Fix XCC discovery issues

A couple of issues could occur during discover that should be
stepped over.

confluent_client/bin/nodeattrib

@@ -50,6 +50,9 @@ argparser.add_option('-c', '--clear', action='store_true',
                      help='Clear attributes')
 argparser.add_option('-p', '--prompt', action='store_true',
                      help='Prompt for attribute values interactively')
+argparser.add_option('-m', '--maxnodes', type='int',
+                     help='Prompt if trying to set attributes on more '
+                          'than specified number of nodes')
 (options, args) = argparser.parse_args()
 
 
@@ -87,6 +90,7 @@ if len(args) > 1:
                     if oneval != twoval:
                         print('Values did not match.')
                 argassign[arg] = twoval
+        session.stop_if_noderange_over(noderange, options.maxnodes)
         exitcode=client.updateattrib(session,args,nodetype, noderange, options, argassign)
     try:
         # setting user output to what the user inputs

confluent_client/bin/nodebmcreset

@@ -32,6 +32,8 @@ if path.startswith('/opt'):
 import confluent.client as client
 
 argparser = optparse.OptionParser(usage="Usage: %prog <noderange>")
+argparser.add_option('-m', '--maxnodes', type='int',
+                     help='Number of nodes to affect before prompting for confirmation')
 (options, args) = argparser.parse_args()
 try:
     noderange = args[0]
@@ -43,7 +45,7 @@ session = client.Command()
 exitcode = 0
 
 errorNodes = set([])
-
+session.stop_if_noderange_over(noderange, options.maxnodes)
 success = session.simple_noderange_command(noderange, 'configuration/management_controller/reset', 'reset', key='state', errnodes=errorNodes)   # = 0 if successful
 
 # Determine which nodes were successful and print them

confluent_client/bin/nodeboot

@@ -42,6 +42,10 @@ argparser.add_option('-p', '--persist', dest='persist', action='store_true',
                      default=False,
                      help='Request the boot device be persistent rather than '
                           'one time')
+argparser.add_option('-m', '--maxnodes', type='int',
+                     help='Specify a maximum number of '
+                          'nodes to boot, '
+                          'prompting if over the threshold')                        
 
 (options, args) = argparser.parse_args()
 
@@ -66,6 +70,7 @@ else:
     bootmode = 'uefi'
 
 errnodes = set([])
+session.stop_if_noderange_over(noderange, options.maxnodes)
 rc = session.simple_noderange_command(noderange, '/boot/nextdevice', bootdev,
                                      bootmode=bootmode,
                                      persistent=options.persist,

confluent_client/bin/nodeconfig

@@ -68,6 +68,10 @@ argparser.add_option('-r', '--restoredefault', default=False,
                      help='Restore the configuration of the node '
                           'to factory default for given component. '
                           'Currently only uefi is supported')
+argparser.add_option('-m', '--maxnodes', type='int',
+                     help='Specify a maximum number of '
+                          'nodes to configure, '
+                          'prompting if over the threshold')                            
 (options, args) = argparser.parse_args()
 
 cfgpaths = {
@@ -205,6 +209,7 @@ else:
 session = client.Command()
 rcode = 0
 if options.restoredefault:
+    session.stop_if_noderange_over(noderange, options.maxnodes)
     if options.restoredefault.lower() in (
         'sys', 'system', 'uefi', 'bios'):
         for fr in session.update(
@@ -225,6 +230,7 @@ if options.restoredefault:
                 options.restoredefault))
         sys.exit(1)
 if setmode:
+    session.stop_if_noderange_over(noderange, options.maxnodes)
     if options.exclude:
         sys.stderr.write('Cannot use exclude and assign at the same time\n')
         sys.exit(1)

confluent_client/bin/nodeeventlog

@@ -38,6 +38,10 @@ if sys.version_info[0] < 3:
 
 argparser = optparse.OptionParser(
     usage="Usage: %prog [options] noderange [clear]")
+argparser.add_option('-m', '--maxnodes', type='int',
+                     help='Specify a maximum number of '
+                          'nodes to clear if clearing log, '
+                          'prompting if over the threshold')            
 (options, args) = argparser.parse_args()
 try:
     noderange = args[0]
@@ -46,11 +50,11 @@ except IndexError:
     sys.exit(1)
 client.check_globbing(noderange)
 deletemode = False
-if len(sys.argv) > 3:
+if len(args) > 2:
     argparser.print_help()
     sys.exit(1)
-if len(sys.argv) == 3:
-    if sys.argv[2] == 'clear':
+if len(args) == 2:
+    if args[1] == 'clear':
         deletemode = True
     else:
         argparser.print_help()
@@ -88,6 +92,7 @@ def format_event(evt):
 
 if deletemode:
     func = session.delete
+    session.stop_if_noderange_over(noderange, options.maxnodes)
 else:
     func = session.read
 for rsp in func('/noderange/{0}/events/hardware/log'.format(noderange)):

confluent_client/bin/nodefirmware

@@ -59,6 +59,10 @@ argparser = optparse.OptionParser(
           "%prog <noderange> [list][update [--backup <file>]]|[<components>]")
 argparser.add_option('-b', '--backup', action='store_true',
                      help='Target a backup bank rather than primary')
+argparser.add_option('-m', '--maxnodes', type='int',
+                     help='When updating, prompt if more than the specified '
+                          'number of servers will be affected')
+             
 (options, args) = argparser.parse_args()
 upfile = None
 try:
@@ -95,6 +99,7 @@ def get_update_progress(session, url):
 
 def update_firmware(session, filename):
     global exitcode
+    session.stop_if_noderange_over(noderange, options.maxnodes)
     output = sq.ScreenPrinter(noderange, session)
     nodeurls = {}
     filename = os.path.abspath(filename)

confluent_client/bin/nodelicense

@@ -37,6 +37,10 @@ exitcode = 0
 argparser = optparse.OptionParser(
     usage="Usage: "
           "%prog <noderange> [list][install <file>|save <directory>|delete <name>]")
+argparser.add_option('-m', '--maxnodes', type='int',
+                     help='Specify a maximum number of '
+                          'nodes to delete licenses from, '
+                          'prompting if over the threshold')                  
 (options, args) = argparser.parse_args()
 upfile = None
 downdir = None
@@ -52,7 +56,7 @@ try:
             delete = args[2]
         elif args[1] != 'list':
             argparser.print_help()
-            sys.exit(1)
+            sys.exit(1)        
 except IndexError:
     argparser.print_help()
     sys.exit(1)
@@ -138,6 +142,7 @@ try:
     elif downdir:
         save_licenses(session, downdir)
     elif delete:
+        session.stop_if_noderange_over(noderange, options.maxnodes)
         delete_license(session, delete)
     else:
         show_licenses(session)

confluent_client/bin/nodepower

@@ -37,6 +37,11 @@ argparser = optparse.OptionParser(
 argparser.add_option('-p', '--showprevious', dest='previous',
                      action='store_true', default=False,
                      help='Show previous power state')
+argparser.add_option('-m', '--maxnodes', type='int',
+                     help='Specify a maximum number of '
+                          'nodes to change power state, '
+                          'prompting if over the threshold')
+
 (options, args) = argparser.parse_args()
 try:
     noderange = args[0]
@@ -72,4 +77,4 @@ if options.previous:
     # add dictionary to session
     session.add_precede_dict(prev)
 
-sys.exit(session.simple_noderange_command(noderange, '/power/state', setstate))
+sys.exit(session.simple_noderange_command(noderange, '/power/state', setstate, promptover=options.maxnodes))

confluent_client/bin/noderemove

@@ -35,6 +35,10 @@ import confluent.client as client
 argparser = optparse.OptionParser(
     usage='''\n       %prog noderange
              \n ''')
+argparser.add_option('-m', '--maxnodes', type='int',
+                     help='Specify a maximum number of '
+                          'nodes to delete, '
+                          'prompting if over the threshold')      
 (options, args) = argparser.parse_args()
 if len(args) != 1:
     argparser.print_help()
@@ -43,6 +47,7 @@ noderange = args[0]
 client.check_globbing(noderange)
 session = client.Command()
 exitcode = 0
+session.stop_if_noderange_over(noderange, options.maxnodes)
 for r in session.delete('/noderange/{0}'.format(noderange)):
     if 'error' in r:
         sys.stderr.write(r['error'] + '\n')

confluent_client/bin/nodereseat

@@ -32,6 +32,10 @@ if path.startswith('/opt'):
 import confluent.client as client
 
 argparser = optparse.OptionParser(usage="Usage: %prog <noderange>")
+argparser.add_option('-m', '--maxnodes', type='int',
+                     help='Specify a maximum number of '
+                          'nodes to reseat, '
+                          'prompting if over the threshold')   
 (options, args) = argparser.parse_args()
 try:
     noderange = args[0]
@@ -43,7 +47,7 @@ session = client.Command()
 exitcode = 0
 
 errorNodes = set([])
-
+session.stop_if_noderange_over(noderange, options.maxnodes)
 success = session.simple_noderange_command(noderange, 'power/reseat', 'reseat', key='reseat', errnodes=errorNodes)   # = 0 if successful
 
 # Determine which nodes were successful and print them

confluent_client/bin/nodersync

@@ -42,6 +42,10 @@ def run():
     argparser = optparse.OptionParser(
         usage="Usage: %prog location noderange:location",
     )
+    argparser.add_option('-m', '--maxnodes', type='int',
+                     help='Specify a maximum number of '
+                          'nodes to run rsync to, '
+                          'prompting if over the threshold')   
     argparser.add_option('-f', '-c', '--count', type='int', default=168,
                          help='Number of nodes to concurrently rsync')
     # among other things, FD_SETSIZE limits.  Besides, spawning too many
@@ -64,7 +68,7 @@ def run():
     pipedesc = {}
     pendingexecs = deque()
     exitcode = 0
-
+    c.stop_if_noderange_over(noderange, options.maxnodes)
     for exp in c.create('/noderange/{0}/attributes/expression'.format(noderange),
             {'expression': cmdstr}):
         if 'error' in exp:
@@ -100,7 +104,7 @@ def run():
                     if desc['type'] == 'stdout':
                         if node not in pernodeout:
                             pernodeout[node] = ''
-                        pernodeout[node] += stringify(data)
+                        pernodeout[node] += client.stringify(data)
                         if '\n' in pernodeout[node]:
                             currout, pernodeout[node] = pernodeout[node].split('\n', 1)
                             if currout:

confluent_client/bin/noderun

@@ -46,6 +46,10 @@ def run():
                          help='Number of commands to run at a time')
     argparser.add_option('-n', '--nonodeprefix', action='store_true',
                          help='Do not prefix output with node names')
+    argparser.add_option('-m', '--maxnodes', type='int',
+                     help='Specify a maximum number of '
+                          'nodes to run the command with, '
+                          'prompting if over the threshold')                           
     # among other things, FD_SETSIZE limits.  Besides, spawning too many
     # processes can be unkind for the unaware on memory pressure and such...
     argparser.disable_interspersed_args()
@@ -63,7 +67,7 @@ def run():
     pipedesc = {}
     pendingexecs = deque()
     exitcode = 0
-
+    c.stop_if_noderange_over(args[0], options.maxnodes)
     for exp in c.create('/noderange/{0}/attributes/expression'.format(args[0]),
                         {'expression': cmdstr}):
         if 'error' in exp:

confluent_client/bin/nodesetboot

@@ -43,7 +43,10 @@ argparser.add_option('-p', '--persist', dest='persist', action='store_true',
 argparser.add_option('-u', '--uefi', dest='uefi', action='store_true',
                      default=True,
                      help='Request UEFI style boot (rather than BIOS)')
-
+argparser.add_option('-m', '--maxnodes', type='int',
+                     help='Specify a maximum number of '
+                          'nodes to modify next boot device, '
+                          'prompting if over the threshold')                        
 (options, args) = argparser.parse_args()
 
 try:
@@ -63,6 +66,7 @@ if options.biosmode:
     bootmode = 'bios'
 else:
     bootmode = 'uefi'
+session.stop_if_noderange_over(noderange, options.maxnodes)
 sys.exit(session.simple_noderange_command(noderange, '/boot/nextdevice', bootdev,
                                      bootmode=bootmode,
                                      persistent=options.persist))

confluent_client/bin/nodeshell

@@ -46,6 +46,10 @@ def run():
                          help='Number of commands to run at a time')
     argparser.add_option('-n', '--nonodeprefix', action='store_true',
                          help='Do not prefix output with node names')
+    argparser.add_option('-m', '--maxnodes', type='int',
+                     help='Specify a maximum number of '
+                          'nodes to run remote ssh command to, '
+                          'prompting if over the threshold')   
     # among other things, FD_SETSIZE limits.  Besides, spawning too many
     # processes can be unkind for the unaware on memory pressure and such...
     argparser.disable_interspersed_args()
@@ -55,7 +59,7 @@ def run():
         sys.exit(1)
     client.check_globbing(args[0])
     concurrentprocs = options.count
-    c = client.Command()
+    c = client.Command()    
     cmdstr = " ".join(args[1:])
 
     currprocs = 0
@@ -64,7 +68,7 @@ def run():
     pendingexecs = deque()
     exitcode = 0
 
-
+    c.stop_if_noderange_over(args[0], options.maxnodes)
     for exp in c.create('/noderange/{0}/attributes/expression'.format(args[0]),
                         {'expression': cmdstr}):
         if 'error' in exp:

confluent_client/bin/nodestorage

@@ -63,10 +63,16 @@ def _print_cfg(scfg):
             sys.stderr.write(e['error'] + '\n')
             exitcode = e.get('errorcode', 1)
         for node in e.get('databynode', {}):
+            curr = e['databynode'][node]
+            if 'error' in curr:
+                if 'no available drives' in curr['error']:
+                    curr['error'] += ' (drives must be in unconfigured state to be available, they must not be in jbod or online state)'
+                sys.stderr.write('{0}: {1}\n'.format(node, curr['error']))
+                exitcode = curr.get('errorcode', 1)
+                continue
             if node not in storagebynode:
                 storagebynode[node] = {'disks': [], 'arrays': [],
                                        'volumes': []}
-            curr = e['databynode'][node]
             storagebynode[node][curr['type'] + 's'].append(curr)
     for node in storagebynode:
         for disk in sorted(storagebynode[node]['disks'],
@@ -108,6 +114,7 @@ def createstorage(noderange, options, args):
         sys.stderr.write('-r and -d are required arguments to create array\n')
         sys.exit(1)
     session = client.Command()
+    session.stop_if_noderange_over(noderange, options.maxnodes)
     names = options.name
     if names is None:
         names = ''.join(args)
@@ -132,6 +139,7 @@ def deletestorage(noderange, options, args):
     else:
         names = options.name
     session = client.Command()
+    session.stop_if_noderange_over(noderange, options.maxnodes)
     for rsp in session.delete(
             '/noderange/{0}/configuration/storage/volumes/{1}'.format(
                 noderange, names)):
@@ -162,6 +170,7 @@ def setdisk(noderange, options, args):
         sys.stderr.write('diskset requires valid state as argument (hotspare, jbod, unconfigured)\n')
         sys.exit(1)
     session = client.Command()
+    session.stop_if_noderange_over(noderange, options.maxnodes)
     scfg = session.update('/noderange/{0}/configuration/storage/disks/{1}'.format(noderange, names), {'state': args[0]})
     _print_cfg(scfg)
 
@@ -202,6 +211,10 @@ def main():
                          help='Comma separated list of stripsizes to use when creating volumes. '
                             'This value is in kilobytes.  The default behavior is to allow the '
                             'storage controller to decide.')
+    argparser.add_option('-m', '--maxnodes', type='int',
+                     help='Specify a maximum number of '
+                          'nodes to configure storage on, '
+                          'prompting if over the threshold') 
     (options, args) = argparser.parse_args()
     if len(args) == 1:
         args.append('show')

confluent_client/bin/nodesupport

@@ -64,7 +64,7 @@ def printerror(res, node=None):
 
 
 
-def download_servicedata(noderange, media):
+def download_servicedata(noderange, media, options):
     global exitcode
     session = client.Command()
     output = sq.ScreenPrinter(noderange, session)
@@ -73,6 +73,7 @@ def download_servicedata(noderange, media):
     upargs = {'filename': filename}
     noderrs = {}
     nodeurls = {}
+    session.stop_if_noderange_over(noderange, options.maxnodes)
     for res in session.create(resource, upargs):
         if 'created' not in res:
             for nodename in res.get('databynode', ()):
@@ -121,6 +122,10 @@ def main():
                'management server (the confluent server if running remote, '
                'and the collective.manager if in collective)\n'
                '\n\nSee `man %prog` for more info.\n')
+    argparser.add_option('-m', '--maxnodes', type='int',
+                     help='Specify a maximum number of '
+                          'nodes to download diagnostic data from, '
+                          'prompting if over the threshold') 
     (options, args) = argparser.parse_args()
     media = None
     try:
@@ -142,6 +147,6 @@ def main():
     except KeyError:
         argparser.print_help()
         sys.exit(1)
-    handler(noderange, media)
+    handler(noderange, media, options)
 if __name__ == '__main__':
     main()

confluent_client/confluent/client.py

@@ -39,6 +39,10 @@ _attraliases = {
     'bmcpass': 'secret.hardwaremanagementpassword',
 }
 
+try:
+    input = raw_input
+except NameError:
+    pass
 
 def stringify(instr):
     # Normalize unicode and bytes to 'str', correcting for
@@ -219,7 +223,7 @@ class Command(object):
         return rc
 
     def simple_noderange_command(self, noderange, resource, input=None,
-                                 key=None, errnodes=None, **kwargs):
+                                 key=None, errnodes=None, promptover=None, **kwargs):
         try:
             self._currnoderange = noderange
             rc = 0
@@ -235,6 +239,7 @@ class Command(object):
                         noderange, resource)):
                     rc = self.handle_results(ikey, rc, res, errnodes)
             else:
+                self.stop_if_noderange_over(noderange, promptover)
                 kwargs[ikey] = input
                 for res in self.update('/noderange/{0}/{1}'.format(
                         noderange, resource), kwargs):
@@ -244,6 +249,33 @@ class Command(object):
         except KeyboardInterrupt:
             cprint('')
             return 0
+    
+    def stop_if_noderange_over(self, noderange, maxnodes):
+        if maxnodes is None:
+            return
+        nsize = self.get_noderange_size(noderange)
+        if nsize > maxnodes:
+            if nsize == 1:
+                nodename = list(self.read(
+                    '/noderange/{0}/nodes/'.format(noderange)))[0].get('item', {}).get('href', None)
+                nodename = nodename[:-1]
+                p = input('Command is about to affect node {0}, continue (y/n)? '.format(nodename))
+            else:
+                p = input('Command is about to affect {0} nodes, continue (y/n)? '.format(nsize))
+            if p.lower() != 'y':
+                sys.stderr.write('Aborting at user request\n')
+                sys.exit(1)
+                raise Exception("Aborting at user request")
+        
+
+    def get_noderange_size(self, noderange):
+        numnodes = 0
+        for node in self.read('/noderange/{0}/nodes/'.format(noderange)):
+            if node.get('item', {}).get('href', None):
+                numnodes += 1
+            else:
+                raise Exception("Error trying to size noderange {0}".format(noderange))
+        return numnodes
 
     def simple_nodegroups_command(self, noderange, resource, input=None, key=None, **kwargs):
         try:
@@ -344,12 +376,8 @@ class Command(object):
                 if fingerprint == khf[hostid]:
                     return
                 else:
-                    try:
-                        replace = raw_input(
-                            "MISMATCHED CERTIFICATE DATA, ACCEPT NEW? (y/n):")
-                    except NameError:
-                        replace = input(
-                            "MISMATCHED CERTIFICATE DATA, ACCEPT NEW? (y/n):")
+                    replace = input(
+                        "MISMATCHED CERTIFICATE DATA, ACCEPT NEW? (y/n):")
                     if replace not in ('y', 'Y'):
                         raise Exception("BAD CERTIFICATE")
             cprint('Adding new key for %s:%s' % (server, port))
@@ -401,6 +429,10 @@ def printattributes(session, requestargs, showtype, nodetype, noderange, options
     path = '/{0}/{1}/attributes/{2}'.format(nodetype, noderange, showtype)
     return print_attrib_path(path, session, requestargs, options)
 
+def _sort_attrib(k):
+    if isinstance(k[1], dict) and k[1].get('sortid', None) is not None:
+        return k[1]['sortid']
+    return k[0]
 
 def print_attrib_path(path, session, requestargs, options, rename=None, attrprefix=None):
     exitcode = 0
@@ -411,9 +443,7 @@ def print_attrib_path(path, session, requestargs, options, rename=None, attrpref
             exitcode = 1
             continue
         for node in sorted(res['databynode']):
-            for attr, val in sorted(
-                    res['databynode'][node].items(),
-                    key=lambda k: k[1].get('sortid', k[0]) if isinstance(k[1], dict) else k[0]):
+            for attr, val in sorted(res['databynode'][node].items(), key=_sort_attrib):
                 if attr == 'error':
                     sys.stderr.write('{0}: Error: {1}\n'.format(node, val))
                     continue

confluent_client/confluent_env.sh

@@ -46,9 +46,8 @@ alias nodelicense='CURRENT_CMDLINE=$(HISTTIMEFORMAT= builtin history 1); export
 _confluent_get_args()
 {
     CMPARGS=($COMP_LINE)
-    NUMARGS=${#CMPARGS[@]}
-    if [ "${COMP_WORDS[-1]}" == '' ]; then
-        NUMARGS=$((NUMARGS+1))
+    NUMARGS=$((COMP_CWORD+1))
+    if [ "${COMP_WORDS[COMP_CWORD]}" == '' ]; then
         CMPARGS+=("")
     fi
     GENNED=""
@@ -75,7 +74,7 @@ function _confluent_generic_completion()
 {
     _confluent_get_args
     if [ $NUMARGS -ge 3 ] && [ ! -z "$GENNED" ]; then
-        COMPREPLY=($(compgen -W "$GENNED" -- ${COMP_WORDS[-1]}))
+        COMPREPLY=($(compgen -W "$GENNED" -- ${COMP_WORDS[COMP_CWORD]}))
     fi
     if [ $NUMARGS -lt 3 ]; then
         _confluent_nr_completion
@@ -111,7 +110,7 @@ _confluent_nodemedia_completion()
         return
     fi
     if [ $NUMARGS -ge 3 ] && [ ! -z "$GENNED" ]; then
-        COMPREPLY=($(compgen -W "$GENNED" -- ${COMP_WORDS[-1]}))
+        COMPREPLY=($(compgen -W "$GENNED" -- ${COMP_WORDS[COMP_CWORD]}))
         return;
     fi
     if [ $NUMARGS -lt 3 ]; then
@@ -124,7 +123,7 @@ _confluent_nodefirmware_completion()
 {
     _confluent_get_args
     if [ $NUMARGS == 3 ]; then
-        COMPREPLY=($(compgen -W "list update" -- ${COMP_WORDS[-1]}))
+        COMPREPLY=($(compgen -W "list update" -- ${COMP_WORDS[COMP_CWORD]}))
         return;
     fi
     if [ $NUMARGS -gt 3 ] && [ ${CMPARGS[2]} == 'update' ]; then
@@ -142,7 +141,7 @@ _confluent_nodeshell_completion()
 {
     _confluent_get_args
     if [ $NUMARGS == 3 ]; then
-        COMPREPLY=($(compgen -c -- ${COMP_WORDS[-1]}))
+        COMPREPLY=($(compgen -c -- ${COMP_WORDS[COMP_CWORD]}))
         return
     fi
     if [ $NUMARGS -gt 3 ]; then
@@ -160,7 +159,7 @@ _confluent_nodelicense_completion()
 {
     _confluent_get_args
     if [ $NUMARGS == 3 ]; then
-        COMPREPLY=($(compgen -W "install list save delete" -- ${COMP_WORDS[-1]}))
+        COMPREPLY=($(compgen -W "install list save delete" -- ${COMP_WORDS[COMP_CWORD]}))
         return;
     fi
     if [ $NUMARGS == 4 ] && [ ${CMPARGS[2]} == 'install' ]; then
@@ -183,7 +182,7 @@ _confluent_nodesupport_completion()
 {
     _confluent_get_args
     if [ $NUMARGS == 3 ]; then
-        COMPREPLY=($(compgen -W "servicedata" -- ${COMP_WORDS[-1]}))
+        COMPREPLY=($(compgen -W "servicedata" -- ${COMP_WORDS[COMP_CWORD]}))
         return;
     fi
     if [ $NUMARGS == 4 ] && [ ${CMPARGS[2]} == 'servicedata' ]; then
@@ -210,41 +209,36 @@ _confluent_nn_completion()
     if [ $NUMARGS -gt 2 ]; then
         return;
     fi
-    INPUT=${COMP_WORDS[-1]}
+    INPUT=${COMP_WORDS[COMP_CWORD]}
     INPUT=${INPUT##*,-}
     INPUT=${INPUT##*,}
     INPUT=${INPUT##*@}
     PREFIX=""
-    if [ "$INPUT" != "${COMP_WORDS[-1]}" ]; then
-        PREFIX=${COMP_WORDS[-1]}
+    if [ "$INPUT" != "${COMP_WORDS[COMP_CWORD]}" ]; then
+        PREFIX=${COMP_WORDS[COMP_CWORD]}
         PREFIX=$(echo $PREFIX | sed -e 's/,[^,@-]*$/,/' -e 's/,-[^,@]*$/,-/' -e 's/@[^,@]*/@/')
     fi
 
-    COMPREPLY=($(compgen -W "$(nodelist | sed -e s/^/$PREFIX/)" -- "${COMP_WORDS[-1]}"))
+    COMPREPLY=($(compgen -W "$(nodelist | sed -e s/^/$PREFIX/)" -- "${COMP_WORDS[COMP_CWORD]}"))
 }
 _confluent_nr_completion()
 {
     CMPARGS=($COMP_LINE)
-    NUMARGS=${#CMPARGS[@]}
-    if [ "${COMP_WORDS[-1]}" == '' ]; then
-        NUMARGS=$((NUMARGS+1))
-    fi
     _confluent_get_args
     if [ $NUMARGS -gt 2 ]; then
         return;
     fi
-    INPUT=${COMP_WORDS[-1]}
+    INPUT=${COMP_WORDS[COMP_CWORD]}
     INPUT=${INPUT##*,-}
     INPUT=${INPUT##*,}
     INPUT=${INPUT##*@}
     PREFIX=""
-    if [ "$INPUT" != "${COMP_WORDS[-1]}" ]; then
-        PREFIX=${COMP_WORDS[-1]}
+    if [ "$INPUT" != "${COMP_WORDS[COMP_CWORD]}" ]; then
+        PREFIX=${COMP_WORDS[COMP_CWORD]}
         PREFIX=$(echo $PREFIX | sed -e 's/,[^,@-]*$/,/' -e 's/,-[^,@]*$/,-/' -e 's/@[^,@]*/@/')
     fi
 
-    #COMPREPLY=($(compgen -W "$(confetty show /nodegroups|sed -e 's/\///' -e s/^/$PREFIX/;nodelist | sed -e s/^/$PREFIX/)" -- "${COMP_WORDS[-1]}"))
-    COMPREPLY=($(compgen -W "$(confetty show /nodegroups|sed -e 's/\///' -e s/^/$PREFIX/;nodelist | sed -e s/^/$PREFIX/)" -- "${COMP_WORDS[-1]}"))
+    COMPREPLY=($(compgen -W "$(confetty show /nodegroups|sed -e 's/\///' -e s/^/$PREFIX/;nodelist | sed -e s/^/$PREFIX/)" -- "${COMP_WORDS[COMP_CWORD]}"))
 }
 _confluent_ng_completion()
 {
@@ -252,17 +246,17 @@ _confluent_ng_completion()
     if [ $NUMARGS -gt 2 ]; then
         return;
     fi
-    INPUT=${COMP_WORDS[-1]}
+    INPUT=${COMP_WORDS[COMP_CWORD]}
     INPUT=${INPUT##*,-}
     INPUT=${INPUT##*,}
     INPUT=${INPUT##*@}
     PREFIX=""
-    if [ "$INPUT" != "${COMP_WORDS[-1]}" ]; then
-        PREFIX=${COMP_WORDS[-1]}
+    if [ "$INPUT" != "${COMP_WORDS[COMP_CWORD]}" ]; then
+        PREFIX=${COMP_WORDS[COMP_CWORD]}
         PREFIX=$(echo $PREFIX | sed -e 's/,[^,@-]*$/,/' -e 's/,-[^,@]*$/,-/' -e 's/@[^,@]*/@/')
     fi
 
-    COMPREPLY=($(compgen -W "$(confetty show /nodegroups|sed -e 's/\///' -e s/^/$PREFIX/)" -- "${COMP_WORDS[-1]}"))
+    COMPREPLY=($(compgen -W "$(confetty show /nodegroups|sed -e 's/\///' -e s/^/$PREFIX/)" -- "${COMP_WORDS[COMP_CWORD]}"))
 }
 complete -F _confluent_nodeattrib_completion nodeattrib
 complete -F _confluent_nodeattrib_completion nodegroupattrib

confluent_client/doc/man/nodefirmware.ronn

@@ -13,7 +13,11 @@ nodefirmware(8) -- Report firmware information on confluent nodes
 will retrieve all firmware, but can be directed to fetch specific firmware by
 calling out the name of the firmware (e.g. uefi or xcc) or request reading only
 core firmware firmware by using the word 'core', which is generally a quicker
-operation.
+operation.  Different hardwaremanagement.method indicated plugins may have
+different capabilities available.  For example, the 'core' distinction may
+not be relevant to redfish.  Additionally, the Lenovo XCC makes certain
+information available over IPMI that is not otherwise available (for example
+the FPGA version where applicable).
 
 In the update form, it accepts a single file and attempts to update it using
 the out of band facilities.  Firmware updates can end in one of three states:

confluent_client/doc/man/nodesupport.ronn

@@ -7,7 +7,7 @@ nodesupport(8) -- Utilities for interacting with vendor support
 
 ## DESCRIPTION
 
-`nodesupport` provides capabilities associated with interactiong with support.
+`nodesupport` provides capabilities associated with interacting with support.
 Currently it only has the `servicedata` subcommand.  `servicedata` takes
 an argument that is either a directory name (that can be used for a single node
 or multiple nodes) or a file name (only to be used with single node noderange).
@@ -16,6 +16,9 @@ connects to the managed system, so it will download to the remote system if runn
 remotely and will download to the collective.manager indicated system if
 running in collective mode.
 
+Note that due to vendor filename requirements, any filename may have vendor
+specific suffixes added to any file produced.
+
 ## EXAMPLES
 
 * Download support data from a single node to a specific filename  

confluent_server/builddeb

@@ -3,8 +3,12 @@ cd `dirname $0`
 PKGNAME=$(basename $(pwd))
 DPKGNAME=$(basename $(pwd) | sed -e s/_/-/)
 OPKGNAME=$(basename $(pwd) | sed -e s/_/-/)
+PYEXEC=python3
+DSCARGS="--with-python3=True --with-python2=False"
 if grep wheezy /etc/os-release; then
     DPKGNAME=python-$DPKGNAME
+    PYEXEC=python
+    DSCARGS="--with-python3=False --with-python2=True"
 fi
 cd ..
 mkdir -p /tmp/confluent # $DPKGNAME
@@ -24,15 +28,15 @@ install-scripts=/opt/confluent/bin
 package=$DPKGNAME
 EOF
 
-python setup.py sdist > /dev/null 2>&1
-py2dsc dist/*.tar.gz
+$PYEXEC setup.py sdist > /dev/null 2>&1
+py2dsc $DSCARGS dist/*.tar.gz
 shopt -s extglob
 cd deb_dist/!(*.orig)/
 if [ "$OPKGNAME" = "confluent-server" ]; then
     if grep wheezy /etc/os-release; then
         sed -i 's/^\(Depends:.*\)/\1, python-confluent-client, python-lxml, python-eficompressor, python-pycryptodomex, python-dateutil, python-pyopenssl/' debian/control
     else
-        sed -i 's/^\(Depends:.*\)/\1, confluent-client, python-lxml, python-eficompressor, python-pycryptodome, python-dateutil/' debian/control
+        sed -i 's/^\(Depends:.*\)/\1, confluent-client, python3-lxml, python3-eficompressor, python3-pycryptodome, python3-websocket/' debian/control
     fi
     if grep wheezy /etc/os-release; then
         echo 'confluent_client python-confluent-client' >> debian/pydist-overrides

confluent_server/confluent/auth.py

@@ -58,9 +58,9 @@ _allowedbyrole = {
             '/node*/configuration/*',
         ],
         'update': [
-            '/discovery/*', 
+            '/discovery/*',
             '/networking/macs/rescan',
-            '/node*/power/state', 
+            '/node*/power/state',
             '/node*/power/reseat',
             '/node*/attributes/*',
             '/node*/media/*tach',

confluent_server/confluent/collective/manager.py

@@ -160,11 +160,11 @@ def connect_to_leader(cert=None, name=None, leader=None):
                 raise
             currentleader = leader
         #spawn this as a thread...
-        follower = eventlet.spawn(follow_leader, remote, pvers)
+        follower = eventlet.spawn(follow_leader, remote, pvers, leader)
     return True
 
 
-def follow_leader(remote, proto):
+def follow_leader(remote, proto, leader):
     global currentleader
     cleanexit = False
     try:
@@ -176,8 +176,8 @@ def follow_leader(remote, proto):
             log.log({'info': 'Previous following cleanly closed',
                      'subsystem': 'collective'})
             return
-        log.log({'info': 'Current leader has disappeared, restarting '
-                         'collective membership', 'subsystem': 'collective'})
+        log.log({'info': 'Current leader ({0}) has disappeared, restarting '
+                         'collective membership'.format(leader), 'subsystem': 'collective'})
         # The leader has folded, time to startup again...
         cfm.stop_following()
         currentleader = None

confluent_server/confluent/config/attributes.py

@@ -271,7 +271,7 @@ node = {
     'console.method': {
         'description': ('Indicate the method used to access the console of '
                         'the managed node.'),
-        'validvalues': ('ssh', 'ipmi'),
+        'validvalues': ('ssh', 'ipmi', 'tsmsol'),
     },
 #    'virtualization.host': {
 #        'description': ('Hypervisor where this node does/should reside'),

confluent_server/confluent/config/configmanager.py

@@ -328,6 +328,11 @@ def exec_on_leader(function, *args):
 
 
 def exec_on_followers(fnname, *args):
+    pushes = eventlet.GreenPool()
+    # Check health of collective prior to attempting
+    for _ in pushes.starmap(
+        _push_rpc, [(cfgstreams[s], b'') for s in cfgstreams]):
+        pass
     if len(cfgstreams) < (len(_cfgstore['collective']) // 2):
         # the leader counts in addition to registered streams
         raise exc.DegradedCollective()
@@ -386,9 +391,15 @@ def init_masterkey(password=None, autogen=True):
 
 def _push_rpc(stream, payload):
     with _rpclock:
-        stream.sendall(struct.pack('!Q', len(payload)))
-        if len(payload):
-            stream.sendall(payload)
+        try:
+            stream.sendall(struct.pack('!Q', len(payload)))
+            if len(payload):
+                stream.sendall(payload)
+            return True
+        except Exception:
+            logException()
+            del cfgstreams[stream]
+            stream.close()
 
 
 def decrypt_value(cryptvalue,
@@ -578,11 +589,18 @@ def relay_slaved_requests(name, listener, vers):
             lh = StreamHandler(listener)
             _hasquorum = len(cfgstreams) >= (
                     len(_cfgstore['collective']) // 2)
-            payload = cPickle.dumps({'quorum': _hasquorum}, protocol=lowestver)
-            for _ in pushes.starmap(
-                    _push_rpc,
-                    [(cfgstreams[s], payload) for s in cfgstreams]):
-                pass
+            _newquorum = None
+            while _hasquorum != _newquorum:
+                if _newquorum is not None:
+                    _hasquorum = _newquorum
+                payload = cPickle.dumps({'quorum': _hasquorum}, protocol=lowestver)
+                for _ in pushes.starmap(
+                        _push_rpc,
+                        [(cfgstreams[s], payload) for s in cfgstreams]):
+                    pass
+                _newquorum = len(cfgstreams) >= (
+                        len(_cfgstore['collective']) // 2)
+            _hasquorum = _newquorum
             if _hasquorum and _pending_collective_updates:
                 apply_pending_collective_updates()
             msg = lh.get_next_msg()
@@ -604,8 +622,10 @@ def relay_slaved_requests(name, listener, vers):
                     except Exception as e:
                         exc = e
                     if 'xid' in rpc:
-                        _push_rpc(listener, cPickle.dumps({'xid': rpc['xid'],
+                        res = _push_rpc(listener, cPickle.dumps({'xid': rpc['xid'],
                                                            'exc': exc}, protocol=vers))
+                        if not res:
+                            break
                 try:
                     msg = lh.get_next_msg()
                 except Exception:
@@ -650,7 +670,9 @@ class StreamHandler(object):
                 if confluent.util.monotonic_time() > self.expiry:
                     return None
                 if confluent.util.monotonic_time() > self.keepalive:
-                    _push_rpc(self.sock, b'')  # nulls are a keepalive
+                    res = _push_rpc(self.sock, b'')  # nulls are a keepalive
+                    if not res:
+                        return None
                     self.keepalive = confluent.util.monotonic_time() + 20
             self.expiry = confluent.util.monotonic_time() + 60
             msg = self.sock.recv(8)
@@ -767,7 +789,9 @@ def follow_channel(channel, proto=2):
                         _pendingchangesets[rpc['xid']].send()
                 if 'quorum' in rpc:
                     _hasquorum = rpc['quorum']
-                _push_rpc(channel, b'')  # use null as ACK
+                res = _push_rpc(channel, b'')  # use null as ACK
+                if not res:
+                    break
             msg = lh.get_next_msg()
     finally:
         # mark the connection as broken
@@ -1137,6 +1161,8 @@ class ConfigManager(object):
             attribute, match = expression.split('=')
         else:
             raise Exception('Invalid Expression')
+        if attribute.startswith('secret.'):
+            raise Exception('Filter by secret attributes is not supported')
         for node in nodes:
             try:
                 currvals = [self._cfgstore['nodes'][node][attribute]['value']]
@@ -1626,6 +1652,7 @@ class ConfigManager(object):
                 if group in self._cfgstore['nodes'][node]['groups']:
                     self._cfgstore['nodes'][node]['groups'].remove(group)
                     self._node_removed_from_group(node, group, changeset)
+                    _mark_dirtykey('nodes', node, self.tenant)
         for node in nodes:
             if node not in self._cfgstore['nodes']:
                 self._cfgstore['nodes'][node] = {'groups': [group]}

confluent_server/confluent/consoleserver.py

@@ -362,13 +362,17 @@ class ConsoleHandler(object):
         if self.reconnect:
             self.reconnect.cancel()
             self.reconnect = None
+        strerror = ('The console.method attribute for this node is '
+                'not configured,\r\nset it to a valid value for console '
+                'function')
         try:
             self._console = list(plugin.handle_path(
                 self._plugin_path.format(self.node),
                 "create", self.cfgmgr))[0]
         except (exc.NotImplementedException, exc.NotFoundException):
             self._console = None
-        except:
+        except Exception as e:
+            strerror = str(e)
             if _tracelog:
                 _tracelog.log(traceback.format_exc(), ltype=log.DataTypes.event,
                               event=log.Events.stacktrace)
@@ -381,13 +385,9 @@ class ConsoleHandler(object):
             self._send_rcpts({'connectstate': self.connectstate,
                               'error': self.error})
             self.feedbuffer(
-                '\x1bc\x1b[2J\x1b[1;1H[The console.method attribute for this node is '
-                'not configured,\r\nset it to a valid value for console '
-                'function]')
+                '\x1bc\x1b[2J\x1b[1;1H[{0}]'.format(strerror))
             self._send_rcpts(
-                '\x1bc\x1b[2J\x1b[1;1H[The console.method attribute for this node is '
-                'not configured,\r\nset it to a valid value for console '
-                'function]')
+                '\x1bc\x1b[2J\x1b[1;1H[{0}]'.format(strerror))
             self.clearerror = True
             return
         if self.clearerror:

confluent_server/confluent/core.py

@@ -72,7 +72,7 @@ import struct
 import sys
 
 pluginmap = {}
-dispatch_plugins = (b'ipmi', u'ipmi')
+dispatch_plugins = (b'ipmi', u'ipmi', b'redfish', u'redfish', b'tsmsol', u'tsmsol')
 
 
 def seek_element(currplace, currkey):
@@ -224,7 +224,7 @@ def _init_core():
                     'advanced': PluginRoute({
                         'pluginattrs': ['hardwaremanagement.method'],
                         'default': 'ipmi',
-                    }),                    
+                    }),
                 },
             },
             'storage': {

confluent_server/confluent/discovery/core.py

@@ -949,9 +949,9 @@ def eval_node(cfg, handler, info, nodename, manual=False):
                     #     raise exc.InvalidArgumentException(errorstr)
                     # log.log({'error': errorstr})
                     if encuuid in pending_by_uuid:
-                        pending_by_uuid[encuuid].add(info)
+                        pending_by_uuid[encuuid].append(info)
                     else:
-                        pending_by_uuid[encuuid] = set([info])
+                        pending_by_uuid[encuuid] = [info]
                     return
                 # We found the real smm, replace the list with the actual smm
                 # to continue
@@ -1094,6 +1094,10 @@ def discover_node(cfg, handler, info, nodename, manual):
         info['discostatus'] = 'discovered'
         for i in pending_by_uuid.get(curruuid, []):
             eventlet.spawn_n(_recheck_single_unknown_info, cfg, i)
+        try:
+            del pending_by_uuid[curruuid]
+        except KeyError:
+            pass
         return True
     log.log({'info': 'Detected {0}, but discovery.policy is not set to a '
                      'value allowing discovery (open or permissive)'.format(

confluent_server/confluent/discovery/handlers/generic.py

@@ -68,17 +68,21 @@ class NodeHandler(object):
     def _savecert(self, certificate):
         self._fp = certificate
         return True
-    
+
     def get_node_credentials(self, nodename, creds, defuser, defpass):
         user = creds.get(nodename, {}).get(
             'secret.hardwaremanagementuser', {}).get('value', None)
         havecustomcreds = False
+        if user and not isinstance(user, str):
+            user = user.decode('utf8')
         if user is not None and user != defuser:
             havecustomcreds = True
         else:
             user = defuser
         passwd = creds.get(nodename, {}).get(
             'secret.hardwaremanagementpassword', {}).get('value', None)
+        if passwd and not isinstance(passwd, str):
+            passwd = passwd.decode('utf8')
         if passwd is not None and passwd != defpass:
             havecustomcreds = True
         else:

confluent_server/confluent/discovery/handlers/smm.py

@@ -98,7 +98,7 @@ class NodeHandler(bmchandler.NodeHandler):
                 setdata += ',v4Gateway:{0}'.format(gateway)
             wc.request('POST', '/data', setdata)
             rsp = wc.getresponse()
-            rspdata = rsp.read()
+            rspdata = util.stringify(rsp.read())
             if '<statusCode>0' not in rspdata:
                 raise Exception("Error configuring SMM Network")
             return
@@ -145,7 +145,7 @@ class NodeHandler(bmchandler.NodeHandler):
             authdata['password'] = password
             wc.request('POST', '/data/login', urlencode(authdata), headers)
             rsp = wc.getresponse()
-            rspdata = rsp.read()
+            rspdata = util.stringify(rsp.read())
         if 'authResult>0' in rspdata:
             tokens = fromstring(rspdata)
             st2 = tokens.findall('st2')[0].text
@@ -181,6 +181,10 @@ class NodeHandler(bmchandler.NodeHandler):
             'secret.hardwaremanagementuser', {}).get('value', 'USERID')
         passwd = creds.get(nodename, {}).get(
             'secret.hardwaremanagementpassword', {}).get('value', 'PASSW0RD')
+        if not isinstance(username, str):
+            username = username.decode('utf8')
+        if not isinstance(passwd, str):
+            passwd = passwd.decode('utf8')
         if passwd == 'PASSW0RD' and self.ruleset:
             raise Exception('Cannot support default password and setting password rules at same time')
         if passwd == 'PASSW0RD':

confluent_server/confluent/discovery/handlers/tsm.py

@@ -22,6 +22,7 @@ try:
     from urllib import urlencode
 except ImportError:
     from urllib.parse import urlencode
+
 getaddrinfo = eventlet.support.greendns.getaddrinfo
 
 webclient = eventlet.import_patched('pyghmi.util.webclient')
@@ -43,6 +44,13 @@ class NodeHandler(generic.NodeHandler):
         self.atdefault = True
         super(NodeHandler, self).__init__(info, configmanager)
 
+    def scan(self):
+        c = webclient.SecureHTTPConnection(self.ipaddr, 443, verifycallback=self.validate_cert)
+        i = c.grab_json_response('/redfish/v1/')
+        uuid = i.get('UUID', None)
+        if uuid:
+            self.info['uuid'] = uuid
+
     def validate_cert(self, certificate):
         # broadly speaking, merely checks consistency moment to moment,
         # but if https_cert gets stricter, this check means something
@@ -54,9 +62,17 @@ class NodeHandler(generic.NodeHandler):
             'username': self.DEFAULT_USER,
             'password': self.DEFAULT_PASS,
         }
+        wc = webclient.SecureHTTPConnection(self.ipaddr, 443, verifycallback=self.validate_cert)
+        wc.set_header('Content-Type', 'application/json')
+        authmode = 0
         if not self.trieddefault:
-            wc = webclient.SecureHTTPConnection(self.ipaddr, 443, verifycallback=self.validate_cert)
-            rsp, status = wc.grab_json_response_with_status('/api/session', urlencode(authdata))
+            rsp, status = wc.grab_json_response_with_status('/api/session', authdata)
+            if status == 403:
+                wc.set_header('Content-Type', 'application/x-www-form-urlencoded')
+                authmode = 1
+                rsp, status = wc.grab_json_response_with_status('/api/session', urlencode(authdata))
+            else:
+                authmode = 2
             if status > 400:
                 rsp = util.stringify(rsp)
                 self.trieddefault = True
@@ -68,9 +84,15 @@ class NodeHandler(generic.NodeHandler):
                         'default_password': self.DEFAULT_PASS,
                         'username': self.DEFAULT_USER
                         }
-                    rsp, status = wc.grab_json_response_with_status('/api/reset-pass', urlencode(passchange))
+                    if authmode == 2:
+                        rsp, status = wc.grab_json_response_with_status('/api/reset-pass', passchange)
+                    else:
+                        rsp, status = wc.grab_json_response_with_status('/api/reset-pass', urlencode(passchange))
                     authdata['password'] = self.targpass
-                    rsp, status = wc.grab_json_response_with_status('/api/session', urlencode(authdata))
+                    if authmode == 2:
+                        rsp, status = wc.grab_json_response_with_status('/api/session', authdata)
+                    else:
+                        rsp, status = wc.grab_json_response_with_status('/api/session', urlencode(authdata))
                     self.csrftok = rsp['CSRFToken']
                     self.channel = rsp['channel']
                     self.curruser = self.DEFAULT_USER
@@ -85,15 +107,23 @@ class NodeHandler(generic.NodeHandler):
         if self.curruser:
             authdata['username'] = self.curruser
             authdata['password'] = self.currpass
-            rsp, status = wc.grab_json_response_with_status('/api/session', urlencode(authdata))
-            if rsp.status != 200:
+            if authmode != 1:
+                rsp, status = wc.grab_json_response_with_status('/api/session', authdata)
+            if authmode == 1 or status == 403:
+                wc.set_header('Content-Type', 'application/x-www-form-urlencoded')
+                rsp, status = wc.grab_json_response_with_status('/api/session', urlencode(authdata))
+            if status != 200:
                 return None
             self.csrftok = rsp['CSRFToken']
             self.channel = rsp['channel']
             return wc
         authdata['username'] = self.targuser
         authdata['password'] = self.targpass
-        rsp, status = wc.grab_json_response_with_status('/api/session', urlencode(authdata))
+        if authmode != 1:
+            rsp, status = wc.grab_json_response_with_status('/api/session', authdata)
+        if authmode == 1 or status == 403:
+            wc.set_header('Content-Type', 'application/x-www-form-urlencoded')
+            rsp, status = wc.grab_json_response_with_status('/api/session', urlencode(authdata))
         if status != 200:
             return None
         self.curruser = self.targuser

confluent_server/confluent/discovery/handlers/xcc.py

@@ -15,6 +15,7 @@
 import base64
 import codecs
 import confluent.discovery.handlers.imm as immhandler
+import confluent.exceptions as exc
 import confluent.netutil as netutil
 import confluent.util as util
 import errno
@@ -95,7 +96,8 @@ class NodeHandler(immhandler.NodeHandler):
             ipmicmd.xraw_command(netfn=0x3a, command=0xf1, data=(1,))
         except pygexc.IpmiException as e:
             if (e.ipmicode != 193 and 'Unauthorized name' not in str(e) and
-                    'Incorrect password' not in str(e)):
+                    'Incorrect password' not in str(e) and 
+                    str(e) != 'Session no longer connected'):
                 # raise an issue if anything other than to be expected
                 if disableipmi:
                     _, _ = wc.grab_json_response_with_status(
@@ -170,7 +172,12 @@ class NodeHandler(immhandler.NodeHandler):
                 pwdchanged = True
             if '_csrf_token' in wc.cookies:
                 wc.set_header('X-XSRF-TOKEN', wc.cookies['_csrf_token'])
+            if pwdchanged:
+                # Remove the minimum change interval, to allow sane 
+                # password changes after provisional changes
+                self.set_password_policy('')
             return (wc, pwdchanged)
+        return (None, None)
 
     @property
     def wc(self):
@@ -204,7 +211,7 @@ class NodeHandler(immhandler.NodeHandler):
                 # however the target *will* demand a new password... if it's currently
                 # PASSW0RD
                 # use TempW0rd42 to avoid divulging a real password on the line
-                # This is replacing one well known password (PASSW0RD) with another 
+                # This is replacing one well known password (PASSW0RD) with another
                 # (TempW0rd42)
                 passwd = 'TempW0rd42'
             wc, pwdchanged = self.get_webclient('USERID', 'PASSW0RD', passwd)
@@ -226,9 +233,9 @@ class NodeHandler(immhandler.NodeHandler):
         if wc:
             return wc
 
-    def set_password_policy(self):
+    def set_password_policy(self, strruleset):
         ruleset = {'USER_GlobalMinPassChgInt': '0'}
-        for rule in self.ruleset.split(','):
+        for rule in strruleset.split(','):
             if '=' not in rule:
                 continue
             name, value = rule.split('=')
@@ -350,15 +357,18 @@ class NodeHandler(immhandler.NodeHandler):
         # between hypothetical secure path and today.
         dpp = self.configmanager.get_node_attributes(
             nodename, 'discovery.passwordrules')
-        self.ruleset = dpp.get(nodename, {}).get(
+        strruleset = dpp.get(nodename, {}).get(
             'discovery.passwordrules', {}).get('value', '')
         wc = self.wc
         creds = self.configmanager.get_node_attributes(
             self.nodename, ['secret.hardwaremanagementuser',
             'secret.hardwaremanagementpassword'], decrypt=True)
         user, passwd, isdefault = self.get_node_credentials(nodename, creds, 'USERID', 'PASSW0RD')
-        self.set_password_policy()
+        self.set_password_policy(strruleset)
         if self._atdefaultcreds:
+            if isdefault and self.tmppasswd:
+                raise Exception(
+                    'Request to use default credentials, but refused by target after it has been changed to {0}'.format(self.tmppasswd))
             if not isdefault:
                 self._setup_xcc_account(user, passwd, wc)
         self._convert_sha256account(user, passwd, wc)

confluent_server/confluent/discovery/protocols/slp.py

@@ -84,6 +84,8 @@ def _parse_SrvRply(parsed):
     :return:
     """
     payload = parsed['payload']
+    if len(payload) < 4:
+        return
     ecode, ucount = struct.unpack('!HH', bytes(payload[0:4]))
     if ecode:
         parsed['errorcode'] = ecode
@@ -234,13 +236,20 @@ def _find_srvtype(net, net4, srvtype, addresses, xid):
 
 
 def _grab_rsps(socks, rsps, interval, xidmap):
-    r, _, _ = select.select(socks, (), (), interval)
+    r = None
+    res = select.select(socks, (), (), interval)
+    if res:
+        r = res[0]
     while r:
         for s in r:
             (rsp, peer) = s.recvfrom(9000)
             neighutil.refresh_neigh()
             _parse_slp_packet(rsp, peer, rsps, xidmap)
-            r, _, _ = select.select(socks, (), (), interval)
+            res = select.select(socks, (), (), interval)
+            if not res:
+                r = None
+            else:
+                r = res[0]
 
 
 
@@ -560,7 +569,7 @@ def scan(srvtypes=_slp_services, addresses=None, localonly=False):
     # now to analyze and flesh out the responses
     for id in rsps:
         if 'service:ipmi' in rsps[id]['services']:
-            if 'service:ipmi://Athena:623' in rsps[id]['urls']:
+            if 'service:ipmi://Athena:623' in rsps[id].get('urls', ''):
                 rsps[id]['services'] = ['service:thinkagile-storage']
             else:
                 continue

confluent_server/confluent/exceptions.py

@@ -107,11 +107,14 @@ class PubkeyInvalid(ConfluentException):
         self.fingerprint = fingerprint
         self.attrname = attribname
         self.message = text
+        certtxt = base64.b64encode(certificate)
+        if not isinstance(certtxt, str):
+            certtxt = certtxt.decode('utf8')
         bodydata = {'message': text,
                     'event': event,
                     'fingerprint': fingerprint,
                     'fingerprintfield': attribname,
-                    'certificate': base64.b64encode(certificate)}
+                    'certificate': certtxt}
         self.errorbody = json.dumps(bodydata)
 
     def get_error_body(self):

confluent_server/confluent/firmwaremanager.py

@@ -43,13 +43,16 @@ def execupdate(handler, filename, updateobj, type, owner, node):
                                    'detail': errstr})
         return
     if type == 'ffdc' and os.path.isdir(filename):
-        filename += '/' + node + '.svcdata'
+        filename += '/' + node
     try:
         if type == 'firmware':
             completion = handler(filename, progress=updateobj.handle_progress,
                                  bank=updateobj.bank)
         else:
             completion = handler(filename, progress=updateobj.handle_progress)
+        if type == 'ffdc' and completion:
+            filename = completion
+            completion = None
         if completion is None:
             completion = 'complete'
         if owner:

confluent_server/confluent/messages.py

@@ -76,7 +76,7 @@ def _htmlify_structure(indict):
                     if datum is None:
                         nd.append('')
                     else:
-                        nd.append(datum)
+                        nd.append(util.stringify(datum))
                 ret += ",".join(nd)
             else:
                 for v in indict:
@@ -138,6 +138,7 @@ class ConfluentMessage(object):
         snippet = ""
         for key in pairs:
             val = pairs[key]
+            key = util.stringify(key)
             value = self.defaultvalue
             if isinstance(val, dict) and 'type' in val:
                 valtype = val['type']
@@ -338,6 +339,7 @@ class ConfluentChoiceMessage(ConfluentMessage):
         snippet = ''
         for key in pairdata:
             val = pairdata[key]
+            key = util.stringify(key)
             snippet += key + ':<select name="%s">' % key
             valid_values = self.valid_values
             if key in self.valid_paramset:
@@ -1377,6 +1379,8 @@ class InputAlertDestination(ConfluentMessage):
         self.alertcfg = {}
         if multinode:  # keys are node names
             for node in inputdata:
+                if not isinstance(inputdata[node], dict):
+                    break
                 self.alertcfg[node] = inputdata[node]
                 for key in inputdata[node]:
                     if key not in self.valid_alert_params:
@@ -1389,7 +1393,8 @@ class InputAlertDestination(ConfluentMessage):
                     else:
                         self.alertcfg[node][key] = \
                             self.valid_alert_params[key](inputdata[node][key])
-        else:
+            else:
+                return
             for key in inputdata:
                 if key not in self.valid_alert_params:
                     raise exc.InvalidArgumentException(
@@ -1487,12 +1492,15 @@ class Volume(ConfluentMessage):
 
 class Disk(ConfluentMessage):
     valid_states = set([
+        'fault',
         'jbod',
         'unconfigured',
         'hotspare',
+        'rebuilding',
         'online',
     ])
     state_aliases = {
+        'unconfigured bad': 'fault',
         'unconfigured good': 'unconfigured',
         'global hot spare': 'hotspare',
         'dedicated hot spare': 'hotspare',

confluent_server/confluent/networking/lldp.py

@@ -125,7 +125,8 @@ def _dump_neighbordatum(info):
 
 def b64tohex(b64str):
     bd = base64.b64decode(b64str)
-    return ''.join(['{0:02x}'.format(ord(x)) for x in bd])
+    bd = bytearray(bd)
+    return ''.join(['{0:02x}'.format(x) for x in bd])
 
 def get_fingerprint(switch, port, configmanager, portmatch):
     update_switch_data(switch, configmanager)

confluent_server/confluent/plugins/console/tsmsol.py

@@ -0,0 +1,160 @@
+# vim: tabstop=4 shiftwidth=4 softtabstop=4
+
+# Copyright 2015-2019 Lenovo
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+
+# This plugin provides an ssh implementation comforming to the 'console'
+# specification.  consoleserver or shellserver would be equally likely
+# to use this.
+
+import confluent.exceptions as cexc
+import confluent.interface.console as conapi
+import confluent.log as log
+import confluent.util as util
+import pyghmi.exceptions as pygexc
+import pyghmi.redfish.command as rcmd
+import eventlet
+import eventlet.green.ssl as ssl
+try:
+    websocket = eventlet.import_patched('websocket')
+    wso = websocket.WebSocket
+except Exception:
+    wso = object
+
+def get_conn_params(node, configdata):
+    if 'secret.hardwaremanagementuser' in configdata:
+        username = configdata['secret.hardwaremanagementuser']['value']
+    else:
+        username = 'USERID'
+    if 'secret.hardwaremanagementpassword' in configdata:
+        passphrase = configdata['secret.hardwaremanagementpassword']['value']
+    else:
+        passphrase = 'PASSW0RD'  # for lack of a better guess
+    if 'hardwaremanagement.manager' in configdata:
+        bmc = configdata['hardwaremanagement.manager']['value']
+    else:
+        bmc = node
+    return {
+        'username': username,
+        'passphrase': passphrase,
+        'bmc': bmc,
+    }
+_configattributes = ('secret.hardwaremanagementuser',
+                     'secret.hardwaremanagementpassword',
+                     'hardwaremanagement.manager')
+
+class WrappedWebSocket(wso):
+
+    def set_verify_callback(self, callback):
+        self._certverify = callback
+
+    def connect(self, url, **options):
+        add_tls = url.startswith('wss://')
+        if add_tls:
+            hostname, port, resource, _ = websocket._url.parse_url(url)
+            if hostname[0] != '[' and ':' in hostname:
+                hostname = '[{0}]'.format(hostname)
+            if resource[0] != '/':
+                resource = '/{0}'.format(resource)
+            url = 'ws://{0}:443{1}'.format(hostname,resource)
+        else:
+            return super(WrappedWebSocket, self).connect(url, **options)
+        self.sock_opt.timeout = options.get('timeout', self.sock_opt.timeout)
+        self.sock, addrs = websocket._http.connect(url, self.sock_opt, websocket._http.proxy_info(**options),
+                                           options.pop('socket', None))
+        self.sock = ssl.wrap_socket(self.sock, cert_reqs=ssl.CERT_NONE)
+        # The above is supersedeed by the _certverify, which provides
+        # known-hosts style cert validaiton
+        bincert = self.sock.getpeercert(binary_form=True)
+        if not self._certverify(bincert):
+            raise pygexc.UnrecognizedCertificate('Unknown certificate', bincert)
+        try:
+            self.handshake_response = websocket._handshake.handshake(self.sock, *addrs, **options)
+            if self.handshake_response.status in websocket._handshake.SUPPORTED_REDIRECT_STATUSES:
+                options['redirect_limit'] = options.pop('redirect_limit', 3) - 1
+                if options['redirect_limit'] < 0:
+                     raise Exception('Redirect limit hit')
+                url = self.handshake_response.headers['location']
+                self.sock.close()
+                return self.connect(url, **options)
+            self.connected = True
+        except:
+            if self.sock:
+                self.sock.close()
+                self.sock = None
+            raise
+
+
+         
+
+
+
+class TsmConsole(conapi.Console):
+
+    def __init__(self, node, config):
+        self.node = node
+        self.ws = None
+        configdata = config.get_node_attributes([node], _configattributes, decrypt=True)
+        connparams = get_conn_params(node, configdata[node])
+        self.username = connparams['username']
+        self.password = connparams['passphrase']
+        self.bmc = connparams['bmc']
+        self.origbmc = connparams['bmc']
+        if ':' in self.bmc:
+            self.bmc = '[{0}]'.format(self.bmc)
+        self.datacallback = None
+        self.nodeconfig = config
+        self.connected = False
+
+
+    def recvdata(self):
+        while self.connected:
+            pendingdata = self.ws.recv()
+            if pendingdata == '':
+                self.datacallback(conapi.ConsoleEvent.Disconnect)
+                return
+            self.datacallback(pendingdata)
+
+    def connect(self, callback):
+        self.datacallback = callback
+        rc = rcmd.Command(self.origbmc, self.username,
+                          self.password,
+                          verifycallback=lambda x: True)
+        wc = rc.oem.wc
+        bmc = self.bmc
+        if '%' in self.bmc:
+            prefix = self.bmc.split('%')[0]
+            bmc = prefix + ']'
+        self.ws = WrappedWebSocket(host=bmc)
+        kv = util.TLSCertVerifier(
+            self.nodeconfig, self.node, 'pubkeys.tls_hardwaremanager').verify_cert
+        self.ws.set_verify_callback(kv)
+        self.ws.connect('wss://{0}/sol?CSRFTOKEN={1}'.format(self.bmc, rc.oem.csrftok), host=bmc, cookie='QSESSIONID={0}'.format(wc.cookies['QSESSIONID']))
+        self.connected = True
+        eventlet.spawn_n(self.recvdata)
+        return
+
+    def write(self, data):
+        self.ws.send(data)
+
+    def close(self):
+        if self.ws:
+            self.ws.close()
+        self.connected = False
+        self.datacallback = None
+
+def create(nodes, element, configmanager, inputdata):
+    if len(nodes) == 1:
+        return TsmConsole(nodes[0], configmanager)

confluent_server/confluent/plugins/hardwaremanagement/cnos.py

@@ -78,6 +78,18 @@ def retrieve(nodes, element, configmanager, inputdata):
         for node in nodes:
             workers.add(eventlet.spawn(retrieve_health, configmanager, creds,
                                        node, results))
+    elif element[:3] == ['inventory', 'hardware', 'all']:
+        creds = configmanager.get_node_attributes(
+                nodes, ['secret.hardwaremanagementuser', 'secret.hardwaremanagementpassword'], decrypt=True)
+        for node in nodes:
+            workers.add(eventlet.spawn(retrieve_inventory, configmanager,
+                                       creds, node, results, element))
+    elif element[:3] == ['inventory', 'firmware', 'all']:
+        creds = configmanager.get_node_attributes(
+                nodes, ['secret.hardwaremanagementuser', 'secret.hardwaremanagementpassword'], decrypt=True)
+        for node in nodes:
+            workers.add(eventlet.spawn(retrieve_firmware, configmanager,
+                                       creds, node, results, element))
     else:
         for node in nodes:
             yield msg.ConfluentNodeError(node, 'Not Implemented')
@@ -105,6 +117,44 @@ def retrieve(nodes, element, configmanager, inputdata):
         pass
 
 
+def retrieve_inventory(configmanager, creds, node, results, element):
+    if len(element) == 3:
+        results.put(msg.ChildCollection('all'))
+        results.put(msg.ChildCollection('system'))
+        return
+    wc = cnos_login(node, configmanager, creds)
+    sysinfo = wc.grab_json_response('/nos/api/sysinfo/inventory')
+    invinfo = {
+        'inventory': [{
+            'name': 'System',
+            'present': True,
+            'information': {
+                'Product name': sysinfo['Model'],
+                'Serial Number': sysinfo['Electronic Serial Number'],
+                'Board Serial Number': sysinfo['Serial Number'],
+                'Manufacturer': 'Lenovo',
+                'Model': sysinfo['Machine Type Model'],
+                'FRU Number': sysinfo['FRU'].strip(),
+            }
+        }]
+    }
+    results.put(msg.KeyValueData(invinfo, node))
+
+
+def retrieve_firmware(configmanager, creds, node, results, element):
+    if len(element) == 3:
+        results.put(msg.ChildCollection('all'))
+        return
+    wc = cnos_login(node, configmanager, creds)
+    sysinfo = wc.grab_json_response('/nos/api/sysinfo/inventory')
+    items = [{
+        'Software': {'version': sysinfo['Software Revision']},
+        },
+        {
+        'BIOS': {'version': sysinfo['BIOS Revision']},
+        }]
+    results.put(msg.Firmware(items, node))
+
 def retrieve_health(configmanager, creds, node, results):
     wc = cnos_login(node, configmanager, creds)
     hinfo = wc.grab_json_response('/nos/api/sysinfo/globalhealthstatus')

confluent_server/confluent/plugins/hardwaremanagement/ipmi.py

@@ -42,6 +42,11 @@ import traceback
 if not hasattr(ssl, 'SSLEOFError'):
     ssl.SSLEOFError = None
 
+try:
+    range = xrange
+except NameError:
+    pass
+
 pci_cache = {}
 
 def get_dns_txt(qstring):
@@ -131,7 +136,7 @@ def hex2bin(hexstring):
     if len(hexvals) < 2:
         hexvals = hexstring.split(' ')
     if len(hexvals) < 2:
-        hexvals = [hexstring[i:i+2] for i in xrange(0, len(hexstring), 2)]
+        hexvals = [hexstring[i:i+2] for i in range(0, len(hexstring), 2)]
     bytedata = [int(i, 16) for i in hexvals]
     return bytearray(bytedata)
 
@@ -278,6 +283,7 @@ def _donothing(data):
 
 class IpmiConsole(conapi.Console):
     configattributes = frozenset(_configattributes)
+    bmctonodemapping = {}
 
     def __init__(self, node, config):
         self.error = None
@@ -295,10 +301,21 @@ class IpmiConsole(conapi.Console):
         self.bmc = connparams['bmc']
         self.port = connparams['port']
         self.connected = False
+        # ok, is self.bmc unique among nodes already
         # Cannot actually create console until 'connect', when we get callback
+        if (self.bmc in self.bmctonodemapping and
+                self.bmctonodemapping[self.bmc] != node):
+            raise Exception(
+                "Duplicate hardwaremanagement.manager attribute for {0} and {1}".format(
+                    node, self.bmctonodemapping[self.bmc]))
+        self.bmctonodemapping[self.bmc] = node
 
     def __del__(self):
         self.solconnection = None
+        try:
+            del self.bmctonodemapping[self.bmc]
+        except KeyError:
+            pass
 
     def handle_data(self, data):
         if type(data) == dict:
@@ -584,9 +601,13 @@ class IpmiHandler(object):
         self.output.put(msg.CreatedResource(
             'nodes/{0}/media/uploads/{1}'.format(self.node, u.name)))
 
+    def get_diags(self, savefile, progress):
+        return self.ipmicmd.get_diagnostic_data(
+            savefile, progress=progress, autosuffix=True)
+
     def handle_servicedata_fetch(self):
         u = firmwaremanager.Updater(
-            self.node, self.ipmicmd.get_diagnostic_data,
+            self.node, self.get_diags,
             self.inputdata.nodefile(self.node), self.tenant, type='ffdc',
             owner=self.current_user)
         self.output.put(msg.CreatedResource(
@@ -658,7 +679,7 @@ class IpmiHandler(object):
             if len(self.element) == 4:
                 # A list of destinations
                 maxdest = self.ipmicmd.get_alert_destination_count()
-                for alertidx in xrange(0, maxdest + 1):
+                for alertidx in range(0, maxdest + 1):
                     self.output.put(msg.ChildCollection(alertidx))
                 return
             elif len(self.element) == 5:

confluent_server/confluent/plugins/hardwaremanagement/redfish.py

@@ -175,12 +175,15 @@ class IpmiCommandWrapper(ipmicommand.Command):
                     raise exc.TargetEndpointUnreachable(se.strerror)
                 else:
                     raise exc.TargetEndpointUnreachable(str(se))
-            if isinstance(se, socket.timeout) or (len(se) > 1 and se[1] == 'EHOSTUNREACH'):
+            if isinstance(se, socket.timeout):
                 raise exc.TargetEndpointUnreachable('timeout')
             raise
         except pygexc.PyghmiException as pe:
             if 'Access Denied' in str(pe):
                 raise exc.TargetEndpointBadCredentials()
+            if 'Redfish not ready' in str(pe):
+                raise exc.TargetEndpointUnreachable('Redfish not yet ready')
+            raise
 
     def close_confluent(self):
         if self._attribwatcher:
@@ -462,9 +465,13 @@ class IpmiHandler(object):
         self.output.put(msg.CreatedResource(
             'nodes/{0}/media/uploads/{1}'.format(self.node, u.name)))
 
+    def get_diags(self, savefile, progress):
+        return self.ipmicmd.get_diagnostic_data(
+            savefile, progress=progress, autosuffix=True)
+
     def handle_servicedata_fetch(self):
         u = firmwaremanager.Updater(
-            self.node, self.ipmicmd.get_diagnostic_data,
+            self.node, self.get_diags,
             self.inputdata.nodefile(self.node), self.tenant, type='ffdc',
             owner=self.current_user)
         self.output.put(msg.CreatedResource(

confluent_server/confluent_server.spec.tmpl

@@ -13,9 +13,9 @@ BuildRoot: %{_tmppath}/%{name}-%{version}-%{release}-buildroot
 Prefix: %{_prefix}
 BuildArch: noarch
 %if "%{dist}" == ".el8"
-Requires: python3-pyghmi >= 1.0.34, python3-eventlet, python3-greenlet, python3-pycryptodomex >= 3.4.7, confluent_client, python3-pyparsing, python3-paramiko, python3-dns, python3-netifaces, python3-pyasn1 >= 0.2.3, python3-pysnmp >= 4.3.4, python3-pyte, python3-lxml, python3-eficompressor, python3-setuptools, python3-dateutil, python3-enum34, python3-asn1crypto, python3-cffi, python3-pyOpenSSL, python3-monotonic
+Requires: python3-pyghmi >= 1.0.34, python3-eventlet, python3-greenlet, python3-pycryptodomex >= 3.4.7, confluent_client, python3-pyparsing, python3-paramiko, python3-dns, python3-netifaces, python3-pyasn1 >= 0.2.3, python3-pysnmp >= 4.3.4, python3-pyte, python3-lxml, python3-eficompressor, python3-setuptools, python3-dateutil, python3-enum34, python3-asn1crypto, python3-cffi, python3-pyOpenSSL, python3-monotonic, python3-websocket-client
 %else
-Requires: python-pyghmi >= 1.0.34, python-eventlet, python-greenlet, python-pycryptodomex >= 3.4.7, confluent_client, python-pyparsing, python-paramiko, python-dns, python-netifaces, python2-pyasn1 >= 0.2.3, python-pysnmp >= 4.3.4, python-pyte, python-lxml, python-eficompressor, python-setuptools, python-dateutil
+Requires: python-pyghmi >= 1.0.34, python-eventlet, python-greenlet, python-pycryptodomex >= 3.4.7, confluent_client, python-pyparsing, python-paramiko, python-dns, python-netifaces, python2-pyasn1 >= 0.2.3, python-pysnmp >= 4.3.4, python-pyte, python-lxml, python-eficompressor, python-setuptools, python-dateutil, python2-websocket-client
 %endif
 Vendor: Jarrod Johnson <jjohnson2@lenovo.com>
 Url: http://xcat.sf.net/
@@ -42,7 +42,10 @@ python2 setup.py install --single-version-externally-managed -O1 --root=$RPM_BUI
 for file in $(grep confluent/__init__.py INSTALLED_FILES.bare); do
     rm $RPM_BUILD_ROOT/$file
 done
-grep -v confluent/__init__.py INSTALLED_FILES.bare > INSTALLED_FILES
+grep -v confluent/__init__.py INSTALLED_FILES.bare | grep -v etc/init.d/confluent > INSTALLED_FILES
+rm $RPM_BUILD_ROOT/etc/init.d/confluent
+rmdir $RPM_BUILD_ROOT/etc/init.d
+rmdir $RPM_BUILD_ROOT/etc
 cat INSTALLED_FILES
 
 %triggerin -- python-pyghmi

confluent_server/setup.py.tmpl

@@ -14,6 +14,7 @@ setup(
               'confluent/discovery/handlers/',
               'confluent/networking/',
               'confluent/plugins/hardwaremanagement/',
+              'confluent/plugins/console/',
               'confluent/plugins/shell/',
               'confluent/collective/',
               'confluent/plugins/configuration/'],