c_pgms/arequal: Add arequal tool to the git

Adding arequal to the git, which can be used to calculate the checksum
of path given to it as an argument

Change-Id: Iefd283511e6ff2f3bc1359e718d1922506d7a24a
Signed-off-by: Raghavendra Bhat <raghavendrabhat@gluster.com>

1 files changed, 77 insertions, 0 deletions

diff --git a/arequal/README b/arequal/README
new file mode 100644
index 0000000..8c9d952
--- /dev/null
+++ b/arequal/README
@@ -0,0 +1,77 @@
+arequal - Tool to test data security of GlusterFS
+=======
+
+0. Install arequal
+------------------
+  sh# ./autogen.sh
+  sh# ./configure
+  sh# make
+  sh# make install
+
+  Perform the above installation steps on both clients and servers
+
+
+1. Generate data set
+--------------------
+  Create a data set to be used for the test. This data set should
+have file sizes and file count similar to the data set to be used
+in production. You could also use existing data (like /usr) as
+your dataset as it will not be modified. This document will use
+/usr as the example source directory.
+
+
+2. Mount GlusterFS
+------------------
+  Install, configure and start glusterfs servers and client. If
+the Replicate module is loaded, this tool can be used to perform
+data consistency check among the replicas. This document will
+use /mnt/glusterfs as the example mount point.
+
+
+3. Start the test
+-----------------
+  sh# arequal-run.sh /usr/ /mnt/gluster/usr
+
+
+4. Verify the output
+--------------------
+   The tool outputs two sets of checksums one after another on
+the standard output. Verify that all the values match against
+each other. This ensures that the data has been copied over
+properly into the GlusterFS mountpoint.
+
+
+5. Extensive Replicate testing
+------------------------------
+   The rest of the document is for testing the high availability
+and healing features of Replicate.
+
+
+6. High availability testing
+----------------------------
+   Restart step 3. While the script is in progress, kill one of the
+servers. Let the script continue to completion. The script should
+not fail because of one of the server getting killed. The checksums
+should still match.
+
+
+7. Consistency testing
+----------------------
+  After step 3, run the following command on both the servers
+
+  sh# arequal-checksum /export/directory
+
+  The output values should match
+
+
+8. Recovery testing
+-------------------
+   If step 7 is performed after step 6, the output values will not
+match since changes performed when one of the servers was down has
+not propagated to the backend.
+
+   Bring back the server up again. On the same mountpoint, run an
+ls -lR to force an access to all the files on the files involved.
+
+   Now calculate the checksums on both the backends as described
+in step 7. The output values should match.