This loop is far too static.
The likelihood of it failing increases with increasing size or after merging content from multiple Confluence instances.
It may not make sense to keep adjusting the retries manually.
|
for (( c=1; c<=${progress_retries}; c++ )); do |
|
log " * waiting for atlassian ${APP} cloud to finish backup..." |
|
PROGRESS_JSON=$(curl -sSfL -u ${EMAIL}:${API_TOKEN} https://${INSTANCE}/wiki/rest/obm/1.0/getprogress.json || die " * Error while requesting backup generation progress, exiting.") |
|
FILE_NAME=$(echo "$PROGRESS_JSON" | sed -n 's/.*"fileName"[ ]*:[ ]*"\([^"]*\).*/\1/p') |
|
|
|
## PRINT BACKUP STATUS INFO ## |
|
# log "$PROGRESS_JSON" |
|
|
|
if [[ ${PROGRESS_JSON} == *"error"* ]]; then |
|
break |
|
fi |
|
|
|
if [[ ! -z "${FILE_NAME}" ]]; then |
|
break |
|
fi |
|
|
|
# Waits for the amount of seconds specified in progress_sleep variable between a check and the other |
|
sleep ${progress_sleep} |
|
done |
An improvement would be the implementation of a dynamic adjustment of the backoff depending on the progress.
