Product Support > SPECvirt_sc2013
runspecvirt fail
zgy:
HI all:
as i run ./runspecvirt.sh , it fail, shows like this:
[root@client1 SPECvirt]# tail -50 primectrl.out
2017-09-07 16:38:05:383 PrimeControl: Master 1: client1:1098
2017-09-07 16:38:05:383 PrimeControl: Master 2: client1:1096
2017-09-07 16:38:05:383 PrimeControl: Master 3: client1:1094
2017-09-07 16:38:05:383 PrimeControl: Master 4: client1:1092
2017-09-07 16:38:05:385 PrimeControl: adding host client1:1098
2017-09-07 16:38:05:394 PrimeControl: adding host client1:1096
2017-09-07 16:38:05:399 PrimeControl: adding host client1:1094
2017-09-07 16:38:05:405 PrimeControl: adding host client1:1092
2017-09-07 16:38:05:420 First client for 0: 172.21.128.219:1091
2017-09-07 16:38:05:434 First client for 1: 172.21.128.219:1010
2017-09-07 16:38:05:444 First client for 2: 172.21.128.219:1200
2017-09-07 16:38:05:450 First client for 3: 172.21.128.219:1902
2017-09-07 16:38:05:454 PrimeControl: starting clients...
2017-09-07 16:38:05:454 PrimeControl: clients.length = 4
2017-09-07 16:38:05:454 PrimeControl: clients[0].length = 1
2017-09-07 16:38:05:454 PrimeControl: starting clients[0][0]: 172.21.128.219:1091
2017-09-07 16:38:05:475 PrimeControl: started client: 172.21.128.219:1091
2017-09-07 16:38:05:475 PrimeControl: clients[1].length = 1
2017-09-07 16:38:05:476 PrimeControl: starting clients[1][0]: 172.21.128.219:1010
2017-09-07 16:38:05:482 PrimeControl: started client: 172.21.128.219:1010
2017-09-07 16:38:05:482 PrimeControl: clients[2].length = 1
2017-09-07 16:38:05:482 PrimeControl: starting clients[2][0]: 172.21.128.219:1200
2017-09-07 16:38:05:491 PrimeControl: started client: 172.21.128.219:1200
2017-09-07 16:38:05:491 PrimeControl: clients[3].length = 1
2017-09-07 16:38:05:491 PrimeControl: starting clients[3][0]: 172.21.128.219:1902
2017-09-07 16:38:05:510 PrimeControl: started client: 172.21.128.219:1902
2017-09-07 16:38:05:517 PrimeControl: PTDs not used for this benchmark run!
2017-09-07 16:38:05:517 PrimeControl: starting 4 masters.
2017-09-07 16:38:39:648 PrimeControl: master[0][0] sleeping 20 sec.
2017-09-07 16:38:39:648 PrimeControl: master[0][1] sleeping 20 sec.
2017-09-07 16:38:39:648 PrimeControl: master[0][2] sleeping 20 sec.
2017-09-07 16:38:39:648 PrimeControl: master[0][3] sleeping 20 sec.
2017-09-07 16:38:59:687 PrimeControl: waiting on 4 prime client(s).
2017-09-07 16:39:00:146 Sending config to client1:1092
2017-09-07 16:39:00:151 Sending config to client1:1098
2017-09-07 16:39:00:158 Sending config to client1:1094
2017-09-07 16:39:00:200 Sending config to client1:1096
2017-09-07 16:39:00:344 specvirt: waiting on 3 prime clients.
2017-09-07 16:39:00:618 specvirt: waiting on 2 prime clients.
.
2017-09-07 16:39:00:721 setting hostsReady = true
2017-09-07 16:39:09:684 specvirt: waiting on 1 prime clients.
2017-09-07 16:41:15:791 [ERROR] Received abort signal from client1:1098. Terminating.
2017-09-07 16:41:15:791 PrimeControl: sending abortTest() to prime clients.
2017-09-07 16:41:15:792 PrimeControl: id=1, abortID=0
2017-09-07 16:41:15:792 PrimeControl: masters[1]=client1:1096
2017-09-07 16:41:15:792 PrimeControl: id=3, abortID=0
2017-09-07 16:41:15:792 PrimeControl: id=2, abortID=0
2017-09-07 16:41:15:792 PrimeControl: masters[2]=client1:1094
2017-09-07 16:41:15:792 PrimeControl: masters[3]=client1:1092
so i touch client-1098 log, and it shows:
prime-client1_1098:
[root@client1 SPECvirt]# tail -50 logs/20170907-163805/prime-client1_1098.log
2017-09-07 16:38:59:890 Looking up SPECvirt controller: client1
2017-09-07 16:39:00:150 masterID: 0, tile: 0, workload: 0
2017-09-07 16:39:00:151 hostname: client1
Hostname of prime client: client1
2017-09-07 16:39:00:219 Thu Sep 07 16:38:05 CST 2017
2017-09-07 16:39:00:272 RMI server started: client1:9900
2017-09-07 16:39:00:273 Total clients: 1
2017-09-07 16:39:00:273 Adding host client1:1091
2017-09-07 16:39:00:291 Setting up clients...
2017-09-07 16:39:00:717 calling getHostVM() on jappclient...
2017-09-07 16:41:15:786 Remote exception in setup() from client1:1091
java.rmi.ServerException: RemoteException occurred in server thread; nested exception is:
java.rmi.RemoteException: Connection refused (Connection refused); nested exception is:
java.lang.Exception: Connection refused (Connection refused)
2017-09-07 16:41:15:787 Aborting...
2017-09-07 16:41:15:791 abortTest() signal sent...
2017-09-07 16:41:15:791 clientsRunning = true
2017-09-07 16:41:15:792 calling client.exit()...
PLS, the pollme is running in ALL vms, iptables is stop, and SSLv3 is removed from the jdk.tls.disabledAlgorithms property in the java.security file( /usr/lib/jvm/java-1.8.0-openjdk-1.8.0.141-3.b16.el6_9.x86_64/jre/lib/security/java.security),
Run pollmecheck.sh from client shows:
[root@client1 SPECvirt]# ./pollmecheck.sh
Tile 1 VMs:
dbserver1:
root@dbserver1's password:
root 3643 1 0 16:31 pts/0 00:00:01 java -jar /opt/SPECpoll/pollme.jar -n dbserver -p 8001
appserver1:
root@appserver1's password:
root 4025 1 0 16:30 pts/0 00:00:01 java -jar /opt/SPECpoll/pollme.jar -n appserver -p 8001
batchserver1:
root@batchserver1's password:
root 2362 1 0 16:27 ? 00:00:01 java -jar /opt/SPECpoll/pollme.jar -n batchserver -p 8001
mailserver1:
root@mailserver1's password:
root 2351 1 0 16:26 ? 00:00:02 java -jar /opt/SPECpoll/pollme.jar -n mailserver -p 8001
infraserver1:
root@infraserver1's password:
root 3375 1 0 16:32 pts/0 00:00:01 java -jar /opt/SPECpoll/pollme.jar -n infraserver -p 8001
webserver1:
root@webserver1's password:
root 3512 1 0 16:29 pts/0 00:00:02 java -jar /opt/SPECpoll/pollme.jar -n webserver -p 8001
Polleme.out:
[root@infraserver ~]# cat /tmp/pollme.out
Creating RMI listener using RMI Registry port 8001
infraserver/172.21.128.220:8001 ready...
infraserver: 1504773548135
[root@webserver ~]# cat /tmp/pollme.out
Creating RMI listener using RMI Registry port 8001
webserver/172.21.128.221:8001 ready...
2017-09-07 16:39:07:755 contacting host 'infraserver'...
webserver: 1504773547774
infraserver: 1504773548135
[root@mailserver ~]# cat /tmp/pollme.out
Creating RMI listener using RMI Registry port 8001
mailserver/172.21.128.222:8001 ready...
mailserver: 1504773534638
[root@mailserver ~]#
[root@appserver ~]# cat /tmp/pollme.out
Creating RMI listener using RMI Registry port 8001
appserver/172.21.128.224:8001 ready...
[root@appserver ~]#
[root@dbserver ~]# cat /tmp/pollme.out
Creating RMI listener using RMI Registry port 8001
dbserver/172.21.128.225:8001 ready...
[root@dbserver ~]#
[root@batchserver ~]# cat /tmp/pollme.out
Creating RMI listener using RMI Registry port 8001
batchserver/172.21.128.227:8001 ready...
batchserver: 1504773537995
[root@batchserver ~]#
My etc/hosts is:
# External VM-to-client communications
172.21.128.220 infraserver infraserver1
172.21.128.221 webserver webserver1
172.21.128.222 mailserver mailserver1
172.21.128.224 appserver appserver1 specdelivery specemulator
172.21.128.225 dbserver dbserver1
172.21.128.227 batchserver batchserver1
172.21.128.219 client1 specdriver
# Internal VM-to-VM only communications
172.21.128.240 infraserver1-int
172.21.128.241 webserver1-int
172.21.128.244 appserver1-int
172.21.128.245 dbserver1-int specdb
java version is 1.8.0_141 expect appserver is 1.7.0_151, because java 1.8 is incompatible with glassfish3, so i use 1.7.0_151 in appserver.
How to deal whit it now?
attachments is the client log and primectl.out, also my Control.config in it , please help have a look, thanks so much.
lroderic:
Thanks for all the logs. The problem is with appserver's internal network connection to dbserver evidenced in prime-client1_1098.log which gets reported into primectrl.out. Let's work on that without the noise of the other workloads. In Control.config, please change NUM_WORKLOADS=1 to run only app/dbserver.
You're not using an internal network since both the internal vNICs and external vNICs are on 172.21.128. The point of using internal vNICs for communications between appserver and dbserver is to move that intra-VM network traffic off of the LAN between the client and workload VMs, reducing network traffic between the client and workload VMs. The error in client-172.21.128.219_1091.log means that SPECjAppServer is trying to connect to a hostname it can't find.
To make the example VM configuration work, please create a "dummy" internal network that lets VMs communicate only with each other. If you use vSphere, see https://communities.vmware.com/thread/234233 to set up a "host-only" network. If you use KVM, see http://www.linux-kvm.org/page/Networking#Private_Virtual_Bridge for setting up a private virtual bridge.
Once you create the internal network and assign a second vNIC on each VM to it, edit the VM's /etc/hosts to reflect the IP addresses for those vNICS. The Example VM guide states in Section 2.0:
--- Code: ---Each VM must have at least one virtual network adapter associated with a physical network that can communicate
with the client. Review the /etc/hosts file for each cloned VM for the IP address assignments. Assign the first network
adapter (eth0) to the external network (for example 192.168). Assign the second network adapter (eth1) to the internal
network (for example 10.10).
--- End code ---
Using this example, we'll use 10.10.1 for the internal network on app/db and web/infra VMs:
--- Code: ---# External VM-to-client communications
172.21.128.220 infraserver infraserver1
172.21.128.221 webserver webserver1
172.21.128.222 mailserver mailserver1
172.21.128.224 appserver appserver1 specdelivery specemulator
172.21.128.225 dbserver dbserver1
172.21.128.227 batchserver batchserver1
# Internal VM-to-VM only communications
10.10.1.220 infraserver1-int
10.10.1.221 webserver1-int
10.10.1.224 appserver1-int
10.10.1.225 dbserver1-int specdb
--- End code ---
* Reboot the workload VMs and check that SPECpoll has started on each of them.
* Make sure that you can ping between appserver1-int and dbserver1-int to see that the internal network is up.
* Run the helper script at /opt/SPECvirt/helper-sc2013/jApp-testappserver.sh to see that SPECjAppServer emulator, delivery, and dbserver are up and working.
* Then rerun the test with only app/dbserver (NUM_WORKLOADS=1). You could also do a search on "SPECpoll" here in the forum to see others who've had problems with getting SPECpoll working.
Lisa
zgy:
HI Lisa, Thanks for your help.
After I use 10.10.1 for the internal network, I rerun runspecvirt.sh , now the error info is:
2017-09-08 22:10:34:097 SslConnection: [ERROR] IOException during SSL handshake: javax.net.ssl.SSLHandshakeException: No appropriate protocol (protocol is disabled or cipher suites are inappropriate)
2017-09-08 22:10:34:097 Connection: [ERROR] createSocket() failed.
2017-09-08 22:10:34:097 HttpRequestSched: Creating socket...
2017-09-08 22:10:34:097 SslConnection: [ERROR] IOException during SSL handshake: javax.net.ssl.SSLHandshakeException: No appropriate protocol (protocol is disabled or cipher suites are inappropriate)
2017-09-08 22:10:34:097 Connection: [ERROR] createSocket() failed.
2017-09-08 22:10:34:097 HttpRequestSched: Creating socket...
2017-09-08 22:10:34:097 SslConnection: [ERROR] IOException during SSL handshake: javax.net.ssl.SSLHandshakeException: No appropriate protocol (protocol is disabled or cipher suites are inappropriate)
2017-09-08 22:10:34:098 Connection: [ERROR] createSocket() failed.
2017-09-08 22:10:34:098 HttpRequestSched: Creating socket...
After I did below, but The error still exists
1:Remove SSLv3 and MD5withRSA from the jdk.tls.disabledAlgorithms property in the $JDK_HOME/jre/lib/security/java.security file (SPECvirt assumes an MD5-based SSLv3 cipher).
2:Add SSL_RSA_WITH_RC4_128_MD5 to the jdk.tls.legacyAlgorithms list in the $JDK_HOME/jre/lib/security/java.security file.
3:Remove !MD5 from the SSLCipherSuite list in /etc/httpd/conf.d/ssl.conf file.
lroderic:
Hi. Your app/dbserver got further, but since you ran with NUM_WORKLOADS=4, we can't tell if your changes to the hosts file for app/dbserver worked. I would've made sure that was working before moving on to web/infraserver, but OK.
Thanks for reading the Technical Support FAQ. You're using Java 1.8.0_141 on all VMs except for the appserver? You made those three SSL changes on the client, correct?
Lisa
zgy:
YES, I using java 1.8.0_141 on all VMs expect appserver,
I made those threre SSL changes on all VMs, it is right?
Or just for the client VM?
Navigation
[0] Message Index
[#] Next page
Go to full version