Cromwell PR #4000 attempts to log and retry a number of issues.
Even as the PR patches other issues, the following have repeatedly popped up in Travis the second week of August 2018.
-- regular delete from union doesn't work | |
delete `METADATA_ENTRY` | |
from `METADATA_ENTRY` | |
where (`METADATA_ENTRY`.`WORKFLOW_EXECUTION_UUID` in | |
((select `WORKFLOW_EXECUTION_UUID` as x2 | |
from `WORKFLOW_METADATA_SUMMARY_ENTRY` | |
where `ROOT_WORKFLOW_EXECUTION_UUID` = 'aaaa') | |
union | |
(select 'aaaa' as x2))) | |
and (`METADATA_ENTRY`.`METADATA_KEY` like 'labels:%'); |
Cromwell PR #4000 attempts to log and retry a number of issues.
Even as the PR patches other issues, the following have repeatedly popped up in Travis the second week of August 2018.
#!/usr/bin/env bash | |
for container in $(gcloud container images list --repository gcr.io/google-containers --format='table[no-heading](name)' | sed 's#gcr.io/google-containers/##'); do | |
gcloud -q container images list-tags gcr.io/google-containers/${container} --sort-by=~tags --limit=unlimited --filter='tags:*' --format=json | jq -r '.[] | . as $image | .tags[] | . as $tag | [$tag, $image.timestamp.datetime] | @tsv' | while IFS=$'\t' read -r tag date; do | |
date="${date:-null}" | |
date="${date/ /T}" | |
echo ${date}$'\t'${container}$'\t'${tag}$'\t'$(curl -s \ | |
-H 'Accept: application/vnd.docker.distribution.manifest.v2+json' \ | |
-I https://gcr.io/v2/google-containers/${container}/manifests/${tag} \ | |
| grep HTTP/1.1) |
========================= | |
= Setup on any session | |
========================= | |
DELETE FROM FAILURE_EVENT; | |
DELETE FROM RUNTIME_ATTRIBUTES; | |
DELETE FROM EXECUTION_INFO; | |
DELETE FROM SYMBOL; | |
DELETE FROM WORKFLOW_EXECUTION_AUX; | |
DELETE FROM EXECUTION_EVENT; |