modifications realted to special account
[tests.git] / system / TestPlc.py
index 6fb76eb..377c4a4 100644 (file)
@@ -3,9 +3,9 @@ import os, os.path
 import datetime
 import time
 import sys
-import datetime
 import traceback
 from types import StringTypes
+import socket
 
 import utils
 from TestSite import TestSite
@@ -17,6 +17,8 @@ from TestSliver import TestSliver
 from TestBox import TestBox
 from TestSsh import TestSsh
 from TestApiserver import TestApiserver
+from TestSliceSfa import TestSliceSfa
+from TestUserSfa import TestUserSfa
 
 # step methods must take (self) and return a boolean (options is a member of the class)
 
@@ -44,6 +46,8 @@ def node_mapper (method):
                 test_node = TestNode (self,test_site,node_spec)
                 if not node_method(test_node): overall=False
         return overall
+    # restore the doc text
+    actual.__doc__=method.__doc__
     return actual
 
 def slice_mapper_options (method):
@@ -56,10 +60,63 @@ def slice_mapper_options (method):
             test_slice=TestSlice(self,test_site,slice_spec)
             if not slice_method(test_slice,self.options): overall=False
         return overall
+    # restore the doc text
+    actual.__doc__=method.__doc__
+    return actual
+
+def slice_mapper_options_sfa (method):
+    def actual(self):
+       test_plc=self
+        overall=True
+        slice_method = TestSliceSfa.__dict__[method.__name__]
+        for slice_spec in self.plc_spec['sfa']['slices_sfa']:
+            site_spec = self.locate_site (slice_spec['sitename'])
+            test_site = TestSite(self,site_spec)
+            test_slice=TestSliceSfa(test_plc,test_site,slice_spec)
+            if not slice_method(test_slice,self.options): overall=False
+        return overall
+    # restore the doc text
+    actual.__doc__=method.__doc__
     return actual
 
+SEP='<sep>'
+
 class TestPlc:
 
+    default_steps = [
+        'display', 'local_pre', SEP,
+        'delete','create','install', 'configure', 'start', SEP,
+        'fetch_keys', 'store_keys', 'clear_known_hosts', SEP,
+        'initscripts', 'sites', 'nodes', 'slices', 'nodegroups', SEP,
+        'reinstall_node', 'init_node','bootcd', 'configure_qemu', 'export_qemu',
+        'kill_all_qemus', 'start_node', SEP,
+        # better use of time: do this now that the nodes are taking off
+        'plcsh_stress_test', SEP,
+        'nodes_ssh_debug', 'nodes_ssh_boot', 'check_slice', 'check_initscripts', SEP,
+       'install_sfa', 'configure_sfa', 'import_sfa', 'start_sfa', SEP,
+        'setup_sfa', 'add_sfa', 'update_sfa', SEP,
+        'view_sfa', 'check_slice_sfa', 'delete_sfa', 'stop_sfa', SEP,
+        'check_tcp',  'check_hooks',  SEP,
+        'force_gather_logs', 'force_local_post',
+        ]
+    other_steps = [ 
+        'fresh_install', 'stop', 'vs_start', SEP,
+        'clean_initscripts', 'clean_nodegroups','clean_all_sites', SEP,
+        'clean_sites', 'clean_nodes', 'clean_slices', 'clean_keys', SEP,
+        'populate' , SEP,
+        'show_boxes', 'list_all_qemus', 'list_qemus', 'kill_qemus', SEP,
+        'db_dump' , 'db_restore', SEP,
+        'local_list','local_cleanup',SEP,
+        'standby_1 through 20',
+        ]
+
+    @staticmethod
+    def printable_steps (list):
+        return " ".join(list).replace(" "+SEP+" "," \\\n")
+    @staticmethod
+    def valid_step (step):
+        return step != SEP
+
     def __init__ (self,plc_spec,options):
        self.plc_spec=plc_spec
         self.options=options
@@ -70,17 +127,12 @@ class TestPlc:
             self.url="https://%s:443/PLCAPI/"%plc_spec['vserverip']
             self.vserver=True
         except:
-            self.vserver=False
-            self.url="https://%s:443/PLCAPI/"%plc_spec['hostname']
-#        utils.header('Using API url %s'%self.url)
+            raise Exception,'chroot-based myplc testing is deprecated'
        self.apiserver=TestApiserver(self.url,options.dry_run)
         
     def name(self):
         name=self.plc_spec['name']
-        if self.vserver:
-            return name+".vserver.%s"%self.vservername
-        else:
-            return name+".chroot"
+        return "%s.%s"%(name,self.vservername)
 
     def hostname(self):
         return self.plc_spec['hostname']
@@ -96,40 +148,26 @@ class TestPlc:
     def actual_command_in_guest (self,command):
         return self.test_ssh.actual_command(self.host_to_guest(command))
     
+    def start_guest (self):
+      return utils.system(self.test_ssh.actual_command(self.start_guest_in_host()))
+    
     def run_in_guest (self,command):
         return utils.system(self.actual_command_in_guest(command))
     
     def run_in_host (self,command):
         return self.test_ssh.run_in_buildname(command)
 
-    #command gets run in the chroot/vserver
+    #command gets run in the vserver
     def host_to_guest(self,command):
-        if self.vserver:
-            return "vserver %s exec %s"%(self.vservername,command)
-        else:
-            return "chroot /plc/root %s"%TestSsh.backslash_shell_specials(command)
+        return "vserver %s exec %s"%(self.vservername,command)
+    
+    #command gets run in the vserver
+    def start_guest_in_host(self):
+        return "vserver %s start"%(self.vservername)
     
-    # copy a file to the myplc root image - pass in_data=True if the file must go in /plc/data
-    def copy_in_guest (self, localfile, remotefile, in_data=False):
-        if in_data:
-            chroot_dest="/plc/data"
-        else:
-            chroot_dest="/plc/root"
-        if self.is_local():
-            if not self.vserver:
-                utils.system("cp %s %s/%s"%(localfile,chroot_dest,remotefile))
-            else:
-                utils.system("cp %s /vservers/%s/%s"%(localfile,self.vservername,remotefile))
-        else:
-            if not self.vserver:
-                utils.system("scp %s %s:%s/%s"%(localfile,self.hostname(),chroot_dest,remotefile))
-            else:
-                utils.system("scp %s %s@/vservers/%s/%s"%(localfile,self.hostname(),self.vservername,remotefile))
-
-
     # xxx quick n dirty
     def run_in_guest_piped (self,local,remote):
-        return utils.system(local+" | "+self.test_ssh.actual_command(self.host_to_guest(remote)))
+        return utils.system(local+" | "+self.test_ssh.actual_command(self.host_to_guest(remote),keep_stdin=True))
 
     def auth_root (self):
        return {'Username':self.plc_spec['PLC_ROOT_USER'],
@@ -171,6 +209,37 @@ class TestPlc:
                 return slice
         raise Exception,"Cannot locate slice %s"%slicename
 
+    def all_sliver_objs (self):
+        result=[]
+        for slice_spec in self.plc_spec['slices']:
+            slicename = slice_spec['slice_fields']['name']
+            for nodename in slice_spec['nodenames']:
+                result.append(self.locate_sliver_obj (nodename,slicename))
+        return result
+
+    def locate_sliver_obj (self,nodename,slicename):
+        (site,node) = self.locate_node(nodename)
+        slice = self.locate_slice (slicename)
+        # build objects
+        test_site = TestSite (self, site)
+        test_node = TestNode (self, test_site,node)
+        # xxx the slice site is assumed to be the node site - mhh - probably harmless
+        test_slice = TestSlice (self, test_site, slice)
+        return TestSliver (self, test_node, test_slice)
+
+    def locate_first_node(self):
+        nodename=self.plc_spec['slices'][0]['nodenames'][0]
+        (site,node) = self.locate_node(nodename)
+        test_site = TestSite (self, site)
+        test_node = TestNode (self, test_site,node)
+        return test_node
+
+    def locate_first_sliver (self):
+        slice_spec=self.plc_spec['slices'][0]
+        slicename=slice_spec['slice_fields']['name']
+        nodename=slice_spec['nodenames'][0]
+        return self.locate_sliver_obj(nodename,slicename)
+
     # all different hostboxes used in this plc
     def gather_hostBoxes(self):
         # maps on sites and nodes, return [ (host_box,test_node) ]
@@ -198,11 +267,12 @@ class TestPlc:
 
     # make this a valid step
     def kill_all_qemus(self):
+        "all qemu boxes: kill all running qemus (even of former runs)"
         # this is the brute force version, kill all qemus on that host box
         for (box,nodes) in self.gather_hostBoxes().iteritems():
             # pass the first nodename, as we don't push template-qemu on testboxes
-            nodename=nodes[0].name()
-            TestBox(box,self.options.buildname).kill_all_qemus(nodename)
+            nodedir=nodes[0].nodedir()
+            TestBox(box,self.options.buildname).kill_all_qemus(nodedir)
         return True
 
     # make this a valid step
@@ -228,92 +298,191 @@ class TestPlc:
                 node.kill_qemu()
         return True
 
-    #################### step methods
-
-    ### uninstall
-    def uninstall_chroot(self):
-        self.run_in_host('service plc safestop')
-        #####detecting the last myplc version installed and remove it
-        self.run_in_host('rpm -e myplc')
-        ##### Clean up the /plc directory
-        self.run_in_host('rm -rf /plc/data')
-        ##### stop any running vservers
-        self.run_in_host('for vserver in $(ls -d /vservers/* | sed -e s,/vservers/,,) ; do case $vserver in vtest*) echo Shutting down vserver $vserver ; vserver $vserver stop ;; esac ; done')
+    #################### display config
+    def display (self):
+        "show test configuration after localization"
+        self.display_pass (1)
+        self.display_pass (2)
         return True
 
-    def uninstall_vserver(self):
-        self.run_in_host("vserver --silent %s delete"%self.vservername)
-        return True
+    # entry point
+    def display_pass (self,passno):
+        for (key,val) in self.plc_spec.iteritems():
+            if passno == 2:
+                if key == 'sites':
+                    for site in val:
+                        self.display_site_spec(site)
+                        for node in site['nodes']:
+                            self.display_node_spec(node)
+                elif key=='initscripts':
+                    for initscript in val:
+                        self.display_initscript_spec (initscript)
+                elif key=='slices':
+                    for slice in val:
+                        self.display_slice_spec (slice)
+                elif key=='keys':
+                    for key in val:
+                        self.display_key_spec (key)
+            elif passno == 1:
+                if key not in ['sites','initscripts','slices','keys']:
+                    print '+   ',key,':',val
 
-    def uninstall(self):
-        # if there's a chroot-based myplc running, and then a native-based myplc is being deployed
-        # it sounds safer to have the former uninstalled too
-        # now the vserver method cannot be invoked for chroot instances as vservername is required
-        if self.vserver:
-            self.uninstall_vserver()
-            self.uninstall_chroot()
-        else:
-            self.uninstall_chroot()
+    def display_site_spec (self,site):
+        print '+ ======== site',site['site_fields']['name']
+        for (k,v) in site.iteritems():
+            if k=='nodes':
+                if v: 
+                    print '+       ','nodes : ',
+                    for node in v:  
+                        print node['node_fields']['hostname'],'',
+                    print ''
+            elif k=='users':
+                if v: 
+                    print '+       users : ',
+                    for user in v:  
+                        print user['name'],'',
+                    print ''
+            elif k == 'site_fields':
+                print '+       login_base',':',v['login_base']
+            elif k == 'address_fields':
+                pass
+            else:
+                print '+       ',k,
+                PrettyPrinter(indent=8,depth=2).pprint(v)
+        
+    def display_initscript_spec (self,initscript):
+        print '+ ======== initscript',initscript['initscript_fields']['name']
+
+    def display_key_spec (self,key):
+        print '+ ======== key',key['name']
+
+    def display_slice_spec (self,slice):
+        print '+ ======== slice',slice['slice_fields']['name']
+        for (k,v) in slice.iteritems():
+            if k=='nodenames':
+                if v: 
+                    print '+       nodes : ',
+                    for nodename in v:  
+                        print nodename,'',
+                    print ''
+            elif k=='usernames':
+                if v: 
+                    print '+       users : ',
+                    for username in v:  
+                        print username,'',
+                    print ''
+            elif k=='slice_fields':
+                print '+       fields',':',
+                print 'max_nodes=',v['max_nodes'],
+                print ''
+            else:
+                print '+       ',k,v
+
+    def display_node_spec (self,node):
+        print "+           node",node['name'],"host_box=",node['host_box'],
+        print "hostname=",node['node_fields']['hostname'],
+        print "ip=",node['interface_fields']['ip']
+    
+
+    # another entry point for just showing the boxes involved
+    def display_mapping (self):
+        TestPlc.display_mapping_plc(self.plc_spec)
         return True
 
-    ### install
-    def install_chroot(self):
-        # nothing to do
+    @staticmethod
+    def display_mapping_plc (plc_spec):
+        print '+ MyPLC',plc_spec['name']
+        print '+\tvserver address = root@%s:/vservers/%s'%(plc_spec['hostname'],plc_spec['vservername'])
+        print '+\tIP = %s/%s'%(plc_spec['PLC_API_HOST'],plc_spec['vserverip'])
+        for site_spec in plc_spec['sites']:
+            for node_spec in site_spec['nodes']:
+                TestPlc.display_mapping_node(node_spec)
+
+    @staticmethod
+    def display_mapping_node (node_spec):
+        print '+   NODE %s'%(node_spec['name'])
+        print '+\tqemu box %s'%node_spec['host_box']
+        print '+\thostname=%s'%node_spec['node_fields']['hostname']
+
+    def local_pre (self):
+        "run site-dependant pre-test script as defined in LocalTestResources"
+        from LocalTestResources import local_resources
+        return local_resources.step_pre(self)
+    def local_post (self):
+        "run site-dependant post-test script as defined in LocalTestResources"
+        from LocalTestResources import local_resources
+        return local_resources.step_post(self)
+    def local_list (self):
+        "run site-dependant list script as defined in LocalTestResources"
+        from LocalTestResources import local_resources
+        return local_resources.step_list(self)
+    def local_cleanup (self):
+        "run site-dependant cleanup script as defined in LocalTestResources"
+        from LocalTestResources import local_resources
+        return local_resources.step_cleanup(self)
+    def delete(self):
+        "vserver delete the test myplc"
+        self.run_in_host("vserver --silent %s delete"%self.vservername)
         return True
 
-    def install_vserver(self):
-        # we need build dir for vtest-init-vserver
+    ### install
+    def create (self):
+        "vserver creation (no install done)"
         if self.is_local():
             # a full path for the local calls
-            build_dir=os.path(sys.argv[0])+"/build"
+            build_dir=os.path.dirname(sys.argv[0])
+            # sometimes this is empty - set to "." in such a case
+            if not build_dir: build_dir="."
+            build_dir += "/build"
         else:
-            # use a standard name - will be relative to HOME 
-            build_dir="options.buildname"
+            # use a standard name - will be relative to remote buildname
+            build_dir="build"
        # run checkout in any case - would do an update if already exists
         build_checkout = "svn checkout %s %s"%(self.options.build_url,build_dir)
         if self.run_in_host(build_checkout) != 0:
             return False
-        # the repo url is taken from myplc-url 
-        # with the last two steps (i386/myplc...) removed
-        repo_url = self.options.myplc_url
-        for level in [ 'rpmname','arch' ]:
+        # the repo url is taken from arch-rpms-url 
+        # with the last step (i386) removed
+        repo_url = self.options.arch_rpms_url
+        for level in [ 'arch' ]:
            repo_url = os.path.dirname(repo_url)
-        create_vserver="%s/vtest-init-vserver.sh %s %s -- --interface eth0:%s"%\
-            (build_dir,self.vservername,repo_url,self.vserverip)
+        # pass the vbuild-nightly options to vtest-init-vserver
+        test_env_options=""
+        test_env_options += " -p %s"%self.options.personality
+        test_env_options += " -d %s"%self.options.pldistro
+        test_env_options += " -f %s"%self.options.fcdistro
+        script="vtest-init-vserver.sh"
+        vserver_name = self.vservername
+        vserver_options="--netdev eth0 --interface %s"%self.vserverip
+        try:
+            vserver_hostname=socket.gethostbyaddr(self.vserverip)[0]
+            vserver_options += " --hostname %s"%vserver_hostname
+        except:
+            pass
+        create_vserver="%(build_dir)s/%(script)s %(test_env_options)s %(vserver_name)s %(repo_url)s -- %(vserver_options)s"%locals()
         return self.run_in_host(create_vserver) == 0
 
+    ### install_rpm 
     def install(self):
-        if self.vserver:
-            return self.install_vserver()
-        else:
-            return self.install_chroot()
-    
-    ### install_rpm - make this an optional step
-    def cache_rpm(self):
-        url = self.options.myplc_url
-        rpm = os.path.basename(url)
-        cache_fetch="pwd;if [ -f %(rpm)s ] ; then echo Using cached rpm %(rpm)s ; else echo Fetching %(url)s ; curl -O %(url)s; fi"%locals()
-       return self.run_in_host(cache_fetch)==0
-
-    def install_rpm_chroot(self):
-        url = self.options.myplc_url
-        rpm = os.path.basename(url)
-       if not self.cache_rpm():
-            return False
-       utils.header('Installing the :  %s'%rpm)
-        return self.run_in_host('rpm -Uvh '+rpm)==0 and self.run_in_host('service plc mount')==0
-
-    def install_rpm_vserver(self):
-        return self.run_in_guest("yum -y install myplc-native")==0
-
-    def install_rpm(self):
-        if self.vserver:
-            return self.install_rpm_vserver()
+        "yum install myplc, noderepo, and the plain bootstrapfs"
+        if self.options.personality == "linux32":
+            arch = "i386"
+        elif self.options.personality == "linux64":
+            arch = "x86_64"
         else:
-            return self.install_rpm_chroot()
+            raise Exception, "Unsupported personality %r"%self.options.personality
+        return \
+            self.run_in_guest("yum -y install myplc")==0 and \
+            self.run_in_guest("yum -y install noderepo-%s-%s"%(self.options.pldistro,arch))==0 and \
+            self.run_in_guest("yum -y install bootstrapfs-%s-%s-plain"%(self.options.pldistro,arch))==0 
 
     ### 
     def configure(self):
+        "run plc-config-tty"
         tmpname='%s.plc-config-tty'%(self.name())
         fileconf=open(tmpname,'w')
         for var in [ 'PLC_NAME',
@@ -322,6 +491,8 @@ class TestPlc:
                      'PLC_MAIL_ENABLED',
                      'PLC_MAIL_SUPPORT_ADDRESS',
                      'PLC_DB_HOST',
+                     'PLC_DB_PASSWORD',
+                    # Above line was added for integrating SFA Testing
                      'PLC_API_HOST',
                      'PLC_WWW_HOST',
                      'PLC_BOOT_HOST',
@@ -336,23 +507,23 @@ class TestPlc:
         utils.system('rm %s'%tmpname)
         return True
 
-    # the chroot install is slightly different to this respect
     def start(self):
-        if self.vserver:
-            self.run_in_guest('service plc start')
-        else:
-            self.run_in_host('service plc start')
+        "service plc start"
+        self.run_in_guest('service plc start')
         return True
-        
+
     def stop(self):
-        if self.vserver:
-            self.run_in_guest('service plc stop')
-        else:
-            self.run_in_host('service plc stop')
+        "service plc stop"
+        self.run_in_guest('service plc stop')
         return True
         
-    # could use a TestKey class
+    def vs_start (self):
+        self.start_guest()
+        return True
+
+    # stores the keys from the config for further use
     def store_keys(self):
+        "stores test users ssh keys in keys/"
         for key_spec in self.plc_spec['keys']:
                TestKey(self,key_spec).store_key()
         return True
@@ -360,10 +531,28 @@ class TestPlc:
     def clean_keys(self):
         utils.system("rm -rf %s/keys/"%os.path(sys.argv[0]))
 
+    # fetches the ssh keys in the plc's /etc/planetlab and stores them in keys/
+    # for later direct access to the nodes
+    def fetch_keys(self):
+        "gets ssh keys in /etc/planetlab/ and stores them locally in keys/"
+        dir="./keys"
+        if not os.path.isdir(dir):
+            os.mkdir(dir)
+        vservername=self.vservername
+        overall=True
+        prefix = 'debug_ssh_key'
+        for ext in [ 'pub', 'rsa' ] :
+            src="/vservers/%(vservername)s/etc/planetlab/%(prefix)s.%(ext)s"%locals()
+            dst="keys/%(vservername)s-debug.%(ext)s"%locals()
+            if self.test_ssh.fetch(src,dst) != 0: overall=False
+        return overall
+
     def sites (self):
+        "create sites with PLCAPI"
         return self.do_sites()
     
     def clean_sites (self):
+        "delete sites with PLCAPI"
         return self.do_sites(action="delete")
     
     def do_sites (self,action="add"):
@@ -381,9 +570,18 @@ class TestPlc:
                 test_site.create_users()
         return True
 
+    def clean_all_sites (self):
+        print 'auth_root',self.auth_root()
+        site_ids = [s['site_id'] for s in self.apiserver.GetSites(self.auth_root(), {}, ['site_id'])]
+        for site_id in site_ids:
+            print 'Deleting site_id',site_id
+            self.apiserver.DeleteSite(self.auth_root(),site_id)
+
     def nodes (self):
+        "create nodes with PLCAPI"
         return self.do_nodes()
     def clean_nodes (self):
+        "delete nodes with PLCAPI"
         return self.do_nodes(action="delete")
 
     def do_nodes (self,action="add"):
@@ -403,9 +601,15 @@ class TestPlc:
                     test_node.create_node ()
         return True
 
-    # create nodegroups if needed, and populate
-    # no need for a clean_nodegroups if we are careful enough
     def nodegroups (self):
+        "create nodegroups with PLCAPI"
+        return self.do_nodegroups("add")
+    def clean_nodegroups (self):
+        "delete nodegroups with PLCAPI"
+        return self.do_nodegroups("delete")
+
+    # create nodegroups if needed, and populate
+    def do_nodegroups (self, action="add"):
         # 1st pass to scan contents
         groups_dict = {}
         for site_spec in self.plc_spec['sites']:
@@ -421,14 +625,54 @@ class TestPlc:
                             groups_dict[nodegroupname]=[]
                         groups_dict[nodegroupname].append(test_node.name())
         auth=self.auth_root()
+        overall = True
         for (nodegroupname,group_nodes) in groups_dict.iteritems():
-            try:
-                self.apiserver.GetNodeGroups(auth,{'name':nodegroupname})[0]
-            except:
-                self.apiserver.AddNodeGroup(auth,{'name':nodegroupname})
-            for node in group_nodes:
-                self.apiserver.AddNodeToNodeGroup(auth,node,nodegroupname)
-        return True
+            if action == "add":
+                print 'nodegroups:','dealing with nodegroup',nodegroupname,'on nodes',group_nodes
+                # first, check if the nodetagtype is here
+                tag_types = self.apiserver.GetTagTypes(auth,{'tagname':nodegroupname})
+                if tag_types:
+                    tag_type_id = tag_types[0]['tag_type_id']
+                else:
+                    tag_type_id = self.apiserver.AddTagType(auth,
+                                                            {'tagname':nodegroupname,
+                                                             'description': 'for nodegroup %s'%nodegroupname,
+                                                             'category':'test',
+                                                             'min_role_id':10})
+                print 'located tag (type)',nodegroupname,'as',tag_type_id
+                # create nodegroup
+                nodegroups = self.apiserver.GetNodeGroups (auth, {'groupname':nodegroupname})
+                if not nodegroups:
+                    self.apiserver.AddNodeGroup(auth, nodegroupname, tag_type_id, 'yes')
+                    print 'created nodegroup',nodegroupname,'from tagname',nodegroupname,'and value','yes'
+                # set node tag on all nodes, value='yes'
+                for nodename in group_nodes:
+                    try:
+                        self.apiserver.AddNodeTag(auth, nodename, nodegroupname, "yes")
+                    except:
+                        traceback.print_exc()
+                        print 'node',nodename,'seems to already have tag',nodegroupname
+                    # check anyway
+                    try:
+                        expect_yes = self.apiserver.GetNodeTags(auth,
+                                                                {'hostname':nodename,
+                                                                 'tagname':nodegroupname},
+                                                                ['value'])[0]['value']
+                        if expect_yes != "yes":
+                            print 'Mismatch node tag on node',nodename,'got',expect_yes
+                            overall=False
+                    except:
+                        if not self.options.dry_run:
+                            print 'Cannot find tag',nodegroupname,'on node',nodename
+                            overall = False
+            else:
+                try:
+                    print 'cleaning nodegroup',nodegroupname
+                    self.apiserver.DeleteNodeGroup(auth,nodegroupname)
+                except:
+                    traceback.print_exc()
+                    overall=False
+        return overall
 
     def all_hostnames (self) :
         hostnames = []
@@ -437,14 +681,14 @@ class TestPlc:
                            for node_spec in site_spec['nodes'] ]
         return hostnames
 
-    # gracetime : during the first <gracetime> minutes nothing gets printed
-    def do_nodes_booted (self, minutes, gracetime=2):
+    # silent_minutes : during the first <silent_minutes> minutes nothing gets printed
+    def nodes_check_boot_state (self, target_boot_state, timeout_minutes, silent_minutes,period=15):
         if self.options.dry_run:
             print 'dry_run'
             return True
         # compute timeout
-        timeout = datetime.datetime.now()+datetime.timedelta(minutes=minutes)
-        graceout = datetime.datetime.now()+datetime.timedelta(minutes=gracetime)
+        timeout = datetime.datetime.now()+datetime.timedelta(minutes=timeout_minutes)
+        graceout = datetime.datetime.now()+datetime.timedelta(minutes=silent_minutes)
         # the nodes that haven't checked yet - start with a full list and shrink over time
         tocheck = self.all_hostnames()
         utils.header("checking nodes %r"%tocheck)
@@ -457,21 +701,21 @@ class TestPlc:
             for array in tocheck_status:
                 hostname=array['hostname']
                 boot_state=array['boot_state']
-                if boot_state == 'boot':
-                    utils.header ("%s has reached the 'boot' state"%hostname)
+                if boot_state == target_boot_state:
+                    utils.header ("%s has reached the %s state"%(hostname,target_boot_state))
                 else:
                     # if it's a real node, never mind
                     (site_spec,node_spec)=self.locate_hostname(hostname)
                     if TestNode.is_real_model(node_spec['node_fields']['model']):
                         utils.header("WARNING - Real node %s in %s - ignored"%(hostname,boot_state))
                         # let's cheat
-                        boot_state = 'boot'
-                    if datetime.datetime.now() > graceout:
+                        boot_state = target_boot_state
+                    elif datetime.datetime.now() > graceout:
                         utils.header ("%s still in '%s' state"%(hostname,boot_state))
                         graceout=datetime.datetime.now()+datetime.timedelta(1)
                 status[hostname] = boot_state
             # refresh tocheck
-            tocheck = [ hostname for (hostname,boot_state) in status.iteritems() if boot_state != 'boot' ]
+            tocheck = [ hostname for (hostname,boot_state) in status.iteritems() if boot_state != target_boot_state ]
             if not tocheck:
                 return True
             if datetime.datetime.now() > timeout:
@@ -479,27 +723,47 @@ class TestPlc:
                     utils.header("FAILURE due to %s in '%s' state"%(hostname,status[hostname]))
                 return False
             # otherwise, sleep for a while
-            time.sleep(15)
+            time.sleep(period)
         # only useful in empty plcs
         return True
 
     def nodes_booted(self):
-        return self.do_nodes_booted(minutes=0)
-    
+        return self.nodes_check_boot_state('boot',timeout_minutes=20,silent_minutes=15)
 
-    def do_nodes_ssh(self,minutes):
+    def check_nodes_ssh(self,debug,timeout_minutes,silent_minutes,period=20):
         # compute timeout
-        timeout = datetime.datetime.now()+datetime.timedelta(minutes=minutes)
+        timeout = datetime.datetime.now()+datetime.timedelta(minutes=timeout_minutes)
+        graceout = datetime.datetime.now()+datetime.timedelta(minutes=silent_minutes)
+        vservername=self.vservername
+        if debug: 
+            message="debug"
+            local_key = "keys/%(vservername)s-debug.rsa"%locals()
+        else: 
+            message="boot"
+           local_key = "keys/key1.rsa"
         tocheck = self.all_hostnames()
-#        self.scan_publicKeys(tocheck)
-        utils.header("checking Connectivity on nodes %r"%tocheck)
+        utils.header("checking ssh access (expected in %s mode) to nodes %r"%(message,tocheck))
+        utils.header("max timeout is %d minutes, silent for %d minutes (period is %s)"%\
+                         (timeout_minutes,silent_minutes,period))
         while tocheck:
             for hostname in tocheck:
-                # try to ssh in nodes
-                node_test_ssh = TestSsh (hostname,key="/etc/planetlab/root_ssh_key.rsa")
-                access=self.run_in_guest(node_test_ssh.actual_command("date"))
-                if not access:
-                    utils.header('The node %s is sshable -->'%hostname)
+                # try to run 'hostname' in the node
+                command = TestSsh (hostname,key=local_key).actual_command("hostname;uname -a")
+                # don't spam logs - show the command only after the grace period 
+                if datetime.datetime.now() > graceout:
+                    success=utils.system(command)
+                else:
+                    # truly silent, just print out a dot to show we're alive
+                    print '.',
+                    sys.stdout.flush()
+                    command += " 2>/dev/null"
+                    if self.options.dry_run:
+                        print 'dry_run',command
+                        success=0
+                    else:
+                        success=os.system(command)
+                if success==0:
+                    utils.header('Successfully entered root@%s (%s)'%(hostname,message))
                     # refresh tocheck
                     tocheck.remove(hostname)
                 else:
@@ -515,22 +779,50 @@ class TestPlc:
                     utils.header("FAILURE to ssh into %s"%hostname)
                 return False
             # otherwise, sleep for a while
-            time.sleep(15)
+            time.sleep(period)
         # only useful in empty plcs
         return True
         
-    def nodes_ssh(self):
-        return self.do_nodes_ssh(minutes=2)
+    def nodes_ssh_debug(self):
+        "Tries to ssh into nodes in debug mode with the debug ssh key"
+        return self.check_nodes_ssh(debug=True,timeout_minutes=30,silent_minutes=10)
     
+    def nodes_ssh_boot(self):
+        "Tries to ssh into nodes in production mode with the root ssh key"
+        return self.check_nodes_ssh(debug=False,timeout_minutes=30,silent_minutes=10)
+    
+    @node_mapper
+    def init_node (self): 
+        "all nodes : init a clean local directory for holding node-dep stuff like iso image..."
+        pass
     @node_mapper
-    def init_node (self): pass
+    def bootcd (self): 
+        "all nodes: invoke GetBootMedium and store result locally"
+        pass
     @node_mapper
-    def bootcd (self): pass
+    def configure_qemu (self): 
+        "all nodes: compute qemu config qemu.conf and store it locally"
+        pass
     @node_mapper
-    def configure_qemu (self): pass
+    def reinstall_node (self): 
+        "all nodes: mark PLCAPI boot_state as reinstall"
+        pass
     @node_mapper
-    def reinstall_node (self): pass
+    def export_qemu (self): 
+        "all nodes: push local node-dep directory on the qemu box"
+        pass
         
+    ### check hooks : invoke scripts from hooks/{node,slice}
+    def check_hooks_node (self): 
+        return self.locate_first_node().check_hooks()
+    def check_hooks_sliver (self) : 
+        return self.locate_first_sliver().check_hooks()
+    
+    def check_hooks (self):
+        "runs unit tests in the node and slice contexts - see hooks/{node,slice}"
+        return self.check_hooks_node() and self.check_hooks_sliver()
+
+    ### initscripts
     def do_check_initscripts(self):
         overall = True
         for slice_spec in self.plc_spec['slices']:
@@ -549,18 +841,35 @@ class TestPlc:
         return overall
            
     def check_initscripts(self):
-           return self.do_check_initscripts()
-                   
+        "check that the initscripts have triggered"
+        return self.do_check_initscripts()
+    
     def initscripts (self):
+        "create initscripts with PLCAPI"
         for initscript in self.plc_spec['initscripts']:
             utils.pprint('Adding Initscript in plc %s'%self.plc_spec['name'],initscript)
             self.apiserver.AddInitScript(self.auth_root(),initscript['initscript_fields'])
         return True
 
+    def clean_initscripts (self):
+        "delete initscripts with PLCAPI"
+        for initscript in self.plc_spec['initscripts']:
+            initscript_name = initscript['initscript_fields']['name']
+            print('Attempting to delete %s in plc %s'%(initscript_name,self.plc_spec['name']))
+            try:
+                self.apiserver.DeleteInitScript(self.auth_root(),initscript_name)
+                print initscript_name,'deleted'
+            except:
+                print 'deletion went wrong - probably did not exist'
+        return True
+
+    ### manage slices
     def slices (self):
+        "create slices with PLCAPI"
         return self.do_slices()
 
     def clean_slices (self):
+        "delete slices with PLCAPI"
         return self.do_slices("delete")
 
     def do_slices (self,  action="add"):
@@ -578,31 +887,22 @@ class TestPlc:
         return True
         
     @slice_mapper_options
-    def check_slice(self): pass
+    def check_slice(self): 
+        "tries to ssh-enter the slice with the user key, to ensure slice creation"
+        pass
 
     @node_mapper
-    def clear_known_hosts (self): pass
+    def clear_known_hosts (self): 
+        "remove test nodes entries from the local known_hosts file"
+        pass
     
     @node_mapper
-    def start_node (self) : pass
-
-    def locate_first_sliver (self):
-        slice_spec = self.plc_spec['slices'][0]
-        slicename = slice_spec['slice_fields']['name']
-        nodename = slice_spec['nodenames'][0]
-        return self.locate_sliver_obj(nodename,slicename)
-
-    def locate_sliver_obj (self,nodename,slicename):
-        (site,node) = self.locate_node(nodename)
-        slice = self.locate_slice (slicename)
-        # build objects
-        test_site = TestSite (self, site)
-        test_node = TestNode (self, test_site,node)
-        # xxx the slice site is assumed to be the node site - mhh - probably harmless
-        test_slice = TestSlice (self, test_site, slice)
-        return TestSliver (self, test_node, test_slice)
+    def start_node (self) : 
+        "all nodes: start the qemu instance (also runs qemu-bridge-init start)"
+        pass
 
     def check_tcp (self):
+        "check TCP connectivity between 2 slices (or in loopback if only one is defined)"
         specs = self.plc_spec['tcp_test']
         overall=True
         for spec in specs:
@@ -618,16 +918,215 @@ class TestPlc:
             if not c_test_sliver.run_tcp_client(s_test_sliver.test_node.name(),port):
                 overall=False
         return overall
-    
+
+    def plcsh_stress_test (self):
+        "runs PLCAPI stress test, that checks Add/Update/Delete on all types - preserves contents"
+        # install the stress-test in the plc image
+        location = "/usr/share/plc_api/plcsh_stress_test.py"
+        remote="/vservers/%s/%s"%(self.vservername,location)
+        self.test_ssh.copy_abs("plcsh_stress_test.py",remote)
+        command = location
+        command += " -- --check"
+        if self.options.size == 1:
+            command +=  " --tiny"
+        return ( self.run_in_guest(command) == 0)
+
+    # populate runs the same utility without slightly different options
+    # in particular runs with --preserve (dont cleanup) and without --check
+    # also it gets run twice, once with the --foreign option for creating fake foreign entries
+
+    ### install_sfa_rpm
+    def install_sfa(self):
+        "yum install sfa, sfa-plc and sfa-client"
+        if self.options.personality == "linux32":
+            arch = "i386"
+        elif self.options.personality == "linux64":
+            arch = "x86_64"
+        else:
+            raise Exception, "Unsupported personality %r"%self.options.personality
+        return \
+            self.run_in_guest("yum -y install sfa")==0 and \
+            self.run_in_guest("yum -y install sfa-client")==0 and \
+            self.run_in_guest("yum -y install sfa-plc")==0 and \
+            self.run_in_guest("yum -y install sfa-sfatables")==0
+    ###
+    def configure_sfa(self):
+        "run sfa-config-tty"
+        tmpname='%s.sfa-config-tty'%(self.name())
+        fileconf=open(tmpname,'w')
+        fileconf.write ('u\n')
+        for var in [ 'SFA_REGISTRY_ROOT_AUTH',
+                     'SFA_REGISTRY_LEVEL1_AUTH',
+                     'SFA_PLC_USER',
+                     'SFA_PLC_PASSWORD',
+                     'SFA_PLC_DB_HOST',
+                     'SFA_PLC_DB_USER',
+                     'SFA_PLC_DB_PASSWORD']:
+            fileconf.write ('%s\n'%(self.plc_spec['sfa'][var]))
+        fileconf.write('w\n')
+        fileconf.write('q\n')
+        fileconf.close()
+        utils.system('cat %s'%tmpname)
+        self.run_in_guest_piped('cat %s'%tmpname,'sfa-config-tty')
+        utils.system('rm %s'%tmpname)
+        return True
+
+    def import_sfa(self):
+        "sfa-import-plc"
+       auth=self.plc_spec['sfa']['SFA_REGISTRY_ROOT_AUTH']
+        self.run_in_guest('sfa-import-plc.py')
+        self.run_in_guest('cp /etc/sfa/authorities/%s/%s.pkey /etc/sfa/authorities/server.key'%(auth,auth))
+        return True
+
+    def start_sfa(self):
+        "service sfa start"
+        self.run_in_guest('service sfa start')
+        return True
+
+    def setup_sfa(self):
+        "sfi client configuration"
+       dir_name=".sfi"
+       if os.path.exists(dir_name):
+           utils.system('rm -rf %s'%dir_name)
+       utils.system('mkdir %s'%dir_name)
+       file_name=dir_name + os.sep + 'fake-pi1.pkey'
+        fileconf=open(file_name,'w')
+        fileconf.write (self.plc_spec['keys'][0]['private'])
+        fileconf.close()
+
+       file_name=dir_name + os.sep + 'sfi_config'
+        fileconf=open(file_name,'w')
+       SFI_AUTH=self.plc_spec['sfa']['SFA_REGISTRY_ROOT_AUTH']+".main"
+        fileconf.write ("SFI_AUTH='%s'"%SFI_AUTH)
+       fileconf.write('\n')
+       SFI_USER=SFI_AUTH+'.fake-pi1'
+        fileconf.write ("SFI_USER='%s'"%SFI_USER)
+       fileconf.write('\n')
+        fileconf.write ("SFI_REGISTRY='http://localhost:12345/'")
+       fileconf.write('\n')
+        fileconf.write ("SFI_SM='http://localhost:12347/'")
+       fileconf.write('\n')
+        fileconf.close()
+
+       file_name=dir_name + os.sep + 'person.xml'
+        fileconf=open(file_name,'w')
+       for record in self.plc_spec['sfa']['sfa_person_xml']:
+          person_record=record
+       fileconf.write(person_record)
+       fileconf.write('\n')
+        fileconf.close()
+
+       file_name=dir_name + os.sep + 'slice.xml'
+        fileconf=open(file_name,'w')
+       for record in self.plc_spec['sfa']['sfa_slice_xml']:
+           slice_record=record
+       #slice_record=self.plc_spec['sfa']['sfa_slice_xml']
+       fileconf.write(slice_record)
+       fileconf.write('\n')
+        fileconf.close()
+
+       file_name=dir_name + os.sep + 'slice.rspec'
+        fileconf=open(file_name,'w')
+       slice_rspec=''
+       for (key, value) in self.plc_spec['sfa']['sfa_slice_rspec'].items():
+           slice_rspec +=value 
+       fileconf.write(slice_rspec)
+       fileconf.write('\n')
+        fileconf.close()
+        location = "root/"
+        remote="/vservers/%s/%s"%(self.vservername,location)
+       self.test_ssh.copy_abs(dir_name, remote, recursive=True)
+
+        #utils.system('cat %s'%tmpname)
+        utils.system('rm -rf %s'%dir_name)
+        return True
+
+    def add_sfa(self):
+        "run sfi.py add (on Registry) and sfi.py create (on SM) to form new objects"
+       test_plc=self
+        test_user_sfa=TestUserSfa(test_plc,self.plc_spec['sfa'])
+        success=test_user_sfa.add_user()
+
+       for slice_spec in self.plc_spec['sfa']['slices_sfa']:
+            site_spec = self.locate_site (slice_spec['sitename'])
+            test_site = TestSite(self,site_spec)
+           test_slice_sfa=TestSliceSfa(test_plc,test_site,slice_spec)
+           success1=test_slice_sfa.add_slice()
+           success2=test_slice_sfa.create_slice()
+       return success and success1 and success2
+
+    def update_sfa(self):
+        "run sfi.py update (on Registry) and sfi.py create (on SM) on existing objects"
+       test_plc=self
+       test_user_sfa=TestUserSfa(test_plc,self.plc_spec['sfa'])
+       success1=test_user_sfa.update_user()
+       
+       for slice_spec in self.plc_spec['sfa']['slices_sfa']:
+           site_spec = self.locate_site (slice_spec['sitename'])
+           test_site = TestSite(self,site_spec)
+           test_slice_sfa=TestSliceSfa(test_plc,test_site,slice_spec)
+           success2=test_slice_sfa.update_slice()
+       return success1 and success2
+
+    def view_sfa(self):
+        "run sfi.py list and sfi.py show (both on Registry) and sfi.py slices and sfi.py resources (both on SM)"
+       auth=self.plc_spec['sfa']['SFA_REGISTRY_ROOT_AUTH']
+       return \
+       self.run_in_guest("sfi.py -d /root/.sfi/ list %s.main"%auth)==0 and \
+       self.run_in_guest("sfi.py -d /root/.sfi/ show %s.main"%auth)==0 and \
+       self.run_in_guest("sfi.py -d /root/.sfi/ slices")==0 and \
+       self.run_in_guest("sfi.py -d /root/.sfi/ resources")==0
+
+    @slice_mapper_options_sfa
+    def check_slice_sfa(self): 
+       "tries to ssh-enter the SFA slice"
+        pass
+
+    def delete_sfa(self):
+       "run sfi.py delete (on SM), sfi.py remove (on Registry)"
+       test_plc=self
+       test_user_sfa=TestUserSfa(test_plc,self.plc_spec['sfa'])
+       success1=test_user_sfa.delete_user()
+       for slice_spec in self.plc_spec['sfa']['slices_sfa']:
+            site_spec = self.locate_site (slice_spec['sitename'])
+            test_site = TestSite(self,site_spec)
+           test_slice_sfa=TestSliceSfa(test_plc,test_site,slice_spec)
+           success2=test_slice_sfa.delete_slice()
+
+       return success1 and success2
+
+    def stop_sfa(self):
+        "service sfa stop"
+        self.run_in_guest('service sfa stop')
+        return True
+
+    def populate (self):
+        "creates random entries in the PLCAPI"
+        # install the stress-test in the plc image
+        location = "/usr/share/plc_api/plcsh_stress_test.py"
+        remote="/vservers/%s/%s"%(self.vservername,location)
+        self.test_ssh.copy_abs("plcsh_stress_test.py",remote)
+        command = location
+        command += " -- --preserve --short-names"
+        local = (self.run_in_guest(command) == 0);
+        # second run with --foreign
+        command += ' --foreign'
+        remote = (self.run_in_guest(command) == 0);
+        return ( local and remote)
 
     def gather_logs (self):
-        # (1) get the plc's /var/log and store it locally in logs/<plcname>-var-log/*
-        # (2) get all the nodes qemu log and store it as logs/<node>-qemu.log
-        # (3) get the nodes /var/log and store is as logs/<node>-var-log/*
-        # (4) as far as possible get the slice's /var/log as logs/<slice>-<node>-var-log/*
-        # (1)
+        "gets all possible logs from plc's/qemu node's/slice's for future reference"
+        # (1.a) get the plc's /var/log/ and store it locally in logs/myplc.var-log.<plcname>/*
+        # (1.b) get the plc's  /var/lib/pgsql/data/pg_log/ -> logs/myplc.pgsql-log.<plcname>/*
+        # (2) get all the nodes qemu log and store it as logs/node.qemu.<node>.log
+        # (3) get the nodes /var/log and store is as logs/node.var-log.<node>/*
+        # (4) as far as possible get the slice's /var/log as logs/sliver.var-log.<sliver>/*
+        # (1.a)
         print "-------------------- TestPlc.gather_logs : PLC's /var/log"
         self.gather_var_logs ()
+        # (1.b)
+        print "-------------------- TestPlc.gather_logs : PLC's /var/lib/psql/data/pg_log/"
+        self.gather_pgsql_logs ()
         # (2) 
         print "-------------------- TestPlc.gather_logs : nodes's QEMU logs"
         for site_spec in self.plc_spec['sites']:
@@ -640,24 +1139,29 @@ class TestPlc:
         self.gather_nodes_var_logs()
         # (4)
         print "-------------------- TestPlc.gather_logs : sample sliver's /var/log"
-        self.gather_first_sliver_logs()
+        self.gather_slivers_var_logs()
         return True
 
-    def gather_first_sliver_logs(self):
-        try:
-            test_sliver = self.locate_first_sliver()
+    def gather_slivers_var_logs(self):
+        for test_sliver in self.all_sliver_objs():
             remote = test_sliver.tar_var_logs()
-            utils.system("mkdir -p logs/%s-var-log"%test_sliver.name())
-            command = remote + " | tar -C logs/%s-var-log -xf -"%test_sliver.name()
+            utils.system("mkdir -p logs/sliver.var-log.%s"%test_sliver.name())
+            command = remote + " | tar -C logs/sliver.var-log.%s -xf -"%test_sliver.name()
             utils.system(command)
-        except Exception,e:
-            print 'Cannot locate first sliver - giving up',e
         return True
 
     def gather_var_logs (self):
+        utils.system("mkdir -p logs/myplc.var-log.%s"%self.name())
         to_plc = self.actual_command_in_guest("tar -C /var/log/ -cf - .")        
-        command = to_plc + "| tar -C logs/%s-var-log -xf -"%self.name()
-        utils.system("mkdir -p logs/%s-var-log"%self.name())
+        command = to_plc + "| tar -C logs/myplc.var-log.%s -xf -"%self.name()
+        utils.system(command)
+        command = "chmod a+r,a+x logs/myplc.var-log.%s/httpd"%self.name()
+        utils.system(command)
+
+    def gather_pgsql_logs (self):
+        utils.system("mkdir -p logs/myplc.pgsql-log.%s"%self.name())
+        to_plc = self.actual_command_in_guest("tar -C /var/lib/pgsql/data/pg_log/ -cf - .")        
+        command = to_plc + "| tar -C logs/myplc.pgsql-log.%s -xf -"%self.name()
         utils.system(command)
 
     def gather_nodes_var_logs (self):
@@ -665,10 +1169,10 @@ class TestPlc:
             test_site = TestSite (self,site_spec)
             for node_spec in site_spec['nodes']:
                 test_node=TestNode(self,test_site,node_spec)
-                test_ssh = TestSsh (test_node.name(),key="/etc/planetlab/root_ssh_key.rsa")
-                to_plc = self.actual_command_in_guest ( test_ssh.actual_command("tar -C /var/log -cf - ."))
-                command = to_plc + "| tar -C logs/%s-var-log -xf -"%test_node.name()
-                utils.system("mkdir -p logs/%s-var-log"%test_node.name())
+                test_ssh = TestSsh (test_node.name(),key="keys/key1.rsa")
+                command = test_ssh.actual_command("tar -C /var/log -cf - .")
+                command = command + "| tar -C logs/node.var-log.%s -xf -"%test_node.name()
+                utils.system("mkdir -p logs/node.var-log.%s"%test_node.name())
                 utils.system(command)
 
 
@@ -744,4 +1248,3 @@ class TestPlc:
     def standby_19(): pass
     @standby_generic 
     def standby_20(): pass
-