Github user vanzin commented on a diff in the pull request:
https://github.com/apache/spark/pull/5491#discussion_r28297121
--- Diff:
core/src/main/scala/org/apache/spark/deploy/history/FsHistoryProvider.scala ---
@@ -273,35 +273,28 @@ private[history] class FsHistoryProvider(conf:
SparkConf) extends ApplicationHis
*/
private def cleanLogs(): Unit = {
try {
- val statusList = Option(fs.listStatus(new Path(logDir))).map(_.toSeq)
- .getOrElse(Seq[FileStatus]())
val maxAge = conf.getLong("spark.history.fs.cleaner.maxAge.seconds",
DEFAULT_SPARK_HISTORY_FS_MAXAGE_S) * 1000
val now = System.currentTimeMillis()
val appsToRetain = new mutable.LinkedHashMap[String,
FsApplicationHistoryInfo]()
+ // Scan all logs from the log directory.
+ // Only completed applications older than the specified max age will
be deleted.
applications.values.foreach { info =>
- if (now - info.lastUpdated <= maxAge) {
+ if (now - info.lastUpdated <= maxAge || !info.completed) {
appsToRetain += (info.id -> info)
+ } else {
+ try {
+ fs.delete(new Path(logDir + "/" + info.logPath), true)
+ } catch {
+ case t: IOException => logError(s"IOException in cleaning logs
of ${info.logPath}", t)
+ appsToRetain += (info.id -> info)
--- End diff --
Shouldn't this line be indented? And the `logError` above me on its own
line, at the same indent level as this one?
If feels a little weird to leave the app on the live list; in the case of
old-style logs, a few files may have been deleted, while others remain, so the
app would be "un-renderable". On the other hand, this means the code will try
to clean up this app again later on, so I guess it's better this way.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
