Do backups asynchronously #220
Conversation
|
Since this is run on a timer, do we run the risk of backing up the results of a partially finished |
Yeah, we could add exclusions but...eh, doesn't seem worth it. At least for rdgo though we need to teach it to make datestamped-dirs that will sync more naturally. |
Jenkinsfile.backup
Outdated
| string(credentialsId: params.ARTIFACT_SERVER, variable: 'ARTIFACT_SERVER'), | ||
| sshUserPrivateKey(credentialsId: params.ARTIFACT_SSH_CREDS_ID, keyFileVariable: 'KEY_FILE'), | ||
| ]) { | ||
| sh "mkdir -p ${target}" |
There was a problem hiding this comment.
How about keeping this under $WORKSPACE and use the rsync_dir_in_dest() from #211? I really like the move away from /srv.
Jenkinsfile.backup
Outdated
| for (dir in dirs_to_backup) { | ||
| // Expand, but also avoid iteration+closure trap | ||
| def target = "/srv/rhcos/output/${dir}"; | ||
| par_stages["backup-" + target] = { -> stage("Backup " + target) { |
There was a problem hiding this comment.
Let's use (the equivalent of) $(basename $target) for the stage name?
|
It might be as simple as wrapping the "Sync in" here and the "Sync out" in the cloud/rdgo jobs with |
openshift-ci-robot
added
the
size/M
This takes us to a single point of truth for data; the other is clearly secondary.
|
Updated, but not tested (jenkins lost all nodes). |
|
/lgtm |
|
/retest |
This takes us to a single point of truth for data; the other is
clearly secondary.