Update doc about possibility to resume ERROR runs

README.md

@@ -8,7 +8,7 @@ Current version supports running Cassandra cluster repairs in segmented manner,
 opportunistically running multiple parallel repairs at the same time on different nodes
 within the cluster. Basic repair scheduling functionality is also supported.
 
-Cassandra Reaper does not come with a GUI, but please check 
+Cassandra Reaper does not come with a GUI, but please check
 [this project](https://github.com/spodkowinski/cassandra-reaper-ui) if you'd like to use one.
 
 Please see the [Issues](https://github.com/spotify/cassandra-reaper/issues) section for more
@@ -202,6 +202,7 @@ Source code for all the REST resources can be found from package com.spotify.rea
     * *state*: New value for the state of the repair run.
       Possible values for given state are: "PAUSED" or "RUNNING".
   * Starts, pauses, or resumes a repair run identified by the "id" path parameter.
+  * Can also be used to reattempt a repair run in state "ERROR", picking up where it left off.
 
 * DELETE  /repair_run/{id}
   * Expected query parameters:

bin/spreaper

@@ -202,7 +202,7 @@ def _arguments_for_scheduling(parser):
 
 def _arguments_for_resume_repair(parser):
     """Arguments relevant for resuming a repair"""
-    parser.add_argument("run_id", help="ID of the repair run to resume (or start)")
+    parser.add_argument("run_id", help="ID of the repair run to resume, start or reattempt")
 
 
 def _arguments_for_pause_repair(parser):
@@ -271,7 +271,7 @@ Usage: spreaper [<global_args>] <command> [<command_args>]
     schedule        Create a repair schedule, choosing the first activation time and days
                     between repair activations. You need to register a cluster into
                     Reaper (add-cluster) before calling this.
-    resume-repair   Resume paused repair run (or start a not started one).
+    resume-repair   Resume a paused, start a not started or reattempt a failed repair run.
     pause-repair    Pause a repair run.
     abort-repair    Abort a repair run.
     start-schedule  Resume a paused repair schedule.
@@ -500,7 +500,7 @@ class ReaperCLI(object):
     def resume_repair(self):
         reaper, args = ReaperCLI.prepare_reaper(
             "resume-repair",
-            "Resume a paused repair run (or trigger a not started one).",
+            "Resume a paused, start a not started or reattempt a failed repair run.",
             extra_arguments=_arguments_for_resume_repair
         )
         print "# Resuming a repair run with id: {0}".format(args.run_id)