Scripts to identify quota accouting issues

The patch contains 2 scripts:

log_accounting.sh does a du -h on the FS hierarchy and a quota list
on the hierarchy and interleaves the two output. We can then identify
which directory(s) in FS has caused the accounting to go bad and try
to investigate what fops happened on those directories. We can also
limit the set of directories on which we need to set dirty xattr to
correct accounting.
xattr_analysis.py reads all the xattr of a brick and dumps it a human
readable form to ease debugging.

Change-Id: I2155561d10c08dc3ab9e8b09dbd258f0592b4d33
BUG: 1466188
Signed-off-by: Sanoj Unnikrishnan <sunnikri@redhat.com>
Reviewed-on: https://review.gluster.org/17649
CentOS-regression: Gluster Build System <jenkins@build.gluster.org>
Smoke: Gluster Build System <jenkins@build.gluster.org>
Reviewed-by: Raghavendra G <rgowdapp@redhat.com>

3 files changed, 98 insertions, 0 deletions

diff --git a/extras/contri-add.sh b/extras/quota/contri-add.sh
index 7db5edd5d20..7db5edd5d20 100755
--- a/extras/contri-add.sh
+++ b/extras/quota/contri-add.sh
diff --git a/extras/quota/log_accounting.sh b/extras/quota/log_accounting.sh
new file mode 100755
index 00000000000..9c2381f6a95
--- /dev/null
+++ b/extras/quota/log_accounting.sh
@@ -0,0 +1,26 @@
+#!/usr/bin/bash
+# The script does an accounting of all directories using command 'du' and
+# using gluster. We can then compare the two to identify accounting mismatch
+# THere can be minor mismatch because gluster only accounts for the size of
+# files. Direcotries can take up upto 4kB space on FS per directory. THis
+# size is accounted by du and not by gluster. However the difference would
+# not be significant.
+
+mountpoint=$1
+volname=$2
+
+usage ()
+{
+    echo >&2 "usage: $0 <mountpoint> <volume name>"
+    exit
+}
+
+[ $# -lt 2 ] && usage
+
+cd $mountpoint
+du -h | head -n -1 | tr -d '.' |awk  '{ for (i = 2; i <= NF; i++) { printf("%s ", $i);}  print "" }' > /tmp/gluster_quota_1
+cat /tmp/gluster_quota_1 | sed 's/ $//' | sed 's/ /\\ /g' | sed 's/(/\\(/g' | sed 's/)/\\)/g' |xargs gluster v quota $volname list > /tmp/gluster_quota_2
+du -h | head -n -1 |awk  '{ for (i = 2; i <= NF; i++) { printf("%s %s", $i, $1);}  print "" }' | tr -d '.' >  /tmp/gluster_quota_3
+cat /tmp/gluster_quota_2 /tmp/gluster_quota_3 | sort > /tmp/gluster_quota_4
+find . -type d > /tmp/gluster_quota_5
+tar -cvf /tmp/gluster_quota_files.tar /tmp/gluster_quota_*
diff --git a/extras/quota/xattr_analysis.py b/extras/quota/xattr_analysis.py
new file mode 100755
index 00000000000..d3d1a74170b
--- /dev/null
+++ b/extras/quota/xattr_analysis.py
@@ -0,0 +1,72 @@
+#!/usr/bin/python
+# Below script has to purpose
+#  1. Display xattr of entire FS tree in a human readable form
+#  2. Display all the directory where contri and size mismatch.
+#      (If there are any directory with contri and size mismatch that are not dirty
+#       then that highlights a propogation issue)
+#  The script takes only one input LOG _FILE generated from the command,
+#  find <brick_path> | xargs  getfattr -d -m. -e hex  > log_gluster_xattr
+
+import re
+import subprocess
+import sys
+from hurry.filesize import size
+
+if len(sys.argv) < 2:
+    sys.exit('Usage: %s log_gluster_xattr \n'
+              'to genereate log_gluster_xattr use: \n'
+              'find <brick_path> | xargs  getfattr -d -m. -e hex  > log_gluster_xattr'
+              % sys.argv[0])
+LOG_FILE=sys.argv[1]
+
+def get_quota_xattr_brick():
+    out = subprocess.check_output (["/usr/bin/cat", LOG_FILE])
+    pairs = out.splitlines()
+
+    xdict = {}
+    mismatch_size = [('====contri_size===', '====size====')]
+    for xattr in pairs:
+        k = xattr.split("=")[0]
+        if re.search("# file:",k):
+            print xdict
+            filename=k
+            print "=====" + filename + "======="
+            xdict = {}
+        elif k is "":
+            pass
+        else:
+            print xattr
+            v = xattr.split("=")[1]
+            if re.search("contri",k):
+                if len(v) == 34:
+                    # for files size is obtained in iatt, file count should be 1, dir count=0
+                    xdict['contri_file_count'] = int(v[18:34], 16)
+                    xdict['contri_dir_count'] = 0
+                else:
+                    xdict['contri_size'] = size(int(v[2:18], 16))
+                    xdict['contri_file_count'] = int(v[18:34], 16)
+                    xdict['contri_dir_count'] = int(v[34:], 16)
+            elif re.search("size",k):
+                xdict['size'] = size(int(v[2:18], 16))
+                xdict['file_count'] = int(v[18:34], 16)
+                xdict['dir_count'] = int(v[34:], 16)
+            elif re.search("dirty",k):
+                if v == '0x3000':
+                    xdict['dirty'] = False
+                elif v == '0x3100':
+                    xdict['dirty'] = True
+            elif re.search("limit_objects",k):
+                xdict['limit_objects'] = int(v[2:18], 16)
+            elif re.search("limit_set",k):
+                xdict['limit_set'] = size(int(v[2:18], 16))
+
+            if 'size' in xdict and 'contri_size' in xdict and xdict['size'] != xdict['contri_size']:
+                mismatch_size.append((xdict['contri_size'], xdict['size'], filename))
+
+    for values in mismatch_size:
+        print values
+
+
+if __name__ == '__main__':
+    get_quota_xattr_brick()
+