diff --git a/test/write-failure.t b/test/write-failure.t
new file mode 100755
index 00000000..27101ee6
--- /dev/null
+++ b/test/write-failure.t
@@ -0,0 +1,144 @@
+#!/usr/bin/env python3
+# write-failure-test.py - Make sure the test will fail t
+
+import sys
+import os
+
+import atexit
+import filecmp
+import shutil
+import tempfile
+import subprocess
+import unittest
+
+from basetest import Timew, TestCase
+from basetest.utils import run_cmd_wait_nofail
+
+# Ensure python finds the local simpletap module
+sys.path.append(os.path.dirname(os.path.abspath(__file__)))
+
+class WrappedTimew(Timew):
+    def __init__(self):
+
+        try:
+            self._fiu_path = subprocess.check_output("command -v fiu-run", shell=True).strip()
+        except:
+            raise unittest.SkipTest("fiu-run is not installed")
+
+        super().__init__()
+
+    def __call__(self, *args, **kwargs):
+        """aka t = Timew() ; t() which is now an alias to t.runSuccess()"""
+        return self.run(*args, **kwargs)
+
+    def run(self, args="", input=None, merge_streams=False, timeout=5):
+        """Invoke timew with given arguments
+
+        If you wish to pass instructions to timew such as confirmations or other
+        input via stdin, you can do so by providing a input string.
+        Such as input="y\ny\n".
+
+        If merge_streams=True stdout and stderr will be merged into stdout.
+
+        timeout = number of seconds the test will wait for every timew call.
+        Defaults to 1 second if not specified. Unit is seconds.
+
+        Returns (exit_code, stdout, stderr) if merge_streams=False
+                (exit_code, output) if merge_streams=True
+        """
+
+        # This will run timewarrior with a random chance that fputs will return
+        # error 28, "No space left on device", in order to simulate running out
+        # of space on the drive the database is on. It is random in order to
+        # catch different places in the code that are writing as part of the
+        # update (tags, undo, datafiles, configs, etc..)
+        command  = [
+                    self._fiu_path, "-x",
+                    "-c", "enable_random name=posix/stdio/gp/fputs,probability=0.05,failinfo=28",
+                    "-c", "enable_random name=posix/io/rw/write,probability=0.05,failinfo=28",
+                    "-c", "enable_random name=posix/io/rw/pwrite,probability=0.05,failinfo=28",
+                    "-c", "enable_random name=posix/io/rw/writev,probability=0.05,failinfo=28",
+                    "-c", "enable_random name=posix/io/rw/pwritev,probability=0.05,failinfo=28",
+                   ]
+
+        command.extend(self._command[:])
+
+        args = self._split_string_args_if_string(args)
+        command.extend(args)
+
+        return run_cmd_wait_nofail(command, input,
+                                   merge_streams=merge_streams,
+                                   env=self.env,
+                                   timeout=timeout)
+
+
+class TestWriteFailure(TestCase):
+
+    def setUp(self):
+        self._goldendir = tempfile.mkdtemp(prefix="timew_golden_")
+        self.t = WrappedTimew()
+
+    def tearDown(self):
+        shutil.rmtree(self._goldendir)
+
+    def save_database(self):
+        shutil.rmtree(self._goldendir)
+        shutil.copytree(self.t.datadir, self._goldendir)
+
+    def compare_dirs(self, dir1, dir2):
+        compared = filecmp.dircmp(dir1, dir2)
+        if (compared.left_only or compared.right_only or
+            compared.diff_files or compared.funny_files):
+            return False
+
+        for subdir in compared.common_dirs:
+            if not self.compare_dirs(os.path.join(dir1, subdir),
+                                     os.path.join(dir2, subdir)):
+                return False
+
+        return True
+
+    def test_write_failures_do_not_corrupt_database(self):
+        """write failures shall not corrupt the database"""
+
+        # First prepopulate the database
+        start = 10
+        for x in range(start, 0, -1):
+            self.t.runSuccess("track tag{0} {1}min ago".format((start + 1)-x, x))
+
+        self.save_database()
+
+        # Now start in with the failures and make sure the database compares
+        # equally
+        TAG_COUNT = 30
+        error_count = 0
+        success_count = 0
+        MAX_ERROR_COUNT=250
+        for x in range(TAG_COUNT, 0, -1):
+            while True: 
+                exitcode, stdout, stderr = self.t("track tag-{0} {1}s ago".format((TAG_COUNT+1)-x, x))
+                if exitcode != 0:
+                    error_count += 1
+                    if error_count >= MAX_ERROR_COUNT:
+                        break
+                    
+                    if not self.compare_dirs(self.t.datadir, self._goldendir):
+                        filecmp.dircmp(self.t.datadir, self._goldendir).report_full_closure()
+                        self.fail("{0} not equal to {1}".format(self.t.datadir, self._goldendir))
+                else:
+                    success_count += 1
+                    break
+
+            if error_count >= MAX_ERROR_COUNT:
+                break
+
+            self.save_database()
+
+        self.t.runSuccess("export")
+        print("Test pass. Timew returned an error on {0} runs and none on {1} runs.".format(error_count, success_count))
+
+if __name__ == "__main__":
+    from simpletap import TAPTestRunner
+
+    unittest.main(testRunner=TAPTestRunner())
+    #unittest.main()