least in Jenkins)
[SDKs] Writable collection files returned by Go SDK should be seekable
job_task creation fails with ApiError - HttpError 422 - ActiveRecord::StatementInvalid: PG::InternalError: ERROR: invalid memory alloc request size 1718630765
Docker load time on compute node is too long
[SDKs] Fix misleading arv-mount/pysdk error messages by removing obsolete "fetch manifest from Keep" code
[Keep] Support multiple storage tiers
[Deployment] Use consul to report health of services/components
[Nodemanager] Packaging for compute node setup scripts
[GA4GH] Participate in workflow interoperability activities
[GA4GH] Support workflow execution API
[GA4GH] Support task execution API using crunch v2 containers
[GA4GH] Expose workflow records via tool registry API
[GA4GH] Import from tool registry API
[arv-put] When debug mode is active, report the cache file name being to be used
[Docs] Document keepstore S3 volume configuration
[Docs] Migrate install docs from command line flags to YAML configs (keepstore, keepproxy, keep-web)
[API] create_token script should allow operator to provide token scopes and opaque comment
[Crunch2] Propagate dispatch error messages (e.g., sbatch fails) to user via logs/websocket
Write scalability test which can be used periodically as needed
[Crunch2] [Workbench] Show slurm queue position of containers submitted to slurm but not yet running
apt-get upgrade arvados-api-server on Ubuntu 12.04 did not update crunch-job
"arv keep_service create" should go to an editor.
Trying to copy a pipeline template to 9tee4 fails
"Run a example test" Button.
[SDKs] [Python] When reading data through Collection et al., signatures should refresh automatically when needed
Document how to choose a suitable blob signature TTL
[WB] 'Used by' tab empty when no graph is available
[keep-web] configurable block cache size
[CWL] Avoid using "+" char in mount paths
arvados-cli gem dependency error
[Workbench] Live stats should use consistent coloring for graphs
Switch from goamz to aws-sdk-go
[Crunch2] [API] return a more specific 422 error message when a client calls containers#unlock without having the lock
[Docs] [SDKs] Mention in API docs and discovery doc that destroy=delete and index=list
[Spike] virtualenv arvados-cwl-runner
[Documentation] we should not use github.com copy of arvados in the tutorial
[cwl] arvados-cwl-runner submitted job fails due to missing dependency futures>=3.0.5 in docker/jobs image
[Crunch2] [Crunch1] Disable crunch1 APIs by default on new installs
[API] [SDKs] When client and server versions differ, the server must either implement the API the client expects, or return an error.
[Workbench] Support array type input parameters to Workflows
[Workbench] Don't create manifest with 0 length chunk
Notification of job completion
[keepstore] Refactor volume Get/Put funcs to BlockReader/BlockWriter interfaces
[Crunch2] Propagate error messages if sbatch command succeeds but crunch-run can't run (or can't log to the Arvados API)
[API] [arvados-ws] remove all websockets and puma related code from the api server codebase
[arvbox] [arvados-ws] Make arvbox use the new arvados-ws
[Workbench] pipeline_instance#show page should offer “Copy to project” and “Move pipeline instance” options like most other show pages
add a --user-script option to install-arvados-jobs-image
[Docs] Missing docker_image link classes from API docs
[Workbench] "Public pipelines" on help menu links to wiki, not local public pipelines
[Crunch2] crunch-run fails trying to set swappiness on ubuntu 1204 kernel
Crunch1 jobs should still work on a slurm setup configured for Crunch2
[FUSE] Determine why bcl2fastq doesn't work with writable keep mount
[Node Manager] Doesn't kill itself when unhandled exceptions are raised during actor setup
Packaged go binaries should report their own version numbers
[Python SDK] Document undocumented methods in RichCollectionBase
Unable to build arvados-cwl-runner in 9tee4 environment
[CWL] a-c-r maximum recursion depth exceeded error
apps/workbench_functionals test failing
apps/workbench_integration test failing
services/nodemanager test_time_timout failing
[crunchstat-summary] should not report 0% CPU and 0 RAM in cases where stats were not reported at all
arvbox destroy fails because some files are owned by root
sometimes `arvbox reboot test` gets stuck asking for a sudo password
Collection file count & size information don't get dynamically updated on collection update
arvados-git-httpd returns error '500' (Internal Server Error) when a request is not authorized
arvbox loops forever when something goes wrong
implement approximate (estimated) counts for API list method
[Workbench] Add support for Trash Bin
[Keep] Share buffers between overlapping/consecutive GET requests for the same block
test suite hangs (in arvbox) in sdk/python test_callback
Multiple links with the same relevant information can be created
[crunchstat-summary] HTML plots should be a superset of text format output
[crunchstat-summary] Add scales to graphs
Integrate CI::Reporter to provide test output in JUnit XML format
[crunch-run] use docker instead of docker.io
[Python SDK] Provide more control over pre-fetch behavior
Multiple virtual machines with the same hostname can be created
Document crunchstat-summary
[SDKs] Fix memory leak in Python SDK Collection class
[Performance][Workbench] While in a resource #show page, switching between tabs results in fetching the object again from API server. This can cause unacceptable amount of latency for large objects such as large pipeline instances and also to collections.
4xphq keep balance api request causes oom on api server
Node manager should fetch pricing information from cloud APIs
[API] Remove unimplemented fields from api_client_authorization
[CWL] Reduce redundant work with subworkflow RunInSingleContainer
crunchstat-summary should explain why "(no report generated)" when possible
Request collections that don't (yet) exist via fuse interface
Faster downloading using arv-get
In the arvados sso database, multiple identical clients can be created
Add varchar_pattern_ops to all relevant PostgreSQL UUID indexes
Get list of Arvados configuration parameters in puppet that could be published by Consul
[Workbench] Enable refresh for "small" pipelines
[Workbench] Import CWL workflow
remove the need of yum reinstall from the installation/upgrade of Rails based components
arvados-cwl-runner fails to find secondaryFiles
[Workbench] Don't assume Switchyard config for SSH explanation
[Workbench] list of logins in the admin page isn't showing all logins
[CWL] a-c-r option to resubmit failed jobs
[Crunch2] crunch-run config option to use arv-mount for output dir
[Workbench] add a configuration flag to hide compute nodes
Output glob "./" doesn't work as output from CWL workflow
[Crunch2] [API] Accept collection UUID or PDH as container_image in a container request
[Centos7] [google_api_python_client] packaging with wrong permissions
[API] [Workbench] Error out when unrecognized keys appear in config files
[Go programs] Error out when unrecognized keys appear in config files
[Workbench] Pipeline instance shows wrong button labels in tab area
[CWL] Tries to run docker inside container where docker is not available, gives IOError
[SDK] Convenience method for resolving project/file paths
[Performance][Workbench] Computing user notifications in topnav is expensive. Can the user's properties be used instead?
[Performance] Fetching the pipeline instance qr1hi-d1hrv-jl9fvzy3x8sdj8v seems to take too long
[Crunch1] Log the node properties at the start of a job
Efficient queries for notification menu items
[CWL] arvados-cwl-runner may not be handling input files on keep mount correctly
Arvados should be able to cat/zcat files in collections easily
[SDK] Writable file-like interface for collections in Go SDK
[Node Manager] Support ephemeral additional storage for Azure node types that do not have sufficient instance storage
Provide more reasonable error messages for memory issues during container dispatch
Crunch job output logging improvement stories
Confusing job duration reporting
[SSO] sso server adds multiple records without checking
[Workbench] Jobs on dashboard say "No output" even when they have output
Additional Crunch job logging controls
[Tests] Fix (occasionally) hanging crunch-run tests
Automate starting/enabling services during Workbench and API package installation
[Packaging] Include sample config file (/etc/arvados/foo/foo.example.yml) with each configurable package
[API] In each request log entry, report time spent getting current permission graph
[Crunch] crunch-dispatch log throttling should not apply to its own stderr
Python RichCollectionbase's find_or_create() method callers may assume that found object is of "create_type" type
RichCollectionBase's _get_src_target method would crash if "source" argument is not basestring and "target_path" is empty.
keep-balance reporting large numbers of replicas lost
queued jobs in Paused pipeline instances start running anyway
pipeline_instance state can be changed from `Complete` to `Paused`
[Workbench] ContainerRequest#show crashes "undefined method `each' for nil:NilClass"
Move Container related documentation from wiki to code
[Workbench] Double-clicking "show version" link causes modal deadlock
[Workbench] Retrieve all log content directly from API instead of pre-rendering in Rails
[SDKs] Python SDK should use a more recent version of the google api client library
[Testing] TestErrorDuringRsync_FakeDstKeepservers
[Testing] test_shutdown_declined_when_idle_and_job_queued flaky
[Crunch2] [API] Permit container [request] with empty command (run default docker command)
[Crunch2] [API] Support container request expiry (expires_at)
[CWL][Crunch][Crunch2] crunchrunner should ignore hidden directories when uploading to keep
[CWL][Crunch2][Crunch] crunchrunner should use CWL globs to output data to keep
[FUSE] arv-mount installation documentation not working
[FUSE] performance issues stat() in python
[FUSE] Undelete collections by moving them out of the TrashDirectory
[keep-balance] Refuse to act when planning to delete data above an admin-specified threshold
[keep-balance] Do not delete blocks referenced by collections with replication_desired=0
login-sync config file & systemd unit
[SDK] Package Python apps as virtualenvs
[Workbench] file upload should provide progress bar for overall upload, not just individual files
Workbench file upload should allow the selection of directories
[SDK] portable_manifest_text() normalizes manifest, breaks CollectionReader() round trip
keep-balance fails with "Malformed index line" error
[SDKs] Python SDK duplicates resource attributes in top level of params
[SDKs] arv-put's default output includes both collection UUID and PDH
[SDKs] Improve metadata created by arv-put
[Workbench UI] Pipeline instances should have a link to view the CWL file used.
[CWL] Creating intermediate folders in working directory (fails in cwl-runner)
[Workbench] UI should advance to new step when user submits their input
[CWL] Secondary files entering into Arvados automatically
[Crunch2][Workbench] List both pipeline_templates and workflows on both pages
[CWL] cwl-runner parses nested arrays repetively
[Workbench] Ability to add group from Groups admin page
[keep-rsync] Rename "replications" to "replicas"
[keep-balance] "Expedited delete" tool: perform garbage collection on some specific (recently deleted) collections, bypassing usual GC race protections
[CWL] arvados-cwl-runner should reuse the most recent run
[CWL] Jobs crash when arvados/jobs:latest is not suitable to run CWL jobs
[Crunch2] crunch-dispatch-slurm can't successfully dispatch work when run from a directory it can't write to
[Crunch2] Document/fix handling of collections with non-nil expires_at field
[Data manager] Verbose reporting
[Node Manager] huge virtual memory size
[Workbench] Times out before API server loading active pipeline instances
[FUSE] rmdir on CollectionDirectory sets expires_at
[FUSE] Trash directory to list, inspect, and un-trash trashed collections
[FUSE] Don't display expiring collections in regular collection listings
[FUSE] llfuse thread can stay alive after umount, hanging tests
[FUSE] Make .arvados#collection available in collection subdirectories
[FUSE] .arvados#collection is missing or outdated after writing inside collection
[FUSE] Add timestamp to arv-mount --logfile option
add key/value support for tags
[Provenance Tracking] jobs refer to collections by uuids, not pdh
get rid of dependencies python-pyasn1 python-pyasn1-modules
[FUSE] arv-mount is too slow for strelka
[Documentation] Write a Crunch2 wiki page
[Documentation] Write a Crunch wiki page
[Node Manager] Aggressively throttles node creation requests when the cloud region is full
[Crunchv2] python thread error after job success
[NodeManager] CloudNodeListMonitorActor stopped reporting, and logs are not helpful enough to diagnose
[SDKs] PySDK fails to load CAs for SSL verification when run inside Conda
[Crunch] crunch-job can miss the final log output of a job
[FUSE] When we give up trying to write a block, the next operation on the file should fail
[Workbench] Bad URL quoting/parsing in File links from the pipeline template Components tab
[Python SDK] Add a Collection method to efficiently concatenate files
[Crunch] Limit maximum wait time between retries
[Crunch] [UX] Standard excepthook to help debug job problems
[SDKs] arv-copy fails to copy a Docker image specified with repo:tag syntax
[Workbench] The "Log" tab should have an notification that the job has been queued
[Node Manager] Less logging of tracebacks for routine exceptions
[API] Bad compute node address detection on private networks
[SDKs] WebSocketClient Exception 'NoneType' object is not callable
[Crunch] Log throttling shouldn't throttle logs from crunchstat, crunch-job
[FUSE] Add --umask, --fmask, --dmask options to control file permissions
[SDKs] Bad use of RetryLoop in events.py
[SDK] arvados.arvfile.readall_decompressed() does not return all blocks on a file
[Crunch] Job was repeatedly retried on same bad compute node until abandoned
[Docs] Many tutorial instructions are wrong outside Curoverse clusters
[Workbench] No user profile should be required in the default configuration
[Docs] Document Git 1.7.10+ dependency for shell node and HTTPS Git cloning
[Docs/Workbench] Document how to clone repositories over HTTPS
[Docs/Workbench] Improve SSH agent documentation
[Workbench] Hide file preview buttons when preview isn't possible with current keep-web config
[PySDK] API client logs a warning when you create an object without an owner_uuid
[Node Manager] Broken nodes always get shut down, exempt from all node-counting considerations
[Node Manager] Shut down nodes in SLURM 'down' state
[NodeManager] Use sys.exc_clear() to release exception tracebacks
[Node Manager] Minimum idle time before node is eligible for shutdown
[Node Manager] qr1hi nodemanager can't start if ulimit is in place
[Node Manager] Doesn't convert units when determining if a job's scratch request is satisfiable
[API] Go code to connect to the API server database and make basic sense of it
[API] Go code to implement the Arvados API permissions model
[API] List queries using "distinct" option return incorrect number of items_available
[API] Should not be necessary to host git repos on the same host as API server
[FUSE] Listing a project directory is slow when there are many subprojects
[FUSE] arv-mount should detect and log any files/dirs that are still open after unmounting
[FUSE] Reset mount-tmp to the state of another collection
[SDKs] Update run-command to use arv-mount --mount-tmp instead of staging directory
[API] Job model validates that the script exists in the repository at script_version
[SDKs] Port arv-keepdocker to use the docker-py library
[SDKs] `arv keep docker` can upload an image already saved as .tar
[SDKs] `arv keep docker` can make links for an image already in a collection
[SDKs] Clients can specify which auth config file to load.
[SDKs] PySDK one_task_per_input_file should support a num_retries argument
[SDKs] arv CLI program accepts "-" and "/path/to/file" as values for parameters that expect JSON (most notably the "body" parameter)
[SDKs] arv-copy continue on error
[SDK] arv-copy better Git repository error reporting
[FUSE] arv-mount exits non-zero if any I/O errors were returned
[API] Asset pipeline does not work unless config/environments/production.rb is present at buildtime and runtime
[API] CSS+images missing on login_failure.html
[Crunch] crunch-dispatch should run the bundled version of crunch-job specified by Gemfile.lock
[API] [Workbench] config:check catches and reports YAML parsing errors
[SSO] [API] [Workbench] Have config:check sanity check secrets
[API] [Workbench] Check for nonexistent config params during rake config:check
[CRUNCH] Fail gracefully when arvados repo doesn't exists
[Keep] keepstore status.json should report runtime environment
[Workbench] Crashes with Fiddlesticks when trying to view a pipeline instance graph copied from another cluster
[Workbench] [Fiddlesticks] Live job log should show recent history
[Workbench] Doesn't explain why a job can't be cancelled
[Workbench] Show all known runtime_constraints in job status pane
[Node Manager] Detect and shut down nodes that fail provisioning in Azure
[Keep] Keepstore and keepproxy support HTTP Range requests
[SDKs] [FUSE] Python SDK and arv-mount use Range requests when a caller requests part of a block that has been ejected from the cache
[API] Job log cleaner should not clean logs when the job's log is the empty collection
[SDKs] Document that arv-copy accepts collection PDHs (ie content addresses) as arguments
[SDKs] Recursive pipeline instance arv-copy doesn't copy job objects, losing metadata
[Crunch] Unhelpful translation of script parameter outside script_parameters
[SSO] Template calls undefined method after fresh local accounts installation
[SDKs] Unhelpful 404 stack trace when trying to arv-copy a collection without access
[git] "git clone https://..." with a bogus token should respond 4xx, not 5xx
[Keep-store] make keep-store lookup token at runtime like keep-proxy does
[SDKs] arv-copy continues copying when a collection's content address is wrong, but warns and exits a special code
[SDK] arv-copy automatically creates new repository(ies) on the destination cluster
[SDKs] Teach arv-copy to copy whole projects
[Workbench] node visibility as admin vs ordinary user does not line up
[Keep] keepproxy should log real IP address
ubuntu package libarvados-perl should depend on libcrypt-ssleay-perl
[Docs] Refresh Google OAuth2 setup in SSO install guide
[Docs] Install guide keepstore runit script has bad syntax
[Documentation] RVM and passenger recipes are missing details
[Deployment] arvados-api-server-upgrade.sh should set up git_repositories_dir
[login-sync] Sets up settings.conf
[Documentation] Various small bugs in the CLI SDK install guide
[Documentation] Various small bugs in the install guide
[Docs] keep-web runit script depends on sudo configuration
[Documentation] Clarify PostgreSQL setup commands for Software Collections users
[Docs] Install guide missing Websockets details
[Wiki] Clean up Computation_and_Pipeline_Processing page
[Docs/Deployment] MUNGE, Nginx, PostgreSQL services don't start on boot on CentOS 6
[Documentation] Install guide incurs RVM's sudo wrath
[Docs] SSO install guide documents default_link_*
[Documentation] Install guide explains a few different deployment options
[Documentation] FUSE documentation should note granting FUSE permissions
[Documentation] Make keepproxy install guide follow nginx conventions
[Tests] Make slow tests optional (e.g., run-tests.sh --short)
[Workbench] Many shell node instructions are wrong outside Curoverse clusters
[Workbench] view log via keep-web test fails in arvbox
[API] specifying an invalid filter should result in an error
[API] Support numeric filter comparisons for numeric fields
[API] Synthetic collections include the "kind" field
[Deployment] Test distribution packages
[Workbench/API] Does not derive key name from comments with whitespace
[API] [Maybe] Update API server to support Git 1.7.2
[Node Manager] Mishandles stop signals
[Node Manager] a node filled up its /tmp which killed slurm and docker, node unusable. Node manager did not kill the node.
[API] Trying to update a collection by PDH raises a NoMethodError
[API] "order" params should be rejected or used (not silently ignored) even when they reference columns missing from the "select" param
[API] Accept collection UUID as a job's docker_image constraint
[API] Finalize and document the collections/provenance and collections/used_by API calls
[Keep] keepstore should notify sysadmin about error conditions (trying harder than just log.Print())
[SDKs] Remove ruby dependency from arv wrapper
[SDKs] arv-copy should not require --dst-git-repo when copying on the same cluster
[SDKs] Python Keep block-r/w errors should indicate which collection/file was unreadable/unwritable as a result
[SDKs] arv-put progress meter scales based on file size, rather than being fixed to MiB
[SDKs] arv-copy raises an unexpected backtrace on ^C
Out of memory when using arv-copy not handled gracefully
[SDKs] Recursive arv-copy should report progress of the entire copy, in addition to individual objects
[SDKs] Refactor arv-get/put/copy into the Python "arv" wrapper using common exception-handling and argument parsing
[SDKs] Ensure `arv keep put` replicates properly
[SDKs] Go Keep client enforces a minimum transfer rate for proxy connections
[FUSE] [SDKs] Ensure FUSE and other SDK clients work with a default replication level != 2
[SDKs] Fetch all results with the CLI `list` command
[SDKs] arv get (and edit) should support a flag that suppresses display of non-writable attributes like "status" and "writable_by"
[SDKs] Add client-side join switches to `arv foo list`
[SDKs] GoSDK arvadosclient parameters are more transparent
[SDKs] PySDK Collection class lets you iterate files and subcollections separately
[SDKs] Git credentials helper in CLI tools
[SDKs] CLI tools install tab completion aids
[SDKs] arv-copy copies all metadata associated with an object
[SDKs] Can't `arv get` a collection by PDH
[Workbench] Pagination sets limit on last page, causing unintended side effects
[Workbench] Copying jobs causes "SubmitIdReused" error if submit_id is not null
[Workbench] Project listings act like file manager directory trees, with expanding+collapsing trees
[Workbench] Remove Keep disks link?
[Workbench] Ability to create a new project when creating a new pipeline instance
[Workbench] User admin interface should show which VM(s) the user can login to
[Workbench] workbench move and copy is incomplete
[Workbench] warn user when navigation will cause input text to be lost
[Workbench] Pipeline template description display is difficult to read when making a pipeline template instance
[Workbench] Visible blanking between partial content reloads
[Workbench] Change project "Jobs and pipelines" tab to just "Pipelines"
[Workbench] Add an "invert selection" button to checklists alongside "un/select all"
[git] Should be able to clone git repositories that are shared with "anonymouspublic" without authentication
[FUSE] With --exec, expose the mount point in an environment variable
[FUSE] arv-mount usage message shows incorrect order for [--exec ...] option
[Crunch] crunch-job does not sufficiently sanitize job parameter names when putting them in the environment
[crunchstat-summary] jobs with tasks_per_node>1 need node size recommendations adjusted
[Keepstore] Support a portable data challenge API
[Keepstore] Don't trash/delete blocks until they appear on multiple trash lists
[API] Do not accept an empty SSH key as valid
[Workbench] Project description styled differently based on whether or not the user can write to the project
[Keep] multiple response.Writeheader calls
[SDKs] `arv create` creates an object from an empty file
[Deployment] Update Rails package dependencies
[Documentation] Add Gitolite sanity check to install guide
[Docs] file_names Collection field is undocumented
[Deployment] .yml.example files should be consistent
[Documentation] Document installing the PySDK at a different PREFIX
[Documentation] A user with an older install may land in a page that requires the "latest" prerequisites.
[Workbench] Displays content at the top of the dashboard, set by the sysadmin
[Workbench] Provide an interface for the sysadmin to set Dashboard banner content
[Workbench] Provide an interface to delete files from collections
[SDKs] Provide a CLI tool to add files to an existing collection
[SDKs] Provide a CLI tool to remove files from a collection
[API] Advertise anonymous token in discovery document
[Crunch] Crunchstat and arv-mount print final stats before exiting
[Workbench] Pipeline template listings should include links to the scripts
[Workbench] Reusing ArvadosResourceList results in inconsistent results
[SDKs] CLI tools support Mac
[GATK Queue] Support a mechanism to redirect stderr for children
[Docker] API server starts before PostgreSQL is ready
[Docker] Bootstrap font fails to load in Workbench Docker image
[Docker] Fix SSO configuration in published images
[Docker] API server uses omniauth.rb for SSO configuration, not application.yml
[Docker] Workbench image doesn't serve HTTPS
[Keep] Write performance tests
[Performance] Compare Collection performance benchmarks across components
[API] [Workbench] Generate debconf template from config yaml
[FUSE][Performance] Job to md5sum 1TiB of 1MiB files in parallel is very slow
"This website is under heavy load" error makes a parsing error
[FUSE] Add a --replication switch
[DRAFT] [Workbench] More encouragement for users to send feedback
[Keep] keepstore reboots GCE host under heavy load
[Workbench] Provide a GUI interface to arv-copy
[SDKs] Whenever `arv` has an invalid resource name argument, it should complain about that
[Workbench] [Webshell] Rough edges in web shell
[Workbench] Anonymous user is listed in the "Share with users" popup when anonymous viewing is disabled, while "Share with groups" popup omits the anonymous group
[Documentation] document installation of node manager
[Documentation] webshell installation
Port arvados-libpam to centos
[SDKs] Add runtime option to SDKs (esp Python and arv-mount) to use a filesystem directory block cache as an alternative to RAM cache.
[DRAFT] Provide one clear way for users to get data from an external source into Arvados
[Workbench] [API] Support "world-readable, but not world-searchable" permission on projects: "anyone with the link can view"
[SDKs] CLI SDK tools use Ruby SDK
[Workbench] Allow additional curation of the public projects listing
[Deployment] [Documentation] Minimize system-wide dependencies for compute node setup
[SDKs] Go SDK locator regexp (and its tests) should agree with the documented format
[Node Manager] Booting nodes shouldn't satisfy min_nodes
[Maybe] [SDKs] Support caching Keep blocks in memcached
[FUSE] Support scaling the internal block cache based on number of open files
[Workbench] Public project listing should exclude subprojects
[API] [Workbench] Home project should be a real object (group with group_class="project"), not current_user.
[Workbench] Add social sharing buttons to public projects [and other data?]
[Workbench] Propagate old search engine meta tags in sync with Open Graph data
[Workbench] Make public data crawlable
[Workbench] Breadcrumbs can occupy multiple rows when the screen is narrow
[Workbench] Support searching collections by their collection name
[API] Include a mechanism to automatically trust some clients
[API] walk api server installations and ensure modified_at for collections is unique + ensure modified_at is enforced to be unique at the api level.
[Keep] Design: using Arvados to browse, and run jobs on, input data that is stored remotely (not on local Keep disks)
[Keep] Use data stored remotely (not in local keepstore)
[SSO] Authenticate users with OAuth2
[Workbench] The collection information box values for content address and content size should be updated automatically after new file(s) are added via browser upload.
[FUSE] `ls` in a large home directory (4400+ items) takes too long
[SSOs] Support authentication against multiple backend LDAP servers
[Keep] Use bufferpool in keepproxy to limit memory use when there are many concurrent PUTs
Slow to log-in to Workbench
[DRAFT] [Workbench] Users can be notified about changes to important pipelines
[Workbench] Provide additional guidance to users running their first pipeline
[Workbench] "other collections have the same content" page is confusing
arv-get is slow for this collection
Syntax highlighting of viewing source in workbench
Fiddlesticks when trying to view a log file
[SDKs] PySDK's internal decorators lose method metadata
[Workbench] collection#show performing too many unnecessary API server requests when accessing qr1hi-4zz18-zn1m6s2le7dqjfs.
[API] crunch-dispatch should mark a job failed when its repository cannot be fetched
[SDKs] Improve certificate validation in Python SDK
[Workbench] collection#show is unresponsive for a large collection
[API] Improve performance of large requests in parallel
[SDK] Implement arvados.gettaskparam(*args) to match arvados.getjobparam(*args)
[FUSE] Catch and recover from any API client exceptions
[FUSE] Ensure all I/O errors properly distinguish "not found" and other API errors
[Workbench] collections#index returns 504 Gateway Timeout
[Workbench] PipelineInstancesTest#test_Create_and_run_a_pipeline is not reliable
[Docker] Needs smarter dependency ordering when restarting existing containers
[Workbench] Mobile menus lack contrast (gray text on aubergine)
[Crunch] Run a local job using a docker image directly from the shell VM (without writing the docker image to Keep)
[API] [DRAFT] Provide API methods for manipulating and combining collections
All clients authenticate directly against SSO server without going through API server
[Crunch] crunchstat should not report errors during normal timing races
[API] Relax repository name constraints to permit [-_.] -- ideally allow anything github allows
[Workbench] Create new collection with selected files copy streams as well as files
[SDKs] [refactor] Dry up KeepService get() and put() in arvados/keep.py
[SDKs] arv keep docker --download to get image from Arvados
[Workbench] Support entering boolean script_parameters when editing a pipeline instance
[OPS] Routinely check for stale Gemfile.lock
[API] Websockets subscribe API should accept a start timestamp, so clients don't have to miss any events logged between the given start time and websocket connection success.
[Workbench] Link to script sources on github when repository is a github url
[Workbench] [DRAFT] Fix clock/node time reporting on Workbench pipeline page
[OPS] Add diagnostics script for shell nodes
[Node Manager] Reuse node records after shutting down the cloud node set up with them
[Workbench] Replace type-specific #index pages with generic code
[Workbench] Provide a guided workflow to new users who want to upload data and run a public pipeline
[FUSE] cd/ls sometimes takes too long on su92l
[FUSE] File access sometimes takes too long on su92l
[Documentation] Document the minimum docker image / Dockerfile that can be used to run a job.
[SDKs] one_task_per_input_stream fails due to unsupported APIs. Remove it from docs, examples, and SDK.
[Workbench] Missing link target in iframe error message
[Workbench] Make Dashboard contents filterable
[Workbench] Favorite projects, pipelines, collections
[Workbench] Indicate related pipeline instances
[Workbench] Link to log from pipeline and/or job summary page
[Workbench] 504 error returned when trying to create a large new collection
[Workbench] Pipeline instance "Re-run with latest" button occasionally returns Not found page
[FUSE] Support efficient copy at command line
[Workbench] Log Graph - Feature requests
[Workbench] Improve presentation of anonymously-viewed collections (via both "sharing link" and regular copy-and-pasted Location)
[API] [Workbench] [Performance] Use object caching to improvement performance (such as memcache).
[DRAFT] [API] Good API for accessing the old_ and new_attributes in Logs table properties column
[SDKs] PySDK test_rewrite_on_empty_file is not reliable
[Workbench] [Performance] Combining big collections should start returning a response faster (currently you can get a 502 proxy error even if the collection still combines)
[Crunch] Node went away unexpectedly during pipeline run.
[Documentation] Improve SDK documentation
[API] Replace is_trusted attribute with a per-{user,client} flag which is set by the user on first login to a given client
[Crunch] crunch-dispatch should use the API instead of running in the API server's rails environment.
[API] [Workbench] Add read-only ("locked") flag to projects
[OPS] Use packer to generate docker images for our various services
[Workbench] Workbench should support viewing VCF files
[DRAFT] [Crunch] User can configure task retries
[Workbench] Change "recent pipelines" to curated selection of "public pipelines"
[Documentation] Run an uncommitted script straight from a working directory using arv-run-pipeline-instance --run-here
[Crunch] Job logs include node-level resource usage statistics (cpu, ram, swap, etc)
[DRAFT] [API/SSO] Use externally-defined (e.g., LDAP) groups as permission groups
[Tests] Enhance keep testing to support signed locators.
[SDKs] Add Python version test 2.7 <= sys.version_info < 3
Job failure due to 'arv-put' 'ConnectionError'?
[Docs] Note apt key information in Python package READMEs
[SDK] CLI tools should catch socket errors like "Connection refused"
[SDKs] arv-put --description X --properties P should set respective attrs on new collection.
[Testing] Make run-tests.sh able to run only unit tests or integration tests (or both)
[Testing] Add a new unit-test-only build step to the build pipeline, before the current test step that runs integration tests
[Crunch] Contain all job/task containers in a memory-limiting cgroup
[Node Manager] add one integration test for node manager
[SDKs] arv keep put blows up when trying to put a danging symlink
[OPS] Write a wrapper script that calls packer to generate customized docker images for a new installation (with IP ranges, hostnames, etc preseeded)
[Documentation] Fix Go SDK comments so they look reasonable in godoc
[SDKs] arv-run-pipeline-instance should accept project and title parameters
[Documentation] Show footer with date last updated
[Workbench] API token provided in query string should not be ignored just because the session already has a token.
[Tests] Provide a docker image, based on puppet/packer process, that can run the entire test suite quickly.
[DRAFT] [Workbench] Improve provenance/used-by graphs so they are legible even for large pipelines / long provenance trails
[DRAFT] [Workbench] Improve display of pipelines and jobs
Improvements to sortable log viewer
[Workbench] Enable users to share with any other user
[Documentation] arvados VM simulator, demonstrating how it looks to use Arvados from CLI
[DRAFT] [Workbench] Browse change history for records
[Docker] Auto setup as much as possible for new users
[SDK] Go KeepClient offers a "read block" API that recovers from transmission errors by attempting alternate servers.
[Workbench] Buttons should have consistent ctrl-click and right-click behavior
[Docker] Make API server root URL in discovery document easier to resolve
[Workbench] Additional settings for ssh/config from Manage Account page
[Workbench] Extend Sharing tab to collections
[Tests] Implement long-running keepstore services in run-tests, similar to what api got in #3021
[Tests] Reduce dependency on integration tests
[Workbench] Public project page should encourage visitors to try using Arvados themselves
[Workbench] Pipeline re-run dialog should offer docker image choices (latest/same), similar to script version choices
[SDKs] Use sphinx to render for Python SDK docs
[Tests] Add crunch-job functional tests to the CI suite
[SDKs] CLI tools fail or warn when current working directory is gone
[DRAFT] [API] Enrich supported dataclass values for pipeline templates
[Workbench] User can trash and un-trash items
[SDKs] arv-get should support getting multiple files by shell glob
[Crunch] Use uncommitted code in a git work tree to run a dockerized crunch job in a VM (using crunch-job or arv-run-pipeline-instance --run-jobs-here)
[Crunch] arv-run-pipeline-instance --run-jobs-here cannot run jobs with local script path in script_version.
[SDKs] Refactor run-command so it can be used as an SDK by scripts in a git tree
[SDKs] CLI tools support project paths in addition to uuids
[arv-mount] Change default mount type from "fuse" to "fuse.arvados"
[Workbench] improve error reporting for failed AJAX calls
[Workbench] Suppress "ajax failed" error messages when ajax fails due to browser navigation event.
[Deployment] Switchyard should behave correctly when one SSH key has access to multiple VMs using different usernames.
[SDKs] File access on keep mount giving I/O error
[Tests] Eliminate system globals in test suites
[Workbench] after file upload, the files tab should refresh immediately (overriding the 15 second throttle)
[API] Prohibit leaving jobs in an inconsistent state (e.g., state=queued with non-null log or output)
[SDKs] CLI tests should be better isolated from user configuration
[Workbench] Workbench provenance graph tabs appear empty when graph is very large
[Docker] fix @_NOT_FOUND_@ hack in config.rb
[Keep] Keepstore queries API server to verify tokens
[Workbench] Merge "user agreement" and "required profile fields" into a single page
[Crunch] Automatically restart jobs after internal/intermittent errors
[Docker] Make it easy to add data from the host system to containers
[Workbench] Pilot material design theme
[Workbench] Make pipeline rerun UI more like jobs'
[Node Manager] EC2 driver creates nodes configured to terminate on shutdown
[Node Manager] Don't reuse node table records
[Workbench] When showing a shared project, do not include "Home" in breadcrumbs.
[Crunch] Choose worker nodes where the desired docker image is already available.
[Tests] When comparing pipelines, the graph should highlight the differences by using different colors for pipeline A/B/both.
[Workbench] Page for downloading collections with anonymous sharing link does not have any styling.
[Documentation/FUSE] Document '.arvados#collection' in shell VM keep mount
[SDKs] `arv keep docker` should pipe the data directly to Keep instead of store-and-forward, and provide some progress feedback up front in any case.
[Workbench] Make "Compute and job status" pane update dynamically
[Workbench] File upload should start immediately, not wait for user to press "Start"
[Crunch] Quick turnaround time when updating a docker image in order to re-run a job
[Workbench] When using "Run pipeline" button at the top of project A, the pipeline template chooser should start out showing project A, not "All projects".
[Workbench] Editable "name" column is presented but not actually available for "other" objects such as Human and Specimen.
[SDKs] [DRAFT] arv-get with just one locator argument should download the entire collection
[Workbench] Collection content hash should stay up-to-date (e.g., while uploading files)
[Crunch] run-command should show helpful error message when task.foreach argument is not a list
[API] Discovery doc should fix its incorrect "list" method description to be identical to "index"
[Documentation] Run-command docs should remind user how & why to exit non-zero on failure.
[SDKs] arv-put should assume --no-resume when --raw is given
[Tests] Diagnostics suite tests that live logs are displayed while jobs are running.
[API] Prompt user when logging in to a client for the first time.
[DRAFT] [Workbench] Prevent CSRF and XSS attacks
[Workbench] Search box on collections page should use infinite scroll instead of (broken) page number links
[Workbench] Create and edit pipeline templates
[Workbench] Projects dropdown should have a free text filter box
[Crunch] crunch-dispatch's Bundler environment infects non-Docker jobs
[API] Remove success and running attributes from job data model, since we are now using the state attribute.
[Workbench] Websocket client should only subscribe to events about relevant objects, not all events.
[Workbench] Show more (earlier) log entries for a running job
[Documentation] Introduce compute workflow / pipeline concepts in user guide
[Documentation] Best practices for getting the most out of job re-use
[Documentation] Set of introductory tutorials each focused on accomplishing a simple task
[Documentation] Tutorial for adding a crunch script wrapper for an existing tool that is already hosted on github.
[Workbench] Sortable column headings (filename, size) in collections#show
[Workbench] Disable auto-refresh on tabs with infinite-scroll or editable content.
[Crunch] Do more error-checking and show more diagnostic info when installing a Docker image
[SDKs] CLI tool to view real time logs from running jobs/pipelines, much like the Log tab of the Workbench PipelineInstance#show page
[Workbench] "Add data from another project" modal supports regex filtering
[SDKs] Option to watch logs (instead of exiting) when queuing a job or pipeline instance using arv job create, arv pipeline run, etc.
[Workbench] "Download" / "Save as" button for provenance graph
[Workbench] Reveal log collection ID in pipeline/job summary and job Log tab
[Workbench] Project show page should have buttons to add objects besides Collections
[API] API method and CLI shortcut for refreshing the signatures on some block locators (without creating a collection)
[Workbench] Display portable data hash, parent project, file count in Collection tables
[Workbench] Job detail page should show the job status when the job is not running or has completed running (button that shows 'queued', 'not ready', 'failed', 'complete', etc)
[API] Arvados clients can use standard OAuth2 protocol instead of custom token handling mechanism
[Workbench] Each active job/pipeline on the dashboard should show how many worker nodes are allocated to it.
[Workbench] Job view should include the relevant component name(s) from each pipeline where the job was used
[Workbench] Pipelines that are in New or Ready state should show up somewhere on the dashboard, (e.g., in "Active pipelines" or "Recent pipelines" table).
[Workbench] Make duplicate name error message more user-friendly
[Workbench] Provide feedback on dashboard to indicate that NodeManager is booting a node.
[SDKs] "arv recent [objectType]" lists recently modified objects
[Crunch] Pipelines should not fail immediately after one job failure, but continue running as much as possible
[OPS] Improve and automate non-shared VM configuration
[Workbench] Report a Problem" menu item does not like being opened in a new tab/window.
[Workbench] Don't show empty projects/collections when using the chooser dialog to select pipelines, files, etc.
[Tests] Websockets should work out of the box in test/development modes, even when using an ssh tunnel and a self-signed certificate.
[API] websocket clients should not put API tokens in path or query string
[API] [Workbench] Site-wide regex searching
[Workbench] Add a "docker images" page to Workbench that lists available docker images
[API] [SDKs] Include provenance information for uploaded Collections
[Workbench] [API] "Manage" permission should be sufficient to delete subprojects
[Workbench] Order by Collections "Name" column puts "folders" (collections containing files) at the top
[SDKs] When editing a template in arv-edit and faced with a parse error, when trying again, it should automatically take you to the line/column of the error
[API] API method and interface to check if a bugfix is deployed on a cluster
[OPS] Implement new deploy procedure
[SDKs] arv create needs --project-uuid and --name parameters for appropriate classes
[Docker] Fix supervisor.conf to supervise services instead of running init.d start script and hoping everything is OK
[Crunch] run_command should use SDK's retry support
[API] as admin user, permission denied canceling a job (started by another user)