7 from types import StringTypes
11 from TestSite import TestSite
12 from TestNode import TestNode
13 from TestUser import TestUser
14 from TestKey import TestKey
15 from TestSlice import TestSlice
16 from TestSliver import TestSliver
17 from TestBox import TestBox
18 from TestSsh import TestSsh
19 from TestApiserver import TestApiserver
20 from TestSliceSfa import TestSliceSfa
21 from TestUserSfa import TestUserSfa
23 # step methods must take (self) and return a boolean (options is a member of the class)
25 def standby(minutes,dry_run):
26 utils.header('Entering StandBy for %d mn'%minutes)
30 time.sleep(60*minutes)
33 def standby_generic (func):
35 minutes=int(func.__name__.split("_")[1])
36 return standby(minutes,self.options.dry_run)
39 def node_mapper (method):
42 node_method = TestNode.__dict__[method.__name__]
43 for site_spec in self.plc_spec['sites']:
44 test_site = TestSite (self,site_spec)
45 for node_spec in site_spec['nodes']:
46 test_node = TestNode (self,test_site,node_spec)
47 if not node_method(test_node): overall=False
49 # restore the doc text
50 actual.__doc__=method.__doc__
53 def slice_mapper_options (method):
56 slice_method = TestSlice.__dict__[method.__name__]
57 for slice_spec in self.plc_spec['slices']:
58 site_spec = self.locate_site (slice_spec['sitename'])
59 test_site = TestSite(self,site_spec)
60 test_slice=TestSlice(self,test_site,slice_spec)
61 if not slice_method(test_slice,self.options): overall=False
63 # restore the doc text
64 actual.__doc__=method.__doc__
67 def slice_mapper_options_sfa (method):
71 slice_method = TestSliceSfa.__dict__[method.__name__]
72 for slice_spec in self.plc_spec['sfa']['slices_sfa']:
73 site_spec = self.locate_site (slice_spec['sitename'])
74 test_site = TestSite(self,site_spec)
75 test_slice=TestSliceSfa(test_plc,test_site,slice_spec)
76 if not slice_method(test_slice,self.options): overall=False
78 # restore the doc text
79 actual.__doc__=method.__doc__
87 'display', 'resources_pre', SEP,
88 'delete_vs','create_vs','install', 'configure', 'start', SEP,
89 'fetch_keys', 'store_keys', 'clear_known_hosts', SEP,
90 'initscripts', 'sites', 'nodes', 'slices', 'nodegroups', SEP,
91 'reinstall_node', 'init_node','bootcd', 'configure_qemu', 'export_qemu',
92 'kill_all_qemus', 'start_node', SEP,
93 # better use of time: do this now that the nodes are taking off
94 'plcsh_stress_test', SEP,
95 'install_sfa', 'configure_sfa', 'import_sfa', 'start_sfa', SEP,
96 'setup_sfa', 'add_sfa', 'update_sfa', 'view_sfa', SEP,
97 'nodes_ssh_debug', 'nodes_ssh_boot', 'check_slice', 'check_initscripts', SEP,
98 # optionally run sfa later; takes longer, but checks more about nm
99 # 'install_sfa', 'configure_sfa', 'import_sfa', 'start_sfa', SEP,
100 # 'setup_sfa', 'add_sfa', 'update_sfa', 'view_sfa', SEP,
101 'check_slice_sfa', 'delete_sfa', 'stop_sfa', SEP,
102 'check_tcp', 'check_hooks', SEP,
103 'force_gather_logs', 'force_resources_post',
106 'show_boxes', 'resources_list','resources_release','resources_release_plc','resources_release_qemu',SEP,
107 'stop', 'vs_start', SEP,
108 'clean_initscripts', 'clean_nodegroups','clean_all_sites', SEP,
109 'clean_sites', 'clean_nodes', 'clean_slices', 'clean_keys', SEP,
111 'list_all_qemus', 'list_qemus', 'kill_qemus', SEP,
112 'db_dump' , 'db_restore', SEP,
113 'standby_1 through 20',
117 def printable_steps (list):
118 return " ".join(list).replace(" "+SEP+" "," \\\n")
120 def valid_step (step):
123 def __init__ (self,plc_spec,options):
124 self.plc_spec=plc_spec
126 self.test_ssh=TestSsh(self.plc_spec['hostname'],self.options.buildname)
128 self.vserverip=plc_spec['vserverip']
129 self.vservername=plc_spec['vservername']
130 self.url="https://%s:443/PLCAPI/"%plc_spec['vserverip']
133 raise Exception,'chroot-based myplc testing is deprecated'
134 self.apiserver=TestApiserver(self.url,options.dry_run)
137 name=self.plc_spec['name']
138 return "%s.%s"%(name,self.vservername)
141 return self.plc_spec['hostname']
144 return self.test_ssh.is_local()
146 # define the API methods on this object through xmlrpc
147 # would help, but not strictly necessary
151 def actual_command_in_guest (self,command):
152 return self.test_ssh.actual_command(self.host_to_guest(command))
154 def start_guest (self):
155 return utils.system(self.test_ssh.actual_command(self.start_guest_in_host()))
157 def run_in_guest (self,command):
158 return utils.system(self.actual_command_in_guest(command))
160 def run_in_host (self,command):
161 return self.test_ssh.run_in_buildname(command)
163 #command gets run in the vserver
164 def host_to_guest(self,command):
165 return "vserver %s exec %s"%(self.vservername,command)
167 #command gets run in the vserver
168 def start_guest_in_host(self):
169 return "vserver %s start"%(self.vservername)
172 def run_in_guest_piped (self,local,remote):
173 return utils.system(local+" | "+self.test_ssh.actual_command(self.host_to_guest(remote),keep_stdin=True))
175 def auth_root (self):
176 return {'Username':self.plc_spec['PLC_ROOT_USER'],
177 'AuthMethod':'password',
178 'AuthString':self.plc_spec['PLC_ROOT_PASSWORD'],
179 'Role' : self.plc_spec['role']
181 def locate_site (self,sitename):
182 for site in self.plc_spec['sites']:
183 if site['site_fields']['name'] == sitename:
185 if site['site_fields']['login_base'] == sitename:
187 raise Exception,"Cannot locate site %s"%sitename
189 def locate_node (self,nodename):
190 for site in self.plc_spec['sites']:
191 for node in site['nodes']:
192 if node['name'] == nodename:
194 raise Exception,"Cannot locate node %s"%nodename
196 def locate_hostname (self,hostname):
197 for site in self.plc_spec['sites']:
198 for node in site['nodes']:
199 if node['node_fields']['hostname'] == hostname:
201 raise Exception,"Cannot locate hostname %s"%hostname
203 def locate_key (self,keyname):
204 for key in self.plc_spec['keys']:
205 if key['name'] == keyname:
207 raise Exception,"Cannot locate key %s"%keyname
209 def locate_slice (self, slicename):
210 for slice in self.plc_spec['slices']:
211 if slice['slice_fields']['name'] == slicename:
213 raise Exception,"Cannot locate slice %s"%slicename
215 def all_sliver_objs (self):
217 for slice_spec in self.plc_spec['slices']:
218 slicename = slice_spec['slice_fields']['name']
219 for nodename in slice_spec['nodenames']:
220 result.append(self.locate_sliver_obj (nodename,slicename))
223 def locate_sliver_obj (self,nodename,slicename):
224 (site,node) = self.locate_node(nodename)
225 slice = self.locate_slice (slicename)
227 test_site = TestSite (self, site)
228 test_node = TestNode (self, test_site,node)
229 # xxx the slice site is assumed to be the node site - mhh - probably harmless
230 test_slice = TestSlice (self, test_site, slice)
231 return TestSliver (self, test_node, test_slice)
233 def locate_first_node(self):
234 nodename=self.plc_spec['slices'][0]['nodenames'][0]
235 (site,node) = self.locate_node(nodename)
236 test_site = TestSite (self, site)
237 test_node = TestNode (self, test_site,node)
240 def locate_first_sliver (self):
241 slice_spec=self.plc_spec['slices'][0]
242 slicename=slice_spec['slice_fields']['name']
243 nodename=slice_spec['nodenames'][0]
244 return self.locate_sliver_obj(nodename,slicename)
246 # all different hostboxes used in this plc
247 def gather_hostBoxes(self):
248 # maps on sites and nodes, return [ (host_box,test_node) ]
250 for site_spec in self.plc_spec['sites']:
251 test_site = TestSite (self,site_spec)
252 for node_spec in site_spec['nodes']:
253 test_node = TestNode (self, test_site, node_spec)
254 if not test_node.is_real():
255 tuples.append( (test_node.host_box(),test_node) )
256 # transform into a dict { 'host_box' -> [ test_node .. ] }
258 for (box,node) in tuples:
259 if not result.has_key(box):
262 result[box].append(node)
265 # a step for checking this stuff
266 def show_boxes (self):
267 'print summary of nodes location'
268 for (box,nodes) in self.gather_hostBoxes().iteritems():
269 print box,":"," + ".join( [ node.name() for node in nodes ] )
272 # make this a valid step
273 def kill_all_qemus(self):
274 'kill all qemu instances on the qemu boxes involved by this setup'
275 # this is the brute force version, kill all qemus on that host box
276 for (box,nodes) in self.gather_hostBoxes().iteritems():
277 # pass the first nodename, as we don't push template-qemu on testboxes
278 nodedir=nodes[0].nodedir()
279 TestBox(box,self.options.buildname).kill_all_qemus(nodedir)
282 # make this a valid step
283 def list_all_qemus(self):
284 'list all qemu instances on the qemu boxes involved by this setup'
285 for (box,nodes) in self.gather_hostBoxes().iteritems():
286 # this is the brute force version, kill all qemus on that host box
287 TestBox(box,self.options.buildname).list_all_qemus()
290 # kill only the right qemus
291 def list_qemus(self):
292 'list qemu instances for our nodes'
293 for (box,nodes) in self.gather_hostBoxes().iteritems():
294 # the fine-grain version
299 # kill only the right qemus
300 def kill_qemus(self):
301 'kill the qemu instances for our nodes'
302 for (box,nodes) in self.gather_hostBoxes().iteritems():
303 # the fine-grain version
308 #################### display config
310 "show test configuration after localization"
311 self.display_pass (1)
312 self.display_pass (2)
316 always_display_keys=['PLC_WWW_HOST','nodes','sites',]
317 def display_pass (self,passno):
318 for (key,val) in self.plc_spec.iteritems():
319 if not self.options.verbose and key not in TestPlc.always_display_keys: continue
323 self.display_site_spec(site)
324 for node in site['nodes']:
325 self.display_node_spec(node)
326 elif key=='initscripts':
327 for initscript in val:
328 self.display_initscript_spec (initscript)
331 self.display_slice_spec (slice)
334 self.display_key_spec (key)
336 if key not in ['sites','initscripts','slices','keys', 'sfa']:
337 print '+ ',key,':',val
339 def display_site_spec (self,site):
340 print '+ ======== site',site['site_fields']['name']
341 for (k,v) in site.iteritems():
342 if not self.options.verbose and k not in TestPlc.always_display_keys: continue
345 print '+ ','nodes : ',
347 print node['node_fields']['hostname'],'',
353 print user['name'],'',
355 elif k == 'site_fields':
356 print '+ login_base',':',v['login_base']
357 elif k == 'address_fields':
361 PrettyPrinter(indent=8,depth=2).pprint(v)
363 def display_initscript_spec (self,initscript):
364 print '+ ======== initscript',initscript['initscript_fields']['name']
366 def display_key_spec (self,key):
367 print '+ ======== key',key['name']
369 def display_slice_spec (self,slice):
370 print '+ ======== slice',slice['slice_fields']['name']
371 for (k,v) in slice.iteritems():
384 elif k=='slice_fields':
385 print '+ fields',':',
386 print 'max_nodes=',v['max_nodes'],
391 def display_node_spec (self,node):
392 print "+ node",node['name'],"host_box=",node['host_box'],
393 print "hostname=",node['node_fields']['hostname'],
394 print "ip=",node['interface_fields']['ip']
397 # another entry point for just showing the boxes involved
398 def display_mapping (self):
399 TestPlc.display_mapping_plc(self.plc_spec)
403 def display_mapping_plc (plc_spec):
404 print '+ MyPLC',plc_spec['name']
405 print '+\tvserver address = root@%s:/vservers/%s'%(plc_spec['hostname'],plc_spec['vservername'])
406 print '+\tIP = %s/%s'%(plc_spec['PLC_API_HOST'],plc_spec['vserverip'])
407 for site_spec in plc_spec['sites']:
408 for node_spec in site_spec['nodes']:
409 TestPlc.display_mapping_node(node_spec)
412 def display_mapping_node (node_spec):
413 print '+ NODE %s'%(node_spec['name'])
414 print '+\tqemu box %s'%node_spec['host_box']
415 print '+\thostname=%s'%node_spec['node_fields']['hostname']
417 def resources_pre (self):
418 "run site-dependant pre-test script as defined in LocalTestResources"
419 from LocalTestResources import local_resources
420 return local_resources.step_pre(self)
422 def resources_post (self):
423 "run site-dependant post-test script as defined in LocalTestResources"
424 from LocalTestResources import local_resources
425 return local_resources.step_post(self)
427 def resources_list (self):
428 "run site-dependant list script as defined in LocalTestResources"
429 from LocalTestResources import local_resources
430 return local_resources.step_list(self)
432 def resources_release (self):
433 "run site-dependant release script as defined in LocalTestResources"
434 from LocalTestResources import local_resources
435 return local_resources.step_release(self)
437 def resources_release_plc (self):
438 "run site-dependant release script as defined in LocalTestResources"
439 from LocalTestResources import local_resources
440 return local_resources.step_release_plc(self)
442 def resources_release_qemu (self):
443 "run site-dependant release script as defined in LocalTestResources"
444 from LocalTestResources import local_resources
445 return local_resources.step_release_qemu(self)
448 "vserver delete the test myplc"
449 self.run_in_host("vserver --silent %s delete"%self.vservername)
453 # historically the build was being fetched by the tests
454 # now the build pushes itself as a subdir of the tests workdir
455 # so that the tests do not have to worry about extracting the build (svn, git, or whatever)
456 def create_vs (self):
457 "vserver creation (no install done)"
458 # push the local build/ dir to the testplc box
460 # a full path for the local calls
461 build_dir=os.path.dirname(sys.argv[0])
462 # sometimes this is empty - set to "." in such a case
463 if not build_dir: build_dir="."
464 build_dir += "/build"
466 # use a standard name - will be relative to remote buildname
468 # remove for safety; do *not* mkdir first, otherwise we end up with build/build/
469 self.test_ssh.rmdir(build_dir)
470 self.test_ssh.copy(build_dir,recursive=True)
471 # the repo url is taken from arch-rpms-url
472 # with the last step (i386) removed
473 repo_url = self.options.arch_rpms_url
474 for level in [ 'arch' ]:
475 repo_url = os.path.dirname(repo_url)
476 # pass the vbuild-nightly options to vtest-init-vserver
478 test_env_options += " -p %s"%self.options.personality
479 test_env_options += " -d %s"%self.options.pldistro
480 test_env_options += " -f %s"%self.options.fcdistro
481 script="vtest-init-vserver.sh"
482 vserver_name = self.vservername
483 vserver_options="--netdev eth0 --interface %s"%self.vserverip
485 vserver_hostname=socket.gethostbyaddr(self.vserverip)[0]
486 vserver_options += " --hostname %s"%vserver_hostname
488 print "Cannot reverse lookup %s"%self.vserverip
489 print "This is considered fatal, as this might pollute the test results"
491 create_vserver="%(build_dir)s/%(script)s %(test_env_options)s %(vserver_name)s %(repo_url)s -- %(vserver_options)s"%locals()
492 return self.run_in_host(create_vserver) == 0
496 "yum install myplc, noderepo, and the plain bootstrapfs"
498 # workaround for getting pgsql8.2 on centos5
499 if self.options.fcdistro == "centos5":
500 self.run_in_guest("rpm -Uvh http://download.fedora.redhat.com/pub/epel/5/i386/epel-release-5-3.noarch.rpm")
502 if self.options.personality == "linux32":
504 elif self.options.personality == "linux64":
507 raise Exception, "Unsupported personality %r"%self.options.personality
509 nodefamily="%s-%s-%s"%(self.options.pldistro,self.options.fcdistro,arch)
511 # try to install slicerepo - not fatal yet
512 self.run_in_guest("yum -y install slicerepo-%s"%nodefamily)
515 self.run_in_guest("yum -y install myplc")==0 and \
516 self.run_in_guest("yum -y install noderepo-%s"%nodefamily)==0 and \
517 self.run_in_guest("yum -y install bootstrapfs-%s-plain"%nodefamily)==0
522 tmpname='%s.plc-config-tty'%(self.name())
523 fileconf=open(tmpname,'w')
524 for var in [ 'PLC_NAME',
528 'PLC_MAIL_SUPPORT_ADDRESS',
531 # Above line was added for integrating SFA Testing
537 fileconf.write ('e %s\n%s\n'%(var,self.plc_spec[var]))
538 fileconf.write('w\n')
539 fileconf.write('q\n')
541 utils.system('cat %s'%tmpname)
542 self.run_in_guest_piped('cat %s'%tmpname,'plc-config-tty')
543 utils.system('rm %s'%tmpname)
548 self.run_in_guest('service plc start')
553 self.run_in_guest('service plc stop')
557 "start the PLC vserver"
561 # stores the keys from the config for further use
562 def store_keys(self):
563 "stores test users ssh keys in keys/"
564 for key_spec in self.plc_spec['keys']:
565 TestKey(self,key_spec).store_key()
568 def clean_keys(self):
569 "removes keys cached in keys/"
570 utils.system("rm -rf %s/keys/"%os.path(sys.argv[0]))
572 # fetches the ssh keys in the plc's /etc/planetlab and stores them in keys/
573 # for later direct access to the nodes
574 def fetch_keys(self):
575 "gets ssh keys in /etc/planetlab/ and stores them locally in keys/"
577 if not os.path.isdir(dir):
579 vservername=self.vservername
581 prefix = 'debug_ssh_key'
582 for ext in [ 'pub', 'rsa' ] :
583 src="/vservers/%(vservername)s/etc/planetlab/%(prefix)s.%(ext)s"%locals()
584 dst="keys/%(vservername)s-debug.%(ext)s"%locals()
585 if self.test_ssh.fetch(src,dst) != 0: overall=False
589 "create sites with PLCAPI"
590 return self.do_sites()
592 def clean_sites (self):
593 "delete sites with PLCAPI"
594 return self.do_sites(action="delete")
596 def do_sites (self,action="add"):
597 for site_spec in self.plc_spec['sites']:
598 test_site = TestSite (self,site_spec)
599 if (action != "add"):
600 utils.header("Deleting site %s in %s"%(test_site.name(),self.name()))
601 test_site.delete_site()
602 # deleted with the site
603 #test_site.delete_users()
606 utils.header("Creating site %s & users in %s"%(test_site.name(),self.name()))
607 test_site.create_site()
608 test_site.create_users()
611 def clean_all_sites (self):
612 "Delete all sites in PLC, and related objects"
613 print 'auth_root',self.auth_root()
614 site_ids = [s['site_id'] for s in self.apiserver.GetSites(self.auth_root(), {}, ['site_id'])]
615 for site_id in site_ids:
616 print 'Deleting site_id',site_id
617 self.apiserver.DeleteSite(self.auth_root(),site_id)
620 "create nodes with PLCAPI"
621 return self.do_nodes()
622 def clean_nodes (self):
623 "delete nodes with PLCAPI"
624 return self.do_nodes(action="delete")
626 def do_nodes (self,action="add"):
627 for site_spec in self.plc_spec['sites']:
628 test_site = TestSite (self,site_spec)
630 utils.header("Deleting nodes in site %s"%test_site.name())
631 for node_spec in site_spec['nodes']:
632 test_node=TestNode(self,test_site,node_spec)
633 utils.header("Deleting %s"%test_node.name())
634 test_node.delete_node()
636 utils.header("Creating nodes for site %s in %s"%(test_site.name(),self.name()))
637 for node_spec in site_spec['nodes']:
638 utils.pprint('Creating node %s'%node_spec,node_spec)
639 test_node = TestNode (self,test_site,node_spec)
640 test_node.create_node ()
643 def nodegroups (self):
644 "create nodegroups with PLCAPI"
645 return self.do_nodegroups("add")
646 def clean_nodegroups (self):
647 "delete nodegroups with PLCAPI"
648 return self.do_nodegroups("delete")
650 # create nodegroups if needed, and populate
651 def do_nodegroups (self, action="add"):
652 # 1st pass to scan contents
654 for site_spec in self.plc_spec['sites']:
655 test_site = TestSite (self,site_spec)
656 for node_spec in site_spec['nodes']:
657 test_node=TestNode (self,test_site,node_spec)
658 if node_spec.has_key('nodegroups'):
659 nodegroupnames=node_spec['nodegroups']
660 if isinstance(nodegroupnames,StringTypes):
661 nodegroupnames = [ nodegroupnames ]
662 for nodegroupname in nodegroupnames:
663 if not groups_dict.has_key(nodegroupname):
664 groups_dict[nodegroupname]=[]
665 groups_dict[nodegroupname].append(test_node.name())
666 auth=self.auth_root()
668 for (nodegroupname,group_nodes) in groups_dict.iteritems():
670 print 'nodegroups:','dealing with nodegroup',nodegroupname,'on nodes',group_nodes
671 # first, check if the nodetagtype is here
672 tag_types = self.apiserver.GetTagTypes(auth,{'tagname':nodegroupname})
674 tag_type_id = tag_types[0]['tag_type_id']
676 tag_type_id = self.apiserver.AddTagType(auth,
677 {'tagname':nodegroupname,
678 'description': 'for nodegroup %s'%nodegroupname,
681 print 'located tag (type)',nodegroupname,'as',tag_type_id
683 nodegroups = self.apiserver.GetNodeGroups (auth, {'groupname':nodegroupname})
685 self.apiserver.AddNodeGroup(auth, nodegroupname, tag_type_id, 'yes')
686 print 'created nodegroup',nodegroupname,'from tagname',nodegroupname,'and value','yes'
687 # set node tag on all nodes, value='yes'
688 for nodename in group_nodes:
690 self.apiserver.AddNodeTag(auth, nodename, nodegroupname, "yes")
692 traceback.print_exc()
693 print 'node',nodename,'seems to already have tag',nodegroupname
696 expect_yes = self.apiserver.GetNodeTags(auth,
697 {'hostname':nodename,
698 'tagname':nodegroupname},
699 ['value'])[0]['value']
700 if expect_yes != "yes":
701 print 'Mismatch node tag on node',nodename,'got',expect_yes
704 if not self.options.dry_run:
705 print 'Cannot find tag',nodegroupname,'on node',nodename
709 print 'cleaning nodegroup',nodegroupname
710 self.apiserver.DeleteNodeGroup(auth,nodegroupname)
712 traceback.print_exc()
716 # return a list of tuples (nodename,qemuname)
717 def all_node_infos (self) :
719 for site_spec in self.plc_spec['sites']:
720 node_infos += [ (node_spec['node_fields']['hostname'],node_spec['host_box']) \
721 for node_spec in site_spec['nodes'] ]
724 def all_nodenames (self): return [ x[0] for x in self.all_node_infos() ]
726 # silent_minutes : during the first <silent_minutes> minutes nothing gets printed
727 def nodes_check_boot_state (self, target_boot_state, timeout_minutes, silent_minutes,period=15):
728 if self.options.dry_run:
732 timeout = datetime.datetime.now()+datetime.timedelta(minutes=timeout_minutes)
733 graceout = datetime.datetime.now()+datetime.timedelta(minutes=silent_minutes)
734 # the nodes that haven't checked yet - start with a full list and shrink over time
735 tocheck = self.all_hostnames()
736 utils.header("checking nodes %r"%tocheck)
737 # create a dict hostname -> status
738 status = dict ( [ (hostname,'undef') for hostname in tocheck ] )
741 tocheck_status=self.apiserver.GetNodes(self.auth_root(), tocheck, ['hostname','boot_state' ] )
743 for array in tocheck_status:
744 hostname=array['hostname']
745 boot_state=array['boot_state']
746 if boot_state == target_boot_state:
747 utils.header ("%s has reached the %s state"%(hostname,target_boot_state))
749 # if it's a real node, never mind
750 (site_spec,node_spec)=self.locate_hostname(hostname)
751 if TestNode.is_real_model(node_spec['node_fields']['model']):
752 utils.header("WARNING - Real node %s in %s - ignored"%(hostname,boot_state))
754 boot_state = target_boot_state
755 elif datetime.datetime.now() > graceout:
756 utils.header ("%s still in '%s' state"%(hostname,boot_state))
757 graceout=datetime.datetime.now()+datetime.timedelta(1)
758 status[hostname] = boot_state
760 tocheck = [ hostname for (hostname,boot_state) in status.iteritems() if boot_state != target_boot_state ]
763 if datetime.datetime.now() > timeout:
764 for hostname in tocheck:
765 utils.header("FAILURE due to %s in '%s' state"%(hostname,status[hostname]))
767 # otherwise, sleep for a while
769 # only useful in empty plcs
772 def nodes_booted(self):
773 return self.nodes_check_boot_state('boot',timeout_minutes=30,silent_minutes=20)
775 def check_nodes_ssh(self,debug,timeout_minutes,silent_minutes,period=15):
777 timeout = datetime.datetime.now()+datetime.timedelta(minutes=timeout_minutes)
778 graceout = datetime.datetime.now()+datetime.timedelta(minutes=silent_minutes)
779 vservername=self.vservername
782 local_key = "keys/%(vservername)s-debug.rsa"%locals()
785 local_key = "keys/key1.rsa"
786 node_infos = self.all_node_infos()
787 utils.header("checking ssh access (expected in %s mode) to nodes:"%message)
788 for (nodename,qemuname) in node_infos:
789 utils.header("hostname=%s -- qemubox=%s"%(nodename,qemuname))
790 utils.header("max timeout is %d minutes, silent for %d minutes (period is %s)"%\
791 (timeout_minutes,silent_minutes,period))
793 for node_info in node_infos:
794 (hostname,qemuname) = node_info
795 # try to run 'hostname' in the node
796 command = TestSsh (hostname,key=local_key).actual_command("hostname;uname -a")
797 # don't spam logs - show the command only after the grace period
798 success = utils.system ( command, silent=datetime.datetime.now() < graceout)
800 utils.header('Successfully entered root@%s (%s)'%(hostname,message))
802 node_infos.remove(node_info)
804 # we will have tried real nodes once, in case they're up - but if not, just skip
805 (site_spec,node_spec)=self.locate_hostname(hostname)
806 if TestNode.is_real_model(node_spec['node_fields']['model']):
807 utils.header ("WARNING : check ssh access into real node %s - skipped"%hostname)
808 node_infos.remove(node_info)
811 if datetime.datetime.now() > timeout:
812 for (hostname,qemuname) in node_infos:
813 utils.header("FAILURE to ssh into %s (on %s)"%(hostname,qemuname))
815 # otherwise, sleep for a while
817 # only useful in empty plcs
820 def nodes_ssh_debug(self):
821 "Tries to ssh into nodes in debug mode with the debug ssh key"
822 return self.check_nodes_ssh(debug=True,timeout_minutes=30,silent_minutes=5)
824 def nodes_ssh_boot(self):
825 "Tries to ssh into nodes in production mode with the root ssh key"
826 return self.check_nodes_ssh(debug=False,timeout_minutes=30,silent_minutes=15)
829 def init_node (self):
830 "all nodes : init a clean local directory for holding node-dep stuff like iso image..."
834 "all nodes: invoke GetBootMedium and store result locally"
837 def configure_qemu (self):
838 "all nodes: compute qemu config qemu.conf and store it locally"
841 def reinstall_node (self):
842 "all nodes: mark PLCAPI boot_state as reinstall"
845 def export_qemu (self):
846 "all nodes: push local node-dep directory on the qemu box"
849 ### check hooks : invoke scripts from hooks/{node,slice}
850 def check_hooks_node (self):
851 return self.locate_first_node().check_hooks()
852 def check_hooks_sliver (self) :
853 return self.locate_first_sliver().check_hooks()
855 def check_hooks (self):
856 "runs unit tests in the node and slice contexts - see hooks/{node,slice}"
857 return self.check_hooks_node() and self.check_hooks_sliver()
860 def do_check_initscripts(self):
862 for slice_spec in self.plc_spec['slices']:
863 if not slice_spec.has_key('initscriptname'):
865 initscript=slice_spec['initscriptname']
866 for nodename in slice_spec['nodenames']:
867 (site,node) = self.locate_node (nodename)
868 # xxx - passing the wrong site - probably harmless
869 test_site = TestSite (self,site)
870 test_slice = TestSlice (self,test_site,slice_spec)
871 test_node = TestNode (self,test_site,node)
872 test_sliver = TestSliver (self, test_node, test_slice)
873 if not test_sliver.check_initscript(initscript):
877 def check_initscripts(self):
878 "check that the initscripts have triggered"
879 return self.do_check_initscripts()
881 def initscripts (self):
882 "create initscripts with PLCAPI"
883 for initscript in self.plc_spec['initscripts']:
884 utils.pprint('Adding Initscript in plc %s'%self.plc_spec['name'],initscript)
885 self.apiserver.AddInitScript(self.auth_root(),initscript['initscript_fields'])
888 def clean_initscripts (self):
889 "delete initscripts with PLCAPI"
890 for initscript in self.plc_spec['initscripts']:
891 initscript_name = initscript['initscript_fields']['name']
892 print('Attempting to delete %s in plc %s'%(initscript_name,self.plc_spec['name']))
894 self.apiserver.DeleteInitScript(self.auth_root(),initscript_name)
895 print initscript_name,'deleted'
897 print 'deletion went wrong - probably did not exist'
902 "create slices with PLCAPI"
903 return self.do_slices()
905 def clean_slices (self):
906 "delete slices with PLCAPI"
907 return self.do_slices("delete")
909 def do_slices (self, action="add"):
910 for slice in self.plc_spec['slices']:
911 site_spec = self.locate_site (slice['sitename'])
912 test_site = TestSite(self,site_spec)
913 test_slice=TestSlice(self,test_site,slice)
915 utils.header("Deleting slices in site %s"%test_site.name())
916 test_slice.delete_slice()
918 utils.pprint("Creating slice",slice)
919 test_slice.create_slice()
920 utils.header('Created Slice %s'%slice['slice_fields']['name'])
923 @slice_mapper_options
924 def check_slice(self):
925 "tries to ssh-enter the slice with the user key, to ensure slice creation"
929 def clear_known_hosts (self):
930 "remove test nodes entries from the local known_hosts file"
934 def start_node (self) :
935 "all nodes: start the qemu instance (also runs qemu-bridge-init start)"
938 def check_tcp (self):
939 "check TCP connectivity between 2 slices (or in loopback if only one is defined)"
940 specs = self.plc_spec['tcp_test']
945 s_test_sliver = self.locate_sliver_obj (spec['server_node'],spec['server_slice'])
946 if not s_test_sliver.run_tcp_server(port,timeout=10):
950 # idem for the client side
951 c_test_sliver = self.locate_sliver_obj(spec['server_node'],spec['server_slice'])
952 if not c_test_sliver.run_tcp_client(s_test_sliver.test_node.name(),port):
956 def plcsh_stress_test (self):
957 "runs PLCAPI stress test, that checks Add/Update/Delete on all types - preserves contents"
958 # install the stress-test in the plc image
959 location = "/usr/share/plc_api/plcsh_stress_test.py"
960 remote="/vservers/%s/%s"%(self.vservername,location)
961 self.test_ssh.copy_abs("plcsh_stress_test.py",remote)
963 command += " -- --check"
964 if self.options.size == 1:
966 return ( self.run_in_guest(command) == 0)
968 # populate runs the same utility without slightly different options
969 # in particular runs with --preserve (dont cleanup) and without --check
970 # also it gets run twice, once with the --foreign option for creating fake foreign entries
973 def install_sfa(self):
974 "yum install sfa, sfa-plc and sfa-client"
975 if self.options.personality == "linux32":
977 elif self.options.personality == "linux64":
980 raise Exception, "Unsupported personality %r"%self.options.personality
981 return self.run_in_guest("yum -y install sfa sfa-client sfa-plc sfa-sfatables")==0
984 def configure_sfa(self):
986 tmpname='%s.sfa-config-tty'%(self.name())
987 fileconf=open(tmpname,'w')
988 for var in [ 'SFA_REGISTRY_ROOT_AUTH',
989 'SFA_REGISTRY_LEVEL1_AUTH',
991 'SFA_AGGREGATE_HOST',
997 'SFA_PLC_DB_PASSWORD',
999 fileconf.write ('e %s\n%s\n'%(var,self.plc_spec['sfa'][var]))
1000 fileconf.write('w\n')
1001 fileconf.write('R\n')
1002 fileconf.write('q\n')
1004 utils.system('cat %s'%tmpname)
1005 self.run_in_guest_piped('cat %s'%tmpname,'sfa-config-tty')
1006 utils.system('rm %s'%tmpname)
1009 def import_sfa(self):
1011 auth=self.plc_spec['sfa']['SFA_REGISTRY_ROOT_AUTH']
1012 return self.run_in_guest('sfa-import-plc.py')==0
1013 # not needed anymore
1014 # self.run_in_guest('cp /etc/sfa/authorities/%s/%s.pkey /etc/sfa/authorities/server.key'%(auth,auth))
1016 def start_sfa(self):
1018 return self.run_in_guest('service sfa start')==0
1020 def setup_sfa(self):
1021 "sfi client configuration"
1023 if os.path.exists(dir_name):
1024 utils.system('rm -rf %s'%dir_name)
1025 utils.system('mkdir %s'%dir_name)
1026 file_name=dir_name + os.sep + 'fake-pi1.pkey'
1027 fileconf=open(file_name,'w')
1028 fileconf.write (self.plc_spec['keys'][0]['private'])
1031 file_name=dir_name + os.sep + 'sfi_config'
1032 fileconf=open(file_name,'w')
1033 SFI_AUTH=self.plc_spec['sfa']['SFA_REGISTRY_ROOT_AUTH']+".main"
1034 fileconf.write ("SFI_AUTH='%s'"%SFI_AUTH)
1035 fileconf.write('\n')
1036 SFI_USER=SFI_AUTH+'.fake-pi1'
1037 fileconf.write ("SFI_USER='%s'"%SFI_USER)
1038 fileconf.write('\n')
1039 SFI_REGISTRY='http://' + self.plc_spec['sfa']['SFA_PLC_DB_HOST'] + ':12345/'
1040 fileconf.write ("SFI_REGISTRY='%s'"%SFI_REGISTRY)
1041 fileconf.write('\n')
1042 SFI_SM='http://' + self.plc_spec['sfa']['SFA_PLC_DB_HOST'] + ':12347/'
1043 fileconf.write ("SFI_SM='%s'"%SFI_SM)
1044 fileconf.write('\n')
1047 file_name=dir_name + os.sep + 'person.xml'
1048 fileconf=open(file_name,'w')
1049 for record in self.plc_spec['sfa']['sfa_person_xml']:
1050 person_record=record
1051 fileconf.write(person_record)
1052 fileconf.write('\n')
1055 file_name=dir_name + os.sep + 'slice.xml'
1056 fileconf=open(file_name,'w')
1057 for record in self.plc_spec['sfa']['sfa_slice_xml']:
1059 #slice_record=self.plc_spec['sfa']['sfa_slice_xml']
1060 fileconf.write(slice_record)
1061 fileconf.write('\n')
1064 file_name=dir_name + os.sep + 'slice.rspec'
1065 fileconf=open(file_name,'w')
1067 for (key, value) in self.plc_spec['sfa']['sfa_slice_rspec'].items():
1069 fileconf.write(slice_rspec)
1070 fileconf.write('\n')
1073 remote="/vservers/%s/%s"%(self.vservername,location)
1074 self.test_ssh.copy_abs(dir_name, remote, recursive=True)
1076 #utils.system('cat %s'%tmpname)
1077 utils.system('rm -rf %s'%dir_name)
1081 "run sfi.py add (on Registry) and sfi.py create (on SM) to form new objects"
1083 test_user_sfa=TestUserSfa(test_plc,self.plc_spec['sfa'])
1084 success=test_user_sfa.add_user()
1086 for slice_spec in self.plc_spec['sfa']['slices_sfa']:
1087 site_spec = self.locate_site (slice_spec['sitename'])
1088 test_site = TestSite(self,site_spec)
1089 test_slice_sfa=TestSliceSfa(test_plc,test_site,slice_spec)
1090 success1=test_slice_sfa.add_slice()
1091 success2=test_slice_sfa.create_slice()
1092 return success and success1 and success2
1094 def update_sfa(self):
1095 "run sfi.py update (on Registry) and sfi.py create (on SM) on existing objects"
1097 test_user_sfa=TestUserSfa(test_plc,self.plc_spec['sfa'])
1098 success1=test_user_sfa.update_user()
1100 for slice_spec in self.plc_spec['sfa']['slices_sfa']:
1101 site_spec = self.locate_site (slice_spec['sitename'])
1102 test_site = TestSite(self,site_spec)
1103 test_slice_sfa=TestSliceSfa(test_plc,test_site,slice_spec)
1104 success2=test_slice_sfa.update_slice()
1105 return success1 and success2
1108 "run sfi.py list and sfi.py show (both on Registry) and sfi.py slices and sfi.py resources (both on SM)"
1109 auth=self.plc_spec['sfa']['SFA_REGISTRY_ROOT_AUTH']
1111 self.run_in_guest("sfi.py -d /root/.sfi/ list %s.main"%auth)==0 and \
1112 self.run_in_guest("sfi.py -d /root/.sfi/ show %s.main"%auth)==0 and \
1113 self.run_in_guest("sfi.py -d /root/.sfi/ slices")==0 and \
1114 self.run_in_guest("sfi.py -d /root/.sfi/ resources -o resources")==0
1116 @slice_mapper_options_sfa
1117 def check_slice_sfa(self):
1118 "tries to ssh-enter the SFA slice"
1121 def delete_sfa(self):
1122 "run sfi.py delete (on SM), sfi.py remove (on Registry)"
1124 test_user_sfa=TestUserSfa(test_plc,self.plc_spec['sfa'])
1125 success1=test_user_sfa.delete_user()
1126 for slice_spec in self.plc_spec['sfa']['slices_sfa']:
1127 site_spec = self.locate_site (slice_spec['sitename'])
1128 test_site = TestSite(self,site_spec)
1129 test_slice_sfa=TestSliceSfa(test_plc,test_site,slice_spec)
1130 success2=test_slice_sfa.delete_slice()
1132 return success1 and success2
1136 return self.run_in_guest('service sfa stop')==0
1138 def populate (self):
1139 "creates random entries in the PLCAPI"
1140 # install the stress-test in the plc image
1141 location = "/usr/share/plc_api/plcsh_stress_test.py"
1142 remote="/vservers/%s/%s"%(self.vservername,location)
1143 self.test_ssh.copy_abs("plcsh_stress_test.py",remote)
1145 command += " -- --preserve --short-names"
1146 local = (self.run_in_guest(command) == 0);
1147 # second run with --foreign
1148 command += ' --foreign'
1149 remote = (self.run_in_guest(command) == 0);
1150 return ( local and remote)
1152 def gather_logs (self):
1153 "gets all possible logs from plc's/qemu node's/slice's for future reference"
1154 # (1.a) get the plc's /var/log/ and store it locally in logs/myplc.var-log.<plcname>/*
1155 # (1.b) get the plc's /var/lib/pgsql/data/pg_log/ -> logs/myplc.pgsql-log.<plcname>/*
1156 # (2) get all the nodes qemu log and store it as logs/node.qemu.<node>.log
1157 # (3) get the nodes /var/log and store is as logs/node.var-log.<node>/*
1158 # (4) as far as possible get the slice's /var/log as logs/sliver.var-log.<sliver>/*
1160 print "-------------------- TestPlc.gather_logs : PLC's /var/log"
1161 self.gather_var_logs ()
1163 print "-------------------- TestPlc.gather_logs : PLC's /var/lib/psql/data/pg_log/"
1164 self.gather_pgsql_logs ()
1166 print "-------------------- TestPlc.gather_logs : nodes's QEMU logs"
1167 for site_spec in self.plc_spec['sites']:
1168 test_site = TestSite (self,site_spec)
1169 for node_spec in site_spec['nodes']:
1170 test_node=TestNode(self,test_site,node_spec)
1171 test_node.gather_qemu_logs()
1173 print "-------------------- TestPlc.gather_logs : nodes's /var/log"
1174 self.gather_nodes_var_logs()
1176 print "-------------------- TestPlc.gather_logs : sample sliver's /var/log"
1177 self.gather_slivers_var_logs()
1180 def gather_slivers_var_logs(self):
1181 for test_sliver in self.all_sliver_objs():
1182 remote = test_sliver.tar_var_logs()
1183 utils.system("mkdir -p logs/sliver.var-log.%s"%test_sliver.name())
1184 command = remote + " | tar -C logs/sliver.var-log.%s -xf -"%test_sliver.name()
1185 utils.system(command)
1188 def gather_var_logs (self):
1189 utils.system("mkdir -p logs/myplc.var-log.%s"%self.name())
1190 to_plc = self.actual_command_in_guest("tar -C /var/log/ -cf - .")
1191 command = to_plc + "| tar -C logs/myplc.var-log.%s -xf -"%self.name()
1192 utils.system(command)
1193 command = "chmod a+r,a+x logs/myplc.var-log.%s/httpd"%self.name()
1194 utils.system(command)
1196 def gather_pgsql_logs (self):
1197 utils.system("mkdir -p logs/myplc.pgsql-log.%s"%self.name())
1198 to_plc = self.actual_command_in_guest("tar -C /var/lib/pgsql/data/pg_log/ -cf - .")
1199 command = to_plc + "| tar -C logs/myplc.pgsql-log.%s -xf -"%self.name()
1200 utils.system(command)
1202 def gather_nodes_var_logs (self):
1203 for site_spec in self.plc_spec['sites']:
1204 test_site = TestSite (self,site_spec)
1205 for node_spec in site_spec['nodes']:
1206 test_node=TestNode(self,test_site,node_spec)
1207 test_ssh = TestSsh (test_node.name(),key="keys/key1.rsa")
1208 command = test_ssh.actual_command("tar -C /var/log -cf - .")
1209 command = command + "| tar -C logs/node.var-log.%s -xf -"%test_node.name()
1210 utils.system("mkdir -p logs/node.var-log.%s"%test_node.name())
1211 utils.system(command)
1214 # returns the filename to use for sql dump/restore, using options.dbname if set
1215 def dbfile (self, database):
1216 # uses options.dbname if it is found
1218 name=self.options.dbname
1219 if not isinstance(name,StringTypes):
1222 t=datetime.datetime.now()
1225 return "/root/%s-%s.sql"%(database,name)
1228 'dump the planetlab5 DB in /root in the PLC - filename has time'
1229 dump=self.dbfile("planetab5")
1230 self.run_in_guest('pg_dump -U pgsqluser planetlab5 -f '+ dump)
1231 utils.header('Dumped planetlab5 database in %s'%dump)
1234 def db_restore(self):
1235 'restore the planetlab5 DB - looks broken, but run -n might help'
1236 dump=self.dbfile("planetab5")
1237 ##stop httpd service
1238 self.run_in_guest('service httpd stop')
1239 # xxx - need another wrapper
1240 self.run_in_guest_piped('echo drop database planetlab5','psql --user=pgsqluser template1')
1241 self.run_in_guest('createdb -U postgres --encoding=UNICODE --owner=pgsqluser planetlab5')
1242 self.run_in_guest('psql -U pgsqluser planetlab5 -f '+dump)
1243 ##starting httpd service
1244 self.run_in_guest('service httpd start')
1246 utils.header('Database restored from ' + dump)
1249 def standby_1(): pass
1251 def standby_2(): pass
1253 def standby_3(): pass
1255 def standby_4(): pass
1257 def standby_5(): pass
1259 def standby_6(): pass
1261 def standby_7(): pass
1263 def standby_8(): pass
1265 def standby_9(): pass
1267 def standby_10(): pass
1269 def standby_11(): pass
1271 def standby_12(): pass
1273 def standby_13(): pass
1275 def standby_14(): pass
1277 def standby_15(): pass
1279 def standby_16(): pass
1281 def standby_17(): pass
1283 def standby_18(): pass
1285 def standby_19(): pass
1287 def standby_20(): pass