[#HADOOP-2373] Name node silently changes state

Hadoop Common

Name node silently changes state

Created: 07/Dec/07 02:58 AM Updated: 08/Jul/09 04:42 PM

Return to search

Component/s:

None

Affects Version/s:

0.16.0

Fix Version/s:

0.17.0

Time Tracking:

Not Specified

File Attachments:

File Name	Date Attached	Attached By	Size
SafeModeReport.patch	2008-03-18 12:02 AM	Konstantin Shvachko	6 kB
SafeModeReport1.patch	2008-03-19 07:06 PM	Konstantin Shvachko	6 kB

Issue Links:

Reference

This issue relates to:

~~HADOOP-2159~~ Namenode stuck in safemode

Resolution Date:

20/Mar/08 05:30 PM

Description

« Hide

1. The name node should create a log message when entering the "extension" period of safe mode after achieving the minimal replication threshold.

Logging state changes is a Good Idea. When diagnosing 2159, there is no evidence whether the name node is in the extension period or not.

2. When in safe mode, the name node should periodically explain why it is still in safe mode. (2159 again)

All

Comments

Work Log

Change History

Subversion Commits

Sort Order:

[ Permlink | « Hide ]

Konstantin Shvachko added a comment - 18/Mar/08 12:02 AM

This patch provides better reporting of the safe-mode state.

It prints what is the current ratio of the reported blocks.
It reports when safe mode enters the extended wait period and says how many seconds remained till the name-node will leave it.
It logs the name-node state every 20 seconds.

[ Permlink | « Hide ]

Hadoop QA added a comment - 18/Mar/08 06:18 PM

-1 overall. Here are the results of testing the latest attachment
http://issues.apache.org/jira/secure/attachment/12378090/SafeModeReport.patch
against trunk revision 619744.

@author +1. The patch does not contain any @author tags.

tests included -1. The patch doesn't appear to include any new or modified tests.
Please justify why no tests are needed for this patch.

javadoc +1. The javadoc tool did not generate any warning messages.

javac +1. The applied patch does not generate any new javac compiler warnings.

release audit +1. The applied patch does not generate any new release audit warnings.

findbugs +1. The patch does not introduce any new Findbugs warnings.

core tests +1. The patch passed core unit tests.

contrib tests +1. The patch passed contrib unit tests.

Test results: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch/1986/testReport/
Findbugs warnings: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch/1986/artifact/trunk/build/test/findbugs/newPatchFindbugsWarnings.html
Checkstyle results: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch/1986/artifact/trunk/build/test/checkstyle-errors.html
Console output: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch/1986/console

This message is automatically generated.

[ Permlink | « Hide ]

Tsz Wo (Nicholas), SZE added a comment - 18/Mar/08 06:36 PM

reportStatus(String msg, boolean rightNow) need to be synchronized to protect lastStatusReport.

timeInSafemode/1000 => String.format("%0.2d", timeInSafemode/1000.0).

Similarly, use String.format(...) for other floating point output in getTurnOffTip().

[ Permlink | « Hide ]

Konstantin Shvachko added a comment - 19/Mar/08 07:06 PM

> reportStatus(String msg, boolean rightNow) need to be synchronized to protect lastStatusReport.
reportStatus() is called by methods that are synchronized, so there is no need to synchronize itself.

> timeInSafemode/1000 => String.format("%0.2d", timeInSafemode/1000.0).
This prints whole number of seconds, I do not think we want microseconds here.

> Similarly, use String.format(...) for other floating point output in getTurnOffTip().
I did that. The threshold is printed up to the fourth digit.
Two digits is probably not enough since the default has three: 0.999f