source: proiecte/HadoopJUnit/hadoop-0.20.1/src/test/org/apache/hadoop/hdfs/hadoop-dfs-dir.txt @ 120

Last change on this file since 120 was 120, checked in by (none), 14 years ago

Added the mail files for the Hadoop JUNit Project
  • Property svn:executable set to *
File size: 3.0 KB
Line 
1#
2# This is a readme for hadoop-version-dfs-dir.tgz and hadoop-dfs-dir.txt.
3#
4# See HADOOP-1629 for more info if needed.
5# These two files are used by unit test TestDFSUpgradeFromImage.java
6#
7# hadoop-14-dfs-dir.tgz :
8# ---------------------
9# This file contains the HDFS directory structure for one namenode and 4 datanodes.
10# The structure is setup similar to the structure used in MiniDFSCluster.
11# The directory was created with Hadoo-0.14.x.
12#
13# In the test, this directory is unpacked and MiniDFSCluster is run with
14# "-upgrade" option. The test waits for the upgrade to complete
15# (leave safe mode) and then all the files are read. The test checks that the
16# directory structure and file checksums exactly match the information
17# in this file.
18#
19# hadoop-dfs-dir.txt :
20# ---------------------
21# Along with this description this file contains the expected files and
22# checksums or the files in the upgraded DFS.
23#
24# The original DFS directory was created with various types of files and with
25# some recoverable errors (i.e. corrupt or missing .crc files).
26#
27# A similar set of files exist in two different DFS directories.
28# For e.g. "top-dir-1Mb-512" contains files created with dfs.block.size of 1Mb
29# and io.bytes.per.checksum of 512.
30#
31# In the future, when Hadoop project no longer supports upgrade from
32# Hadoop-0.12, then a new DFS directory image must be created.
33#
34# To generate checksum info for new files :
35# ---------------------------------------
36# Uncomment the last coment (starts with "printChecksums") and run the
37# test again. When the test sees this line, it prints the checksum
38# information that should replace the checksum information in
39# this file. When run in this mode, the test will fail with a descriptive IOException.
40#
41# Next, extract the checksum info from the test log like this:
42#  sed -n 's/.*CRC info for reference file : //p' test-log.txt >> this_file
43# This will append a new list of files and checksums to this file.  Be sure to remove the existing checksum info.
44#
45# For your reference, the format of the checksum info below is "filename whitespace*\twhitespace* checksum\n"
46#
47# Uncomment the following line to produce checksum info for a new DFS image.
48#printChecksums
49
50/1kb-multiple-checksum-blocks-64-16      191893480
51/top-dir-120000-60/1Mb-file      4079112547
52/top-dir-120000-60/4k-file       3716287280
53/top-dir-120000-60/5Mb-file      2563834633
54/top-dir-120000-60/directory1/500thousand-file   3036538664
55/top-dir-120000-60/directory1/file-with-corrupt-crc      1984689737
56/top-dir-120000-60/directory1/file-with-no-crc   4004594475
57/top-dir-120000-60/directory1/zero1      0
58/top-dir-120000-60/zerolen       0
59/top-dir-1Mb-512/1Mb-file        4079112547
60/top-dir-1Mb-512/4k-file         3716287280
61/top-dir-1Mb-512/5Mb-file        2563834633
62/top-dir-1Mb-512/directory1/500thousand-file     3036538664
63/top-dir-1Mb-512/directory1/file-with-corrupt-crc        1984689737
64/top-dir-1Mb-512/directory1/file-with-no-crc     4004594475
65/top-dir-1Mb-512/directory1/zero1        0
66/top-dir-1Mb-512/zerolen         0
67overallCRC       1419480698
Note: See TracBrowser for help on using the repository browser.