Github user vanzin commented on a diff in the pull request:

    https://github.com/apache/spark/pull/2471#discussion_r18737534
  
    --- Diff: 
core/src/main/scala/org/apache/spark/deploy/history/FsHistoryProvider.scala ---
    @@ -195,22 +241,68 @@ private[history] class FsHistoryProvider(conf: 
SparkConf) extends ApplicationHis
               }
             }
     
    -        val newIterator = logInfos.iterator.buffered
    -        val oldIterator = applications.values.iterator.buffered
    -        while (newIterator.hasNext && oldIterator.hasNext) {
    -          if (newIterator.head.endTime > oldIterator.head.endTime) {
    -            addIfAbsent(newIterator.next)
    -          } else {
    -            addIfAbsent(oldIterator.next)
    +        applications.synchronized {
    +          val newIterator = logInfos.iterator.buffered
    +          val oldIterator = applications.values.iterator.buffered
    +          while (newIterator.hasNext && oldIterator.hasNext) {
    +            if (newIterator.head.endTime > oldIterator.head.endTime) {
    +              addIfAbsent(newIterator.next)
    +            } else {
    +              addIfAbsent(oldIterator.next)
    +            }
               }
    +          newIterator.foreach(addIfAbsent)
    +          oldIterator.foreach(addIfAbsent)
    +
    +          applications = newApps
             }
    -        newIterator.foreach(addIfAbsent)
    -        oldIterator.foreach(addIfAbsent)
    +      }
    +    } catch {
    +      case t: Throwable => logError("Exception in checking for event log 
updates", t)
    +    }
    +  }
    +
    +  /**
    +   *  Deleting apps if setting cleaner.
    +   */
    +  private def cleanLogs() = {
    +    lastLogCleanTimeMs = getMonotonicTimeMs()
    +    logDebug("Cleaning logs. Time is now %d.".format(lastLogCleanTimeMs))
    +    try {
    +      val logStatus = fs.listStatus(new Path(resolvedLogDir))
    +      val logDirs = if (logStatus != null) logStatus.filter(_.isDir).toSeq 
else Seq[FileStatus]()
    +      val maxAge = conf.getLong("spark.history.fs.maxAge.seconds",
    +        DEFAULT_SPARK_HISTORY_FS_MAXAGE_S) * 1000
    +
    +      val now = System.currentTimeMillis()
    +      fs.synchronized {
    +        // scan all logs from the log directory.
    +        // Only directories older than this many seconds will be deleted .
    +        logDirs.foreach { dir =>
    +          // history file older than this many seconds will be deleted 
    +          // when the history cleaner runs.
    +          if (now - getModificationTime(dir) > maxAge) {
    +            fs.delete(dir.getPath, true)
    +          }
    +        }
    +      }
    +      
    +      val newApps = new mutable.LinkedHashMap[String, 
FsApplicationHistoryInfo]()
    +      def addIfNotExpire(info: FsApplicationHistoryInfo) = {
    +        if(now - info.lastUpdated <= maxAge) {
    +          newApps += (info.id -> info)
    --- End diff --
    
    So, there's the potential that `info.lastUpdated` may be out of sync with 
the file system. This could happen if someone messes with the timestamps of the 
files, for example, so that `info.lastUpdated` is recent but the files in the 
directory were actually cleaned up by the code above.
    
    In that case you'd have stale entries in the applications list and the user 
would see an error when clicking on the link.
    
    That's not a huge issue, since it's not the only way that can happen (as 
with all cases where you're polling things). I was thinking about some way to 
make `checkForLogs` deal with that instead, but not sure if that would be any 
simpler.


---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org

Reply via email to