blob: 3ef9be800797037c7b48311c9d20174932a9be22 [file] [log] [blame]
# SCPFintentRerouteLat
"""
SCPFintentRerouteLat
- Test Intent Reroute Latency
- Test Algorithm:
1. Start Null Provider reroute Topology
2. Using Push-test-intents to push batch size intents from switch 1 to switch 7
3. Cut the link between switch 3 and switch 4 (the path will reroute to switch 8)
4. Get the topology time stamp
5. Get Intent reroute(Installed) time stamp from each nodes
6. Use the latest intent time stamp subtract topology time stamp
- This test will run 5 warm up by default, warm up iteration can be setup in Param file
- The intent batch size will default set to 1, 100, and 1000, also can be set in Param file
- The unit of the latency result is milliseconds
"""
class SCPFintentRerouteLat:
def __init__(self):
self.default = ''
def CASE0( self, main ):
import imp
import os
'''
- GIT
- BUILDING ONOS
Pull specific ONOS branch, then Build ONOS ono ONOS Bench.
This step is usually skipped. Because in a Jenkins driven automated
test env. We want Jenkins jobs to pull&build for flexibility to handle
different versions of ONOS.
- Construct tests variables
'''
gitPull = main.params['GIT']['gitPull']
gitBranch = main.params['GIT']['gitBranch']
main.case("Pull onos branch and build onos on Teststation.")
if gitPull == 'True':
main.step("Git Checkout ONOS branch: " + gitBranch)
stepResult = main.ONOSbench.gitCheckout(branch=gitBranch)
utilities.assert_equals(expect=main.TRUE,
actual=stepResult,
onpass="Successfully checkout onos branch.",
onfail="Failed to checkout onos branch. Exiting test...")
if not stepResult:
main.exit()
main.step("Git Pull on ONOS branch:" + gitBranch)
stepResult = main.ONOSbench.gitPull()
utilities.assert_equals(expect=main.TRUE,
actual=stepResult,
onpass="Successfully pull onos. ",
onfail="Failed to pull onos. Exiting test ...")
if not stepResult: main.exit()
else:
main.log.warn("Skipped pulling onos and Skipped building ONOS")
main.onosIp = main.ONOSbench.getOnosIps()
main.apps = main.params['ENV']['cellApps']
main.BENCHUser = main.params['BENCH']['user']
main.BENCHIp = main.params['BENCH']['ip1']
main.MN1Ip = main.params['MN']['ip1']
main.maxNodes = int(main.params['max'])
main.cellName = main.params['ENV']['cellName']
main.scale = (main.params['SCALE']).split(",")
main.timeout = int(main.params['SLEEP']['timeout'])
main.startUpSleep = int(main.params['SLEEP']['startup'])
main.installSleep = int(main.params['SLEEP']['install'])
main.verifySleep = int(main.params['SLEEP']['verify'])
main.setMasterSleep = int(main.params['SLEEP']['setmaster'])
main.verifyAttempts = int(main.params['ATTEMPTS']['verify'])
main.maxInvalidRun = int(main.params['ATTEMPTS']['maxInvalidRun'])
main.sampleSize = int(main.params['TEST']['sampleSize'])
main.warmUp = int(main.params['TEST']['warmUp'])
main.ingress = main.params['TEST']['ingress']
main.egress = main.params['TEST']['egress']
main.debug = main.params['TEST']['debug']
main.flowObj = main.params['TEST']['flowObj']
main.deviceCount = int(main.params['TEST']['deviceCount'])
main.end1 = main.params['TEST']['end1']
main.end2 = main.params['TEST']['end2']
main.searchTerm = main.params['SEARCHTERM']
if main.flowObj == "True":
main.flowObj = True
main.dbFileName = main.params['DATABASE']['dbFlowObj']
main.intentsList = (main.params['TEST']['FObjintents']).split(",")
else:
main.flowObj = False
main.dbFileName = main.params['DATABASE']['dbName']
main.intentsList = (main.params['TEST']['intents']).split(",")
for i in range(0, len(main.intentsList)):
main.intentsList[i] = int(main.intentsList[i])
# Create DataBase file
main.log.info("Create Database file " + main.dbFileName)
resultsDB = open(main.dbFileName, "w+")
resultsDB.close()
file1 = main.params[ "DEPENDENCY" ][ "FILE1" ]
main.dependencyPath = os.path.dirname( os.getcwd() ) + main.params[ "DEPENDENCY" ][ "PATH" ]
main.intentRerouteLatFuncs = imp.load_source(file1, main.dependencyPath + file1 + ".py")
main.record = 0
def CASE1( self, main ):
'''
clean up test environment and set up
'''
import time
main.log.info("Get ONOS cluster IP")
print(main.scale)
main.numCtrls = int(main.scale[0])
main.ONOSip = []
main.maxNumBatch = 0
main.AllONOSip = main.ONOSbench.getOnosIps()
for i in range(main.numCtrls):
main.ONOSip.append(main.AllONOSip[i])
main.log.info(main.ONOSip)
main.CLIs = []
main.log.info("Creating list of ONOS cli handles")
for i in range(main.numCtrls):
main.CLIs.append(getattr(main, 'ONOScli%s' % (i + 1)))
if not main.CLIs:
main.log.error("Failed to create the list of ONOS cli handles")
main.cleanup()
main.exit()
main.commit = main.ONOSbench.getVersion(report=True)
main.commit = main.commit.split(" ")[1]
main.log.info("Starting up %s node(s) ONOS cluster" % main.numCtrls)
main.log.info("Safety check, killing all ONOS processes" +
" before initiating environment setup")
for i in range(main.numCtrls):
main.ONOSbench.onosStop(main.ONOSip[i])
main.ONOSbench.onosKill(main.ONOSip[i])
main.log.info("NODE COUNT = %s" % main.numCtrls)
main.ONOSbench.createCellFile(main.ONOSbench.ip_address,
main.cellName,
main.MN1Ip,
main.apps,
main.ONOSip,
main.ONOScli1.karafUser)
main.step("Apply cell to environment")
cellResult = main.ONOSbench.setCell(main.cellName)
verifyResult = main.ONOSbench.verifyCell()
stepResult = cellResult and verifyResult
utilities.assert_equals(expect=main.TRUE,
actual=stepResult,
onpass="Successfully applied cell to " + \
"environment",
onfail="Failed to apply cell to environment ")
main.step("Creating ONOS package")
packageResult = main.ONOSbench.buckBuild()
stepResult = packageResult
utilities.assert_equals(expect=main.TRUE,
actual=stepResult,
onpass="Successfully created ONOS package",
onfail="Failed to create ONOS package")
main.step("Uninstall ONOS package on all Nodes")
uninstallResult = main.TRUE
for i in range(int(main.numCtrls)):
main.log.info("Uninstalling package on ONOS Node IP: " + main.ONOSip[i])
u_result = main.ONOSbench.onosUninstall(main.ONOSip[i])
utilities.assert_equals(expect=main.TRUE, actual=u_result,
onpass="Test step PASS",
onfail="Test step FAIL")
uninstallResult = (uninstallResult and u_result)
main.step("Install ONOS package on all Nodes")
installResult = main.TRUE
for i in range(int(main.numCtrls)):
main.log.info("Installing package on ONOS Node IP: " + main.ONOSip[i])
i_result = main.ONOSbench.onosInstall(node=main.ONOSip[i])
utilities.assert_equals(expect=main.TRUE, actual=i_result,
onpass="Test step PASS",
onfail="Test step FAIL")
installResult = installResult and i_result
main.step( "Set up ONOS secure SSH" )
secureSshResult = main.TRUE
for i in range( int( main.numCtrls ) ):
secureSshResult = secureSshResult and main.ONOSbench.onosSecureSSH( node=main.ONOSip[i] )
utilities.assert_equals( expect=main.TRUE, actual=secureSshResult,
onpass="Test step PASS",
onfail="Test step FAIL" )
main.step( "Starting ONOS service" )
stopResult = main.TRUE
startResult = main.TRUE
onosIsUp = main.TRUE
for i in range( main.numCtrls ):
onosIsUp = onosIsUp and main.ONOSbench.isup( main.ONOSip[ i ] )
if onosIsUp == main.TRUE:
main.log.report( "ONOS instance is up and ready" )
else:
main.log.report( "ONOS instance may not be up, stop and " +
"start ONOS again " )
for i in range( main.numCtrls ):
stopResult = stopResult and \
main.ONOSbench.onosStop( main.ONOSip[ i ] )
for i in range( main.numCtrls ):
startResult = startResult and \
main.ONOSbench.onosStart( main.ONOSip[ i ] )
stepResult = onosIsUp and stopResult and startResult
utilities.assert_equals( expect=main.TRUE, actual=stepResult,
onpass="Test step PASS",
onfail="Test step FAIL" )
time.sleep(main.startUpSleep)
main.step("Start ONOS CLI on all nodes")
cliResult = main.TRUE
main.step(" Start ONOS cli using thread ")
startCliResult = main.TRUE
pool = []
main.threadID = 0
for i in range(int(main.numCtrls)):
t = main.Thread(target=main.CLIs[i].startOnosCli,
threadID=main.threadID,
name="startOnosCli",
args=[main.ONOSip[i]],
kwargs={"onosStartTimeout": main.timeout})
pool.append(t)
t.start()
main.threadID = main.threadID + 1
for t in pool:
t.join()
startCliResult = startCliResult and t.result
time.sleep(main.startUpSleep)
# configure apps
main.CLIs[0].setCfg("org.onosproject.provider.nil.NullProviders", "deviceCount", value=main.deviceCount)
main.CLIs[0].setCfg("org.onosproject.provider.nil.NullProviders", "topoShape", value="reroute")
main.CLIs[0].setCfg("org.onosproject.provider.nil.NullProviders", "enabled", value="true")
if main.flowObj:
main.CLIs[0].setCfg("org.onosproject.net.intent.impl.compiler.IntentConfigurableRegistrator",
"useFlowObjectives", value="true")
main.CLIs[0].setCfg("org.onosproject.net.intent.impl.compiler.IntentConfigurableRegistrator",
"defaultFlowObjectiveCompiler",
value='org.onosproject.net.intent.impl.compiler.LinkCollectionIntentObjectiveCompiler')
time.sleep( main.startUpSleep )
for i in range( int( main.numCtrls ) ):
main.CLIs[i].logSet( "DEBUG", "org.onosproject.metrics.topology" )
main.CLIs[i].logSet( "DEBUG", "org.onosproject.metrics.intent" )
# Balance Master
main.CLIs[0].balanceMasters()
time.sleep( main.setMasterSleep )
if len(main.ONOSip) > 1:
main.CLIs[0].deviceRole(main.end1[ 'name' ], main.ONOSip[0])
main.CLIs[0].deviceRole(main.end2[ 'name' ], main.ONOSip[0])
time.sleep( main.setMasterSleep )
def CASE2( self, main ):
import time
import numpy
import datetime
import json
# from scipy import stats
print(main.intentsList)
for batchSize in main.intentsList:
main.batchSize = batchSize
main.log.report("Intent Batch size: " + str(batchSize) + "\n ")
firstLocalLatencies = []
lastLocalLatencies = []
firstGlobalLatencies = []
lastGlobalLatencies = []
main.startLine = {}
main.validRun = 0
main.invalidRun = 0
while main.validRun <= main.warmUp + main.sampleSize and main.invalidRun <= main.maxInvalidRun:
if main.validRun >= main.warmUp:
main.log.info("================================================")
main.log.info("Valid iteration: {} ".format( main.validRun - main.warmUp))
main.log.info("Total iteration: {}".format( main.validRun + main.invalidRun))
main.log.info("================================================")
else:
main.log.info("====================Warm Up=====================")
# push intents
main.CLIs[0].pushTestIntents(main.ingress, main.egress, main.batchSize,
offset=1, options="-i", timeout=main.timeout)
# check links, flows and intents
main.intentRerouteLatFuncs.sanityCheck( main, main.deviceCount * 2, batchSize * (main.deviceCount - 1 ), main.batchSize )
if not main.verify:
main.log.warn( "Sanity check failed, skipping this iteration..." )
continue
# Insert one line in karaf.log before link down
for i in range( main.numCtrls ):
main.CLIs[ i ].log( "\'Scale: {}, Batch:{}, Iteration: {}\'".format( main.numCtrls, batchSize, main.validRun + main.invalidRun ) )
# bring link down
main.CLIs[0].link( main.end1[ 'port' ], main.end2[ 'port' ], "down",
timeout=main.timeout, showResponse=False)
# check links, flows and intents
main.intentRerouteLatFuncs.sanityCheck( main, (main.deviceCount - 1) * 2, batchSize * main.deviceCount, main.batchSize )
if not main.verify:
main.log.warn( "Sanity check failed, skipping this iteration..." )
continue
# Get timestamp of last LINK_REMOVED event as separator between iterations
skip = False
for i in range( main.numCtrls ):
logNum = main.intentRerouteLatFuncs.getLogNum( main, i )
timestamp = str( main.CLIs[ i ].getTimeStampFromLog( "last", "LINK_REMOVED", "time = ", " ", logNum=logNum ) )
if timestamp == main.ERROR:
# Try again in case that the log number just increased
logNum = main.intentRerouteLatFuncs.getLogNum( main, i )
timestamp = str( main.CLIs[ i ].getTimeStampFromLog( "last", "LINK_REMOVED", "time = ", " ", logNum=logNum ) )
if timestamp == main.ERROR:
main.log.warn( "Cannot find the event we want in the log, skipping this iteration..." )
main.intentRerouteLatFuncs.bringBackTopology( main )
if main.validRun >= main.warmUp:
main.invalidRun += 1
else:
main.validRun += 1
skip = True
break
else:
main.startLine[ i ] = timestamp
main.log.info( "Timestamp of last LINK_REMOVED event on node {} is {}".format( i+1, main.startLine[ i ] ) )
if skip: continue
# calculate values
topologyTimestamps = main.intentRerouteLatFuncs.getTopologyTimestamps( main )
intentTimestamps = main.intentRerouteLatFuncs.getIntentTimestamps( main )
if intentTimestamps == main.ERROR or topologyTimestamps == main.ERROR:
main.log.info( "Got invalid timestamp, skipping this iteration..." )
main.intentRerouteLatFuncs.bringBackTopology( main )
if main.validRun >= main.warmUp:
main.invalidRun += 1
else:
main.validRun += 1
continue
else:
main.log.info( "Got valid timestamps" )
firstLocalLatnecy, lastLocalLatnecy, firstGlobalLatency, lastGlobalLatnecy = main.intentRerouteLatFuncs.calculateLatency( main, topologyTimestamps, intentTimestamps )
if firstLocalLatnecy < 0:
main.log.info( "Got negative latency, skipping this iteration..." )
main.intentRerouteLatFuncs.bringBackTopology( main )
if main.validRun >= main.warmUp:
main.invalidRun += 1
else:
main.validRun += 1
continue
else:
main.log.info( "Got valid latencies" )
main.validRun += 1
firstLocalLatencies.append( firstLocalLatnecy )
lastLocalLatencies.append( lastLocalLatnecy )
firstGlobalLatencies.append( firstGlobalLatency )
lastGlobalLatencies.append( lastGlobalLatnecy )
# bring up link and withdraw intents
main.CLIs[0].link( main.end1[ 'port' ], main.end2[ 'port' ], "up",
timeout=main.timeout)
main.CLIs[0].pushTestIntents(main.ingress, main.egress, batchSize,
offset=1, options="-w", timeout=main.timeout)
main.CLIs[0].purgeWithdrawnIntents()
# check links, flows and intents
main.intentRerouteLatFuncs.sanityCheck( main, main.deviceCount * 2, 0, 0 )
if not main.verify:
continue
aveLocalLatency = numpy.average( lastLocalLatencies )
aveGlobalLatency = numpy.average( lastGlobalLatencies )
stdLocalLatency = numpy.std( lastLocalLatencies )
stdGlobalLatency = numpy.std( lastGlobalLatencies )
main.log.report( "Scale: " + str( main.numCtrls ) + " \tIntent batch: " + str( batchSize ) )
main.log.report( "Local latency average:................" + str( aveLocalLatency ) )
main.log.report( "Global latency average:................" + str( aveGlobalLatency ) )
main.log.report( "Local latency std:................" + str( stdLocalLatency ) )
main.log.report( "Global latency std:................" + str( stdGlobalLatency ) )
main.log.report( "________________________________________________________" )
if not ( numpy.isnan( aveLocalLatency ) or numpy.isnan( aveGlobalLatency ) ):
# check if got NaN for result
resultsDB = open( main.dbFileName, "a" )
resultsDB.write( "'" + main.commit + "'," )
resultsDB.write( str( main.numCtrls ) + "," )
resultsDB.write( str( batchSize ) + "," )
resultsDB.write( str( aveLocalLatency ) + "," )
resultsDB.write( str( stdLocalLatency ) + "\n" )
resultsDB.close()
del main.scale[ 0 ]