damccorm commented on code in PR #24770:
URL: https://github.com/apache/beam/pull/24770#discussion_r1059154483
##########
.test-infra/tools/stale_bq_datasets_cleaner.sh:
##########
@@ -33,11 +34,13 @@ for dataset in ${BQ_DATASETS[@]}; do
if [[ $dataset =~ $template ]]; then
# The BQ API reports LAST MODIFIED TIME in miliseconds, while unix works
in seconds since epoch
# thus why we need to convert to seconds.
- LAST_MODIFIED_MS=`bq --format=json --project_id=$PROJECT show $dataset |
jq -r .lastModifiedTime`
+ [[ `bq --format=json --project_id=$PROJECT show $dataset` =~
\"lastModifiedTime\":\"([0-9]+)\" ]]
+ LAST_MODIFIED_MS=${BASH_REMATCH[1]}
LAST_MODIFIED=$(($LAST_MODIFIED_MS / 1000))
if [[ $GRACE_PERIOD -gt $LAST_MODIFIED ]]; then
- echo "Deleting $dataset (modified `date -d @$LAST_MODIFIED`) Command
bq --project_id=$PROJECT rm -r -f $dataset"
- bq --project_id=$PROJECT rm -r -f $dataset
+ echo "Deleting $dataset (modified `date -d @$LAST_MODIFIED`)"
+ # do not fail the script if delete dataset fail
+ bq --project_id=$PROJECT rm -r -f $dataset || true
Review Comment:
Yep, that's fair - I more was wondering if we could update the existing
service account to have these permissions (probably neither you nor me have the
correct permission set to do so though).
In the short term, my vote would be to delete the bad datasets manually and
then merge the PR without the `|| true`, or even better with some logic that
does as much cleanup as it can and then fails the job if any of the cleanups
failed. Longer term, if this is a persistent problem we could update the
service account to handle this correctly.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
