696b1e7 | Thomas B. Mooney | 13 March 2024, 13:58:29 UTC | Merge pull request #2077 from tmooney/retry_connections Retry postgres connections for up to 90 seconds. | 13 March 2024, 13:58:29 UTC |
dd19bb8 | Thomas B. Mooney | 12 March 2024, 18:22:29 UTC | Retry postgres connections for up to 90 seconds. We sometimes get our TCP connection failing trying to connect, so let's try harder to successfully connect. (This depends on genome/docker_genome_perl_environment#39 to work on our cluster.) | 12 March 2024, 18:22:29 UTC |
84b1df4 | Thomas B. Mooney | 21 November 2023, 18:53:22 UTC | Merge pull request #2076 from tmooney/smaller_allocation_request_in_somvar_tests Use a smaller allocation size for testing. | 21 November 2023, 18:53:22 UTC |
2f5bfbf | Thomas B. Mooney | 21 November 2023, 18:28:51 UTC | Use a smaller allocation size for testing. The test data is about 50MB so a 100MB allocation provides plenty of room. This prevents failures when running tests on a blade with <32GB of free temp space, since we defintiely don't need that much! | 21 November 2023, 18:28:51 UTC |
2a4735d | Thomas B. Mooney | 21 November 2023, 17:44:54 UTC | Merge pull request #2074 from tmooney/custom_per_job_queue_in_wdl Define a "queue" runtime attribute for WDL tasks. | 21 November 2023, 17:44:54 UTC |
4aea4ff | Thomas B. Mooney | 21 November 2023, 17:44:38 UTC | Merge pull request #2075 from tmooney/avoid_loading_oltp Avoid attempting to connect to OLTP database | 21 November 2023, 17:44:38 UTC |
27d097e | Thomas B. Mooney | 20 November 2023, 22:04:27 UTC | Skip tests that require an OLTP connection to function. | 20 November 2023, 22:04:27 UTC |
9303a5e | Thomas B. Mooney | 20 November 2023, 22:01:58 UTC | Only consider disconnecting datasources that we've loaded. This is a check at an even deeper level than whether the default handle exists--we don't need to worry if the database could be connected to here if we haven't connected to it yet. (With the shutdown of our OLTP server this become relevant--we'll have a backup database we could connect to sometimes but not always.) | 20 November 2023, 22:01:58 UTC |
b59277f | Thomas B. Mooney | 15 November 2023, 18:38:35 UTC | Define a "queue" runtime attribute for WDL tasks. This will let certain jobs be dispatched to different queues. | 20 November 2023, 20:32:46 UTC |
595c5ee | Thomas B. Mooney | 14 November 2023, 20:39:03 UTC | Merge pull request #2073 from tmooney/deprecate_legacy_lims Remove LIMS-specific extensions. | 14 November 2023, 20:39:03 UTC |
9101534 | Thomas B. Mooney | 13 November 2023, 21:04:34 UTC | Remove LIMS-sepcific extensions. With the retirement of the legacy LIMS system these are no longer functional. (Many of the classes under Genome/Site/TGI/Synchronize are also likely candidates for removal, but there will be a final database snapshot we could theoretically query for a time as needed.) | 13 November 2023, 22:23:32 UTC |
4755ca7 | Thomas B. Mooney | 26 October 2023, 20:01:57 UTC | Merge pull request #1987 from tmooney/lsf10_memory_units Use explicit units for memory reservations (LSF10 feature). | 26 October 2023, 20:01:57 UTC |
2ded320 | Thomas B. Mooney | 26 October 2023, 20:01:48 UTC | Merge pull request #2072 from tmooney/only_delete_archive_location_if_we_created_it Only cleanup if we started to try to do work. | 26 October 2023, 20:01:48 UTC |
d4ed2ec | Thomas Mooney | 26 October 2023, 05:30:24 UTC | Only try to cleanup the archive if we started to try to do work. We shouldn't try to clean up if we're erroring out before attempting the archive-especially if the data was already archived! Also Switch to Try::Tiny for better error handling--instead of lamenting the lack of finally blocks, use one! | 26 October 2023, 16:40:48 UTC |
11154ea | Thomas B. Mooney | 18 October 2023, 16:18:14 UTC | Merge pull request #2071 from tmooney/read_only_tie Only tie read-only since we don't need to write this file. | 18 October 2023, 16:18:14 UTC |
44b0fb9 | Thomas B. Mooney | 18 October 2023, 15:21:04 UTC | Only tie read-only since we don't need to write this file. Also, actually report the error when the tie fails instead of mysteriously failing to parse the header later. | 18 October 2023, 15:21:04 UTC |
2662511 | Thomas B. Mooney | 17 October 2023, 16:30:04 UTC | Merge pull request #2065 from tmooney/restart_cloud Support restarting if our local build fails but the cloud VM is still running. | 17 October 2023, 16:30:04 UTC |
2baebf8 | Thomas B. Mooney | 17 October 2023, 16:29:38 UTC | Merge pull request #2070 from tmooney/unstartable_on_bsub_failure Don't crash if `bsub` for inner job fails. | 17 October 2023, 16:29:38 UTC |
ccc6ae4 | Thomas B. Mooney | 17 October 2023, 16:29:18 UTC | Merge pull request #2067 from tmooney/lsf_parsing_hosts Centralize checking for the existence of eligible LSF hosts. | 17 October 2023, 16:29:18 UTC |
9eefa3e | Thomas B. Mooney | 11 April 2023, 16:47:19 UTC | Centralize checking for the existence of eligible LSF hosts. There were a few competing implementations around for this; better that we have one that we can fix up when/if issues arise. | 06 October 2023, 01:41:30 UTC |
ea6c3ae | Thomas B. Mooney | 11 September 2023, 17:38:14 UTC | Don't crash if `bsub` for inner job fails. A misconfigured AnP could result in the bsub failing for various reasons. Better to skip it and move on than to prevent any other queue processing. | 11 September 2023, 17:38:14 UTC |
721185c | Thomas B. Mooney | 11 April 2023, 22:19:00 UTC | Merge pull request #2069 from tmooney/skip_memcache_test Add done_testing so that test-tracker handles skip properly. | 11 April 2023, 22:19:00 UTC |
d208ec0 | Thomas B. Mooney | 11 April 2023, 21:07:44 UTC | Add done_testing so that test-tracker handles skip properly. As also seen in #1980. | 11 April 2023, 21:07:44 UTC |
7756dc6 | Thomas B. Mooney | 11 April 2023, 20:35:26 UTC | Merge pull request #2068 from tmooney/skip_memcache_test Skip Memcache.t that depends on production memcache. | 11 April 2023, 20:35:26 UTC |
c93f871 | Thomas B. Mooney | 11 April 2023, 17:18:35 UTC | Skip Memcache.t that depends on production memcache. It's also not really safe for concurrent test runs! | 11 April 2023, 17:18:35 UTC |
51e4e50 | John Garza | 27 March 2023, 21:23:41 UTC | Merge pull request #2066 from johnegarza/cache_hashing_typo Fix typo | 27 March 2023, 21:23:41 UTC |
3db4c5c | John Garza | 27 March 2023, 19:22:15 UTC | Fix typo | 27 March 2023, 19:22:15 UTC |
6ce0e71 | John Garza | 12 August 2022, 17:29:14 UTC | Merge pull request #2064 from johnegarza/better_pp_installation Add multithreading and suppress output | 12 August 2022, 17:29:14 UTC |
0382498 | Thomas B. Mooney | 04 August 2022, 20:17:16 UTC | If we detect an existing VM for our build, then use it. | 04 August 2022, 20:32:44 UTC |
5551784 | Thomas B. Mooney | 04 August 2022, 20:06:30 UTC | Extract subroutine to check if cloud instance is running. | 04 August 2022, 20:14:53 UTC |
263122c | APIPE Tester | 12 July 2022, 14:44:42 UTC | Updated `gmt` tab completion. | 12 July 2022, 14:44:42 UTC |
97c8dc9 | APIPE Tester | 12 July 2022, 14:44:13 UTC | Updated `genome` tab completion. | 12 July 2022, 14:44:13 UTC |
d425286 | APIPE Tester | 12 July 2022, 14:43:10 UTC | Updated class browser cache. | 12 July 2022, 14:43:10 UTC |
9ed6def | John Garza | 30 May 2022, 07:01:25 UTC | Add multithreading and suppress output | 30 May 2022, 07:01:25 UTC |
e97262e | John Garza | 30 April 2022, 23:20:41 UTC | Merge pull request #2059 from johnegarza/proc_prof_queueing Changes to allow adding PPs from compute1 | 30 April 2022, 23:20:41 UTC |
f021a29 | Thomas B. Mooney | 27 April 2022, 21:04:12 UTC | Merge pull request #2062 from genome/chrisamiller-patch-3 fixing pull_outputs command for cloud workflows | 27 April 2022, 21:04:12 UTC |
44acf63 | Thomas B. Mooney | 27 April 2022, 21:03:55 UTC | Merge pull request #2061 from genome/chrisamiller-patch-2 updating cloudize-workflow image | 27 April 2022, 21:03:55 UTC |
3bb6c5d | Chris Miller | 16 April 2022, 02:59:46 UTC | fixing pull_outputs command for cloud workflows Removing deep option, as it no longer exists in the pull output script | 16 April 2022, 02:59:46 UTC |
53350ae | Chris Miller | 16 April 2022, 02:51:36 UTC | updating cloudize-workflow image updating to 1.2.2 | 16 April 2022, 02:51:36 UTC |
2ad5bb2 | John Garza | 13 April 2022, 19:08:25 UTC | Send user uploaded pp to cloud queue instead of installing direct to filesystem | 13 April 2022, 19:08:25 UTC |
ad02fb6 | John Maruska | 05 April 2022, 16:49:53 UTC | Merge pull request #2058 from johnmaruska/cromwell-gcp-uses-mgibio-image Use mgibio image for cloudize-workflows in cromwell_gcp workflows | 05 April 2022, 16:49:53 UTC |
a6613fc | John Maruska | 05 April 2022, 16:00:06 UTC | Use mgibio image for cloudize-workflows in cromwell_gcp workflows | 05 April 2022, 16:00:06 UTC |
11f99b6 | Thomas B. Mooney | 09 March 2022, 14:58:52 UTC | Merge pull request #2057 from tmooney/import_data_from_lims_return_code Only return true if we succeeded to process all instrument data. | 09 March 2022, 14:58:52 UTC |
a7ec2e0 | Thomas B. Mooney | 08 March 2022, 20:35:51 UTC | Only return true if we succeeded to process all instrument data. | 08 March 2022, 20:35:51 UTC |
2393868 | Thomas B. Mooney | 21 February 2022, 21:40:24 UTC | Merge pull request #2056 from tmooney/requeue_command_enumerates_options Fix enumeration of valid values for --action parameter. | 21 February 2022, 21:40:24 UTC |
49e361e | Thomas B. Mooney | 16 February 2022, 21:19:24 UTC | Fix enumeration of valid values for --action parameter. They're actually defined on the model class and not the derived build class. Also, a little explanation for "0" seems helpful! | 16 February 2022, 21:19:24 UTC |
f2a49d3 | Thomas B. Mooney | 03 February 2022, 21:23:16 UTC | Merge pull request #2055 from tmooney/update_tab_completion_and_class_browser_cache Update repo version of class browser cache and tab-completion spec. | 03 February 2022, 21:23:16 UTC |
7faeec2 | Thomas B. Mooney | 03 February 2022, 20:22:05 UTC | Update repo version of class browser cache and tab-completion spec. We already update these routinely as part of the snapshot process, but the job that was keeping the repo updated has been disabled for awhile. This is just prep for turning it back on. | 03 February 2022, 20:22:05 UTC |
152a147 | Thomas B. Mooney | 02 February 2022, 23:14:23 UTC | Merge pull request #2054 from tmooney/different_name_for_prodbuilder Rename "prod-builder" env option to "automated". | 02 February 2022, 23:14:23 UTC |
24ca45e | Thomas B. Mooney | 26 January 2022, 22:06:39 UTC | Rename "prod-builder" env option to "automated". Since "prod-builder" isn't really descriptive and is also the name of the default runner user, it was suggested we use a neutral term instead, especially for those circumstances where a different runner is being used. | 26 January 2022, 22:06:39 UTC |
ec3edd5 | Thomas B. Mooney | 26 January 2022, 22:02:15 UTC | Merge pull request #2053 from tmooney/make_abandonpriorstatus_test_consistent_with_command Don't use "last_complete_build" in this test. | 26 January 2022, 22:02:15 UTC |
ff09e31 | Thomas B. Mooney | 26 January 2022, 21:18:30 UTC | Don't use "last_complete_build" in this test. Our command uses the creation date to determine which build is most recent for purposes of abandoning prior attempts. However if an older build completed after a later build, then our test would be grabbing the wrong one for being "saved". (Instead let's keep tabs on the last one we made, since it should be the most-recently-created and thus the one to be saved. | 26 January 2022, 21:39:43 UTC |
f654dc0 | Thomas B. Mooney | 26 January 2022, 19:12:37 UTC | Merge pull request #2052 from tmooney/tolerate_multiple_succeeded_cromwell_workflows If there are multiple Cromwell workflows, use the most recent. | 26 January 2022, 19:12:37 UTC |
97c6d99 | Thomas B. Mooney | 26 January 2022, 18:01:51 UTC | If there are multiple Cromwell workflows, use the most recent. If a build fails after the Cromwell workflow has succeeded, then a build `restart` may create a new workflow in which case there may be multiple succeeded workflows. Instead of crashing, we'll trust that the most recent one is the one we care about. | 26 January 2022, 18:01:51 UTC |
7066a6a | Thomas B. Mooney | 25 January 2022, 21:41:22 UTC | Merge pull request #2051 from tmooney/abandon_prior_builds_test_unique_timestamps For test stability, have each build created at a different time. | 25 January 2022, 21:41:22 UTC |
99681a6 | Thomas B. Mooney | 25 January 2022, 20:25:46 UTC | For test stability, have each build created at a different time. It is true that some of our commands will fail if multiple builds were created within the same second. There's locking in place that prevents that from happening in practice, so this test was failing for an edge condition we don't normall encounter. (If we truly have multiple simultaneous builds in the same status, picking one arbitrarily to execute is the "right" behavior here, anyway.) | 25 January 2022, 20:25:46 UTC |
dc4c22f | John Maruska | 25 January 2022, 17:47:01 UTC | Merge pull request #2048 from johnmaruska/gcp-separate-buckets GCP subnet | 25 January 2022, 17:47:01 UTC |
70c0680 | Thomas B. Mooney | 24 January 2022, 22:43:31 UTC | Merge pull request #2050 from tmooney/dont_worry_about_allocation_size_for_cwl_run_test Don't worry about if we have space for allocations in this test. | 24 January 2022, 22:43:31 UTC |
063c375 | Thomas B. Mooney | 24 January 2022, 22:24:39 UTC | Don't worry about if we have space for allocations in this test. Currently CWL scratch directories reserve 50GB, but the test uses local temp and doesn't require nigh that much space anyway! | 24 January 2022, 22:26:30 UTC |
b09344e | Thomas B. Mooney | 24 January 2022, 20:49:29 UTC | Merge pull request #2049 from tmooney/tiebreaker_for_model_input_sort Add a tiebreaker in the sorting algorithm. | 24 January 2022, 20:49:29 UTC |
cadad49 | Thomas B. Mooney | 24 January 2022, 20:25:37 UTC | Add a tiebreaker in the sorting algorithm. If perchance both items have identical timestamps, use ID as a tiebreaker so the ordering is still consistent between them. | 24 January 2022, 20:25:37 UTC |
dfa677e | Maruska | 21 January 2022, 17:43:43 UTC | Bump tag version to reverted cloud-workflows | 21 January 2022, 17:43:43 UTC |
8910335 | Maruska | 21 January 2022, 17:28:35 UTC | How'd this get here? | 21 January 2022, 17:28:35 UTC |
b15735b | Maruska | 21 January 2022, 17:27:11 UTC | Revert separate buckets, we decided against this | 21 January 2022, 17:27:11 UTC |
cb115b2 | Thomas B. Mooney | 20 January 2022, 21:45:39 UTC | Merge pull request #2047 from tmooney/straggling_test_fixes Some straggling test fixes | 20 January 2022, 21:45:39 UTC |
e550e63 | Thomas B. Mooney | 20 January 2022, 18:13:23 UTC | Disable MethRatio test. Version 2.6 was in someone's homedir and we don't really use this any more, so it has been decided to disable this test rather than switch it to the other version. | 20 January 2022, 18:13:23 UTC |
0f9d48e | John Maruska | 20 January 2022, 16:32:55 UTC | Merge branch 'master' of https://github.com/genome/genome into gcp-separate-buckets | 20 January 2022, 16:32:55 UTC |
cee852a | John Maruska | 20 January 2022, 16:31:23 UTC | Bump cloudize-workflow tag | 20 January 2022, 16:31:23 UTC |
ee9d5a9 | John Maruska | 20 January 2022, 16:21:43 UTC | Merge branch 'gcp-specify-subnet' into gcp-separate-buckets | 20 January 2022, 16:21:43 UTC |
a1b7570 | Thomas B. Mooney | 20 January 2022, 05:01:09 UTC | Check the directory we really plan to write to. No need to have write access to the input directory if they gave us an output one. | 20 January 2022, 05:01:09 UTC |
b245bae | John Maruska | 19 January 2022, 23:07:39 UTC | Split cromwell_gcp_bucket into inputs and executions buckets This is related to work in griffith-lab/cloud-workflows repo. The intention behind splitting buckets is to apply separate lifecycle rules so input files stick around but execution files are deleted after 30 days | 19 January 2022, 23:07:39 UTC |
2319304 | Thomas B. Mooney | 18 January 2022, 23:14:43 UTC | Merge pull request #2046 from tmooney/bam_size_for_perlane_tests Claim that .t test per-lane BAMs are all 1024KB. | 18 January 2022, 23:14:43 UTC |
217eed7 | Thomas B. Mooney | 18 January 2022, 21:41:50 UTC | Claim that .t test per-lane BAMs are all 1024KB. The tests try to allocate space when revivifying the per-lane BAMs. When the bam size is unknown, the default size request is 100,000,000KB. When temp on the host running the test has less than this amount of space free, the test can fail. (If temp has less than 1024KB free, we're probably going to have other trouble, anyway!) | 18 January 2022, 21:41:50 UTC |
85c6150 | Thomas B. Mooney | 18 January 2022, 20:04:47 UTC | Merge pull request #2045 from tmooney/no_hardcoding_java Use the Genome::Sys method to get the java executable path. | 18 January 2022, 20:04:47 UTC |
a85d7c7 | Thomas B. Mooney | 18 January 2022, 18:33:45 UTC | Use the Genome::Sys method to get the java executable path. There is no `/gapp` on the new cluster, so use our config instead. | 18 January 2022, 18:33:45 UTC |
8e0e55c | Thomas B. Mooney | 18 January 2022, 15:24:58 UTC | Merge pull request #2043 from tmooney/filetest_fixes Consistently use Genome::Sys to test readability of files. | 18 January 2022, 15:24:58 UTC |
a5255b6 | Thomas B. Mooney | 18 January 2022, 15:24:34 UTC | Merge pull request #2044 from tmooney/no_blade_name Revise check for matching hosts. | 18 January 2022, 15:24:34 UTC |
1324826 | Thomas B. Mooney | 13 January 2022, 20:49:47 UTC | Revise check for matching hosts. We don't care if they're "blades". In fact, we only care that the command tries to return any results at all, so check for the header row! When there aren't any hosts, it'll return this instead: Not enough host(s) currently eligible So looking for the line to start with HOST is good enough. | 13 January 2022, 20:49:47 UTC |
d227aee | Thomas B. Mooney | 13 January 2022, 20:14:27 UTC | Consistently use Genome::Sys to test readability of files. To handle non-POSIX permissions we need to not rely on the default behavior of -r. | 13 January 2022, 20:14:27 UTC |
247509a | Thomas B. Mooney | 12 January 2022, 22:49:59 UTC | Merge pull request #2042 from tmooney/absolute_url_for_db_server Use fully qualified domain name for database server. | 12 January 2022, 22:49:59 UTC |
3c3ad2c | Thomas B. Mooney | 12 January 2022, 20:11:09 UTC | Use fully qualified domain name for database server. Some hosts don't have a friendly search path configured, so this is a little more certain to work as expected. | 12 January 2022, 20:11:09 UTC |
c355ce5 | Thomas B. Mooney | 11 January 2022, 22:23:04 UTC | Merge pull request #2041 from tmooney/only_set_tmp_basedir_if_writable Only modify tmp basedir if new location is writable. | 11 January 2022, 22:23:04 UTC |
7127671 | Thomas B. Mooney | 11 January 2022, 20:03:51 UTC | Only modify tmp basedir if new location is writable. For some reason, some servers we're using create a doubly-nested LSF tmpdir structure, and the inner one is r/w only by root. In this case they're already giving us the first level as $ENV{TMPDIR} so we can just trust it anyway. This code is still useful on our older cluster, though, so not just removing it outright yet. | 11 January 2022, 20:03:51 UTC |
1dee886 | John Maruska | 10 January 2022, 22:57:51 UTC | Use docker image that allows subnet specification | 10 January 2022, 22:57:51 UTC |
8960a82 | John Maruska | 10 January 2022, 19:32:39 UTC | Add cromwell_gcp_subnet which propagates to gms/start.sh | 10 January 2022, 19:32:39 UTC |
a94fc2a | John Maruska | 06 January 2022, 22:04:27 UTC | Merge pull request #2040 from johnmaruska/incorrect-quoting-in-workflow-options Misplaced quote in workflow_options heredoc, fails json parse | 06 January 2022, 22:04:27 UTC |
1dfd431 | John Maruska | 06 January 2022, 21:13:36 UTC | Misplaced quote in workflow_options heredoc, fails json parse | 06 January 2022, 21:13:36 UTC |
9651bb1 | John Maruska | 06 January 2022, 17:09:46 UTC | Merge pull request #2039 from johnmaruska/gcp-project-passed-to-start Propagate cromwell_gcp_project setting to gms/start.sh script | 06 January 2022, 17:09:46 UTC |
7eabe68 | John Maruska | 06 January 2022, 16:44:23 UTC | Propagate cromwell_gcp_project setting to gms/start.sh script | 06 January 2022, 16:44:23 UTC |
f63af58 | Thomas B. Mooney | 22 December 2021, 21:57:26 UTC | Merge pull request #2038 from tmooney/build_queued_models Commands to back BQM and RQB on the new CI server | 22 December 2021, 21:57:26 UTC |
197b09d | Thomas B. Mooney | 22 December 2021, 20:42:16 UTC | Merge pull request #2037 from tmooney/interactive_queue_config Add config for specifying the queue for interactive jobs. | 22 December 2021, 20:42:16 UTC |
9e96fb4 | Thomas B. Mooney | 21 December 2021, 22:16:05 UTC | Don't propagate LSF_DOCKER_NETWORK to builds. If a process needs to expose a port it can still do so via other means, but in general we don't want to have collisions of local processes when more than one lands on the same execution host. | 21 December 2021, 22:16:05 UTC |
62ab163 | Thomas B. Mooney | 21 December 2021, 18:58:36 UTC | Add config for specifying the queue for interactive jobs. | 21 December 2021, 18:58:36 UTC |
6544230 | Thomas B. Mooney | 20 December 2021, 20:40:50 UTC | Merge pull request #2036 from tmooney/toil_take_two Updates to support per-job containers when running with toil. | 20 December 2021, 20:40:50 UTC |
2690b17 | Thomas B. Mooney | 20 December 2021, 18:36:12 UTC | Add rebuild-queued-builds command for automated restarts. | 20 December 2021, 18:39:14 UTC |
682306d | Thomas B. Mooney | 03 November 2021, 14:29:29 UTC | Updates to support per-job containers when running with toil. The container must be capable of running python3 with glibc version of at least 2.16. | 14 December 2021, 19:09:33 UTC |
d47074a | John Maruska | 14 December 2021, 17:26:29 UTC | Merge pull request #2035 from johnmaruska/run-cromwell-gcp-fully-cloud Switch cromwell_gcp workflow runner from central Cromwell server to per-run Cromwell VM | 14 December 2021, 17:26:29 UTC |
de420d9 | Maruska | 13 December 2021, 17:19:56 UTC | Use status message instead of debug | 13 December 2021, 17:19:56 UTC |
5377214 | Maruska | 10 December 2021, 16:57:25 UTC | Download outputs using nested structure that's cromwell's default | 10 December 2021, 16:57:25 UTC |
310e7ad | Maruska | 10 December 2021, 16:53:34 UTC | not sure how these happened | 10 December 2021, 16:53:34 UTC |