[
https://issues.apache.org/jira/browse/HDFS-10713?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15468265#comment-15468265
]
Arpit Agarwal edited comment on HDFS-10713 at 9/6/16 8:55 PM:
--------------------------------------------------------------
Thanks for the updated patch [~hanishakoneru]. The read lock is not exclusive
so the following sequence of operations in readUnlock won't see a consistent
state. This block of code probably needs to synchronize on a separate mutex
after releasing the read lock.
{code}
if (needReport && readLockInterval >= this.readLockReportingThreshold) {
if (readLockInterval > longestReadLockHeldInterval) {
longestReadLockHeldInterval = readLockInterval;
}
if (currentTime - timeStampOfLastReadLockReport > this
.lockSuppressWarningInterval) {
logReport = true;
numSuppressedWarnings = numReadLockWarningsSuppressed;
numReadLockWarningsSuppressed = 0;
longestLockHeldInterval = longestReadLockHeldInterval;
longestReadLockHeldInterval = 0;
timeStampOfLastReadLockReport = currentTime;
} else {
numReadLockWarningsSuppressed++;
}
}
this.fsLock.readLock().unlock();
{code}
-I also had some questions about the _needReport_ logic I posted on
HDFS-10817.-
was (Author: arpitagarwal):
Thanks for the updated patch [~hanishakoneru]. The read lock is not exclusive
so the following sequence of operations in readUnlock won't see a consistent
state. This block of code probably needs to synchronize on a separate mutex
after releasing the read lock.
{code}
if (needReport && readLockInterval >= this.readLockReportingThreshold) {
if (readLockInterval > longestReadLockHeldInterval) {
longestReadLockHeldInterval = readLockInterval;
}
if (currentTime - timeStampOfLastReadLockReport > this
.lockSuppressWarningInterval) {
logReport = true;
numSuppressedWarnings = numReadLockWarningsSuppressed;
numReadLockWarningsSuppressed = 0;
longestLockHeldInterval = longestReadLockHeldInterval;
longestReadLockHeldInterval = 0;
timeStampOfLastReadLockReport = currentTime;
} else {
numReadLockWarningsSuppressed++;
}
}
this.fsLock.readLock().unlock();
{code}
I also had some questions about the _needReport_ logic I posted on HDFS-10817.
> Throttle FsNameSystem lock warnings
> -----------------------------------
>
> Key: HDFS-10713
> URL: https://issues.apache.org/jira/browse/HDFS-10713
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: logging, namenode
> Reporter: Arpit Agarwal
> Assignee: Hanisha Koneru
> Attachments: HDFS-10713.000.patch, HDFS-10713.001.patch,
> HDFS-10713.002.patch, HDFS-10713.003.patch
>
>
> The NameNode logs a message if the FSNamesystem write lock is held by a
> thread for over 1 second. These messages can be throttled to at one most one
> per x minutes to avoid potentially filling up NN logs. We can also log the
> number of suppressed notices since the last log message.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
