Capture more exceptions during runs, collect results by error condition

- Also print a summary at the end.

multiball/__main__.py

@@ -4,6 +4,8 @@
 
 #
 # TODO
+# - keep track of result status, indicate how many and which ones failed with an error, failed with a return code, and
+#   succeeded, print at end.
 # - keep track of previous command outputs for last command, and any previous command with target list and command line
 # - calling commands and alieses from command line
 # - tagging and filtering commands (so #noupgrade gets a tag, and we can filter them)
@@ -23,7 +25,8 @@
 # - implement various commented commands in the command list
 # - implement interactive alias system
 # - Catch more exceptions in fabtools, and also add retries
-
+# - Make the runner aware of multiple commands so that it can combine outputs and make 'overall success' or 'overall failure'
+# - make C-c break the connections not the program
 
 import argparse
 import datetime

multiball/fabtools.py

@@ -1,22 +1,39 @@
+import socket
 import time
-
+from enum import Enum
 from pathlib import Path
-from threading import Thread, Lock
-
-from tqdm import tqdm
-
-from fabric2 import ThreadingGroup, SerialGroup, Config, Connection
+from threading import Lock, Thread
 
 import paramiko
+from fabric2 import Config, Connection, SerialGroup, ThreadingGroup
+from tqdm import tqdm
 
 
+class RunResult(Enum):
+    Success = 0
+    CommandFailure = 1
+    Error = 2
+    Cancelled = 3
+    Timeout = 4 # FIXME support host timeouts
+
 def thread_run(connection, command, result_lock, result_queue):
+    runresult = RunResult.Success
     try:
         res = connection.run(command, warn=True, hide=True)
-    except (paramiko.ssh_exception.NoValidConnectionsError, paramiko.ssh_exception.SSHException) as inst:
+        if (res.exited != 0):
+            runresult = RunResult.CommandFailure
+    except (paramiko.ssh_exception.NoValidConnectionsError, paramiko.ssh_exception.SSHException, socket.gaierror) as inst:
         res = f"Could not connect to host: {inst}"
+        runresult = RunResult.Error
+    except KeyboardInterrupt as inst:
+        res = f"Canceled host {inst} from C-c"
+        runresult = RunResult.Cancelled
+    except Exception as inst:
+        res = f"Unknown error for host: {inst}"
+        runresult = RunResult.Error
     with result_lock:
-        result_queue.append((connection, res))
+        result_queue.append((connection, res, runresult))
+
 
 # A set of hosts we can target with a series of commands and also can collect output of each command
 class HostSet:
@@ -58,24 +75,37 @@ class HostSet:
                 if ((time.time() - tupdate) > 10.0):
                     tupdate = time.time()
                     print("Still waiting on ", [thread.host for thread in threads])
+                    # FIXME have the thread killed if it's still waiting after $TIMEOUT
         prog.close()
 
         # Gather up results by output
         gathered = {}
-        for connection, result in resq:
+        success = []
+        error = []
+        commandfail = []
+
+        for connection, result, runresult in resq:
+            # print(connection, result, runresult)
             rstr = str(result)
             if not rstr in gathered:
                 gathered[rstr] = []
             gathered[rstr].append(connection)
+            if runresult in (RunResult.Error, RunResult.Cancelled):
+                error.append(connection)
+            elif runresult == RunResult.CommandFailure:
+                commandfail.append(connection)
+            else:
+                success.append(connection)
 
         # display results
         for result, connections in gathered.items():
             print('-----> [{}]'.format(' '.join(connection.original_host for connection in connections)))
             print(result)
-        # ## gather_output
-        # for connection in self.connections:
-        #     # import pdb
-        #     # pdb.set_trace()
-        #     res = connection.run(command)
-        #     print('---- {} ----'.format(connection.original_host))
-        #     print(res)
+
+        print('#####> RUN SUMMARY <#####')
+        if (success):
+            print('        Succeeded: ', ', '.join(connection.original_host for connection in success))
+        if (commandfail):
+            print('     Command fail: ', ', '.join(connection.original_host for connection in commandfail))
+        if (error):
+            print('Failed with Error: ', ','.join(connection.original_host for connection in error))