Performance Summary:

Validation Errors:

•   No Validation Errors Found

Physical Configuration:

Virtual Configuration:

Notes:

Physical System Notes

• Firmware settings used:
•   Power.PlatformControlledType=Efficiency - Favor Performance
•   PlatformControlledType=Maximum Performance

Storage Notes

• The 16 internal SSDs are configured into four RAID5 arrays, each having four disks.
• Each array appears as a disk in the host OS. Each disk is used to create an LVM volume group.
• Two LVM logical volumes for the SUT root file system and for the SUT swap space are allocated from one of the volume groups.
• The storage for each tile is allocated as an LVM logical volume from a volume group. Six partitions are created in the tile's logical volume, one for each VM in the tile.

Virtualization Software Notes

• The following boot options were used:
•   intel_iommu=on iommu=pt selinux=0 elevator=deadline cgroup_disable=memory,blkio
• The following services are configured to be disabled during runlevel initialization:
•   abrt-ccpp, abrt-oops, abrtd, atd, auditd, autofs, certmonger, cgconfig, cgred, crond, cups, dnsmasq, ebtables, ip6tables, iptables, irqbalance, iscsi, iscsid, ksm, ksmtuned, mdmonitor, netconsole, ntpdate, numad, oddjobd, postfix, psacct, quota_nld, radvd, rdisc, restorecond, rhnsd, rhsmcertd, rngd, rpcsvcgssd, saslauthd, smartd, sssd, vncserver, winbind, ypbind
• /etc/rc.local was modified to include:
•   echo NO_WAKEUP_PREEMPT >/sys/kernel/debug/sched_features
• The following items were added in /etc/sysctl.conf:
•   kernel.sched_min_granularity_ns = 4000000
•   kernel.sched_wakeup_granularity_ns = 4000000
•   kernel.sched_latency_ns = 32000000
• Network configuration is as follows:
•   Configuration file /etc/modprobe.d/ixgbe.conf for 10Gb Ethernet includes:
•    options ixgbe max_vfs=48
•    blacklist ixgbevf
•   116 virtual functions were used for all the VMs.
•   Ethernet devices eth2, eth4, eth8, and eth10 are the 4 physical functions that are used by
•   the host to provide physical transport for the virtual functions.
•   The PCI devices for the four physical functions are each routed to one of the four sockets (NUMA nodes).
•   Each VM is given a virtual function that is routed to the NUMA node to which the VM's VCPUs are pinned.
• VM configurations are as follows:
•   All VMs are managed with libvirt and created using the virsh utility.
•   All VMs are created with the following XML template (exceptions noted below template):
•    <domain type='kvm'>
•     <name>$vm_name</name>
•     <memory>$vm_memory</memory>
•     <currentMemory>$vm_memory</currentMemory>
•     <vcpu cpuset='$host-cpu-list'>$vm_vcpus</vcpu>
•     <os>
•      <type arch='x86_64' machine='pc'>hvm</type>
•      <boot dev='hd'/>
•     </os>
•     <features>
•      <acpi/>
•      <apic/>
•      <pae/>
•     </features>
•     <cpu match='exact'>
•      <model>Westmere</model>
•      <feature policy='force' name='x2apic'/>
•     </cpu>
•     <clock offset='utc'>
•      <timer name='pit' tickpolicy='delay'/>
•     </clock>
•     <on_poweroff>destroy</on_poweroff>
•     <on_reboot>restart</on_reboot>
•     <on_crash>restart</on_crash>
•     <devices>
•      <emulator>/usr/libexec/qemu-kvm</emulator>
•      <disk type='block' device='disk'>
•       <driver name='qemu' type='raw' cache='none' io='native'/>
•       <source dev='$vm_disk'/>
•       <target dev='vda' bus='virtio'/>
•      </disk>
•      <hostdev mode='subsystem' type='pci' managed='no'>
•       <source>
•        <address domain='0x0000' bus='$bus' slot='$slot' function='$function'/>
•       </source>
•      </hostdev>
•      <serial type='pty'>
•       <target port='0'/>
•      </serial>
•      <console type='pty'>
•       <target type='serial' port='0'/>
•      </console>
•      <input type='mouse' bus='ps2'/>
•      <graphics type='vnc' port='-1' autoport='yes' keymap='en-us'/>
•      <memballoon model='none'/>
•     </devices>
•    </domain>
•   Web VMs have an addition disk device for holding the web server logs, and therefore have the following addition disk declaration:
•      <disk type='block' device='disk'>
•       <driver name='qemu' type='raw' cache='none' io='native'/>
•       <source dev='$web_log_disk'/>
•       <target dev='vdb' bus='virtio'/>
•       <alias name='virtio-disk1'/>
•       <address type='pci' domain='0x0000' bus='0x00' slot='0x05' function='0x0'/>
•      </disk>
•    VMs are bound to host CPUs via their XML configuration in the following manner:
•     All VMs of tiles 1, 2, 3, and 4 are bound to NUMA node 0.
•     All VMs of tiles 5, 6, 7, and 8 are bound to NUMA node 1.
•     All VMs of tiles 9, 10, 11, and 12 are bound to NUMA node 2.
•     All VMs of tiles 13, 14, 15, and 16 are bound to NUMA node 3.
•     Tile 17 App and Db VMs, tile 18 App VM, and tile 21 Web and Infra VMs are bound to NUMA node 0.
•     Tile 17 Web, Infra, Mail, and Batch VMs, tile 18 Web, Infra, Mail, and Batch VMs, and tile 19 App VM are bound to NUMA node 1.
•     Tile 19 Mail VM, tile 20 App, Mail, and Batch VMs, tile 21 App, Db, Mail, and Batch VMs, and tile 22 App, Mail, and Batch VMs are bound to NUMA node 2.
•     Tile 19 Web, Infra, and Batch VMs, tile 20 Web and Infra VMs, and tile 22 Web and Infra VMs are bound to NUMA node 3.

Web Server VM Notes

• The following boot options were used:
•   selinux=0 audit=0 elevator=deadline
• The following services are disabled:
•   abrt-ccpp, abrtd, atd, auditd, autofs, certmonger, cgconfig, cgred, crond, cups, haldaemon, htcacheclean, httpd, ip6tables, iptables, irqbalance, kdump, mcelogd, messagebus, netconsole, nfs, nfslock, ntpd, ntpdate, numad, oddjobd, portreserve, postfix, psacct, quota_nld, rdisc, restorecond, rhnsd, rhsmcertd, rngd, rpcgssd, rpcsvcgssd, saslauthd, smartd, sssd, svnserve, winbind, ypbind
• /etc/rc.local was modified to include:
•   echo tsc >/sys/devices/system/clocksource/clocksource0/current_clocksource
• httpd.conf changes:
•   KeepAlive on
•   KeepAliveTimeout 60
•   MaxKeepAliveRequest 0
•   ServerLimit 10240 (for prefork)
•   MaxClients 10240 (for prefork)
• Web VM used NFS to access files from infra VM.
• No options were specified in the NFS mount.
• The tuned-0.2.19 package was installed and activated with virtual-guest profile.
• After Infra VM is booted and its NFS service is up, the Web VM mounts the NFS file-system via test automation software.
• root file-system options: barrier=0
• The following items were added in /etc/sysctl.conf:
•   fs.file-max = 500000
•   net.core.somaxconn = 81920
•   net.ipv4.tcp_max_tw_buckets = 1500000
•   net.ipv4.conf.default.arp_filter = 1
•   net.ipv4.conf.all.arp_filter = 1
•   net.ipv4.tcp_tw_reuse=1
•   net.ipv4.tcp_keepalive_time = 120
•   net.ipv4.tcp_keepalive_intvl = 30
•   net.ipv4.tcp_timestamps = 0
•   vm.swappiness=0
• The datastore, /dev/vdb, is mounted on /var/log/httpd

Application Server VM Notes

• The following boot options were used:
•   selinux=0 audit=0 elevator=deadline
• The following services are disabled:
•   abrt-ccpp, abrtd, atd, auditd, autofs, certmonger, cgconfig, cgred, crond, cups, haldaemon, ip6tables, iptables, irqbalance, kdump, mcelogd, messagebus, netconsole, nfs, nfslock, ntpd, ntpdate, numad, oddjobd, portreserve, postfix, psacct, quota_nld, rdisc, restorecond, rhnsd, rhsmcertd, rngd, rpcgssd, rpcsvcgssd, saslauthd, smartd, sssd, svnserve, winbind, ypbind
• /etc/rc.local was modified to include:
•   echo tsc >/sys/devices/system/clocksource/clocksource0/current_clocksource
• There are two WebSphere profiles, AppSrv01 for SPECjAppServer2004 and AppSrv02 for SPEC Emulator.
• Heap size for AppSrv01 has been changed to 2048 MB.
• The tuned-0.2.19 package was installed and activated with virtual-guest profile.
• root file-system options: barrier=0
• The following items were added in /etc/sysctl.conf:
•   fs.file-max = 500000
•   net.core.somaxconn = 8192
•   net.ipv4.tcp_max_tw_buckets = 500000
•   net.ipv4.conf.default.arp_filter = 1
•   net.ipv4.conf.all.arp_filter = 1
•   vm.swappiness=0
• The following changes were made to the defaults in run.properties:
•   runLO=4
•   warmUpStyle = 1
•   linearWarmUpPercentage = .25

Mail Server VM Notes

• The following boot options were used:
•   selinux=0 audit=0 elevator=deadline
• The following services are disabled:
•   abrt-ccpp, abrtd, atd, auditd, autofs, certmonger, cgconfig, cgred, crond, cups, haldaemon, ip6tables, iptables, irqbalance, kdump, mcelogd, messagebus, netconsole, nfs, nfslock, ntpd, ntpdate, numad, oddjobd, portreserve, postfix, psacct, quota_nld, rdisc, restorecond, rhnsd, rhsmcertd, rngd, rpcgssd, rpcsvcgssd, saslauthd, smartd, sssd, svnserve, winbind, ypbind
• /etc/rc.local was modified to include:
•   echo tsc >/sys/devices/system/clocksource/clocksource0/current_clocksource
• Dovecot was compiled with default options and the following dovecot.conf was used:
•   base_dir = /var/run/dovecot/
•   ssl = no
•   disable_plaintext_auth = no
•   listen = *
•   log_path = /var/log/dovecot.log
•   mail_debug = no
•   mail_location = dbox:/mailstore/dovecot/maildir/%u
•   mdbox_rotate_size = 2048
•   passdb {
•    args = /etc/passwd.dovecot
•    driver = passwd-file
•   }
•   protocols = imap
•   service auth {
•    user = root
•   }
•   service imap-login {
•    inet_listener imap {
•     port = 143
•    }
•    process_min_avail = 10
•   }
•   userdb {
•    driver = passwd
•   }
• The tuned-0.2.19 package was installed and activated with virtual-guest profile.
• root file-system options: barrier=0
• The following items were added in /etc/sysctl.conf:
•   fs.file-max = 500000
•   net.core.somaxconn = 8192
•   net.ipv4.tcp_max_tw_buckets = 500000
•   net.ipv4.conf.default.arp_filter = 1
•   net.ipv4.conf.all.arp_filter = 1

Database Server VM Notes

• The following boot options were used:
•   selinux=0 audit=0 elevator=deadline
• The following services are disabled:
•   atd, auditd, autofs, cgconfig, cgred, crond, htcacheclean, httpd, ip6tables, iptables, netconsole, nfs, ntpd, ntpdate, numad, oddjobd, portreserve, postfix, psacct, quota_nld, rdisc, restorecond, rhnsd, rhsmcertd, rngd, rpcgssd, rpcsvcgssd, saslauthd, smartd, sssd, svnserve, winbind, ypbind
• /etc/rc.local was modified to include:
•   echo tsc >/sys/devices/system/clocksource/clocksource0/current_clocksource
• Database tuning options:
•   db2 update db cfg for specdb using DATABASE_MEMORY 6100000
•   db2set DB2_LARGE_PAGE_MEM=DB; db2set DB2_PINNED_BP=ON
•   alter bufferpool IBMDEFAULTBP size 5500000
•   alter tablespace roll_space NO FILE SYSTEM CACHING
•   alter tablespace C_cust_space NO FILE SYSTEM CACHING
•   alter tablespace C_supp_space NO FILE SYSTEM CACHING
•   alter tablespace C_site_space NO FILE SYSTEM CACHING
•   alter tablespace C_parts_space NO FILE SYSTEM CACHING
•   alter tablespace C_custinv_space NO FILE SYSTEM CACHING
•   alter tablespace M_parts_space NO FILE SYSTEM CACHING
•   alter tablespace M_bom_space NO FILE SYSTEM CACHING
•   alter tablespace M_wo_space NO FILE SYSTEM CACHING
•   alter tablespace M_lo_space NO FILE SYSTEM CACHING
•   alter tablespace M_inv_space NO FILE SYSTEM CACHING
•   alter tablespace O_cust_space NO FILE SYSTEM CACHING
•   alter tablespace O_ords_space NO FILE SYSTEM CACHING
•   alter tablespace O_ordl_space NO FILE SYSTEM CACHING
•   alter tablespace O_item_space NO FILE SYSTEM CACHING
•   alter tablespace S_comp_space NO FILE SYSTEM CACHING
•   alter tablespace S_sc_space NO FILE SYSTEM CACHING
•   alter tablespace S_po_space NO FILE SYSTEM CACHING
•   alter tablespace S_site_space NO FILE SYSTEM CACHING
•   alter tablespace S_supp_space NO FILE SYSTEM CACHING
•   alter tablespace U_space NO FILE SYSTEM CACHING
•   update database config for specdb using LOGFILSIZ 65536
•   update db cfg for specdb using logprimary 64 logsecond 4
• The tuned-0.2.19 package was installed and activated with virtual-guest profile.
• 12288 2MB huge-pages were allocated prior to starting the database instance.
• root file-system options: barrier=0
• The following items were added in /etc/sysctl.conf:
•   fs.file-max = 500000
•   net.core.somaxconn = 32384
•   net.ipv4.tcp_max_tw_buckets = 1000000
•   net.ipv4.conf.default.arp_filter = 1
•   net.ipv4.conf.all.arp_filter = 1
•   vm.swappiness=0
•   kernel.sem = 500 64000 400 512

Infraserver VM Notes

• The following boot options were used:
•   selinux=0 audit=0 elevator=deadline
• The following services are disabled:
•   abrtd, autofs, certmonger, crond, cups, haldaemon, htcacheclean, httpd, ip6tables, iptables, irqbalance, kdump, messagebus, netconsole, ntpd, ntpdate, oddjobd, portreserve, postfix, psacct, quota_nld, rdisc, restorecond, rhnsd, rhsmcertd, rngd, rpcgssd, rpcsvcgssd, saslauthd, smartd, sssd, tomcat6, winbind, ypbind
• /etc/rc.local was modified to include:
•   echo tsc >/sys/devices/system/clocksource/clocksource0/current_clocksource
• httpd.conf changes:
•   Listen port to 81
•   ServerLimit 10240 (for prefork)
•   MaxClients 10240 (for prefork)
• /etc/httpd/conf/fcgi.conf added:
•   LoadModule fastcgi_module modules/mod_fastcgi.so
•   ScriptAlias /fcgi-bin/ /var/www/fcgi-bin/
•   FastCgiIpcDir /etc/httpd/fastcgi
•   <Directory /var/www/fcgi-bin>
•    AllowOverride None
•    Options +ExecCGI -Includes
•    SetHandler fastcgi-script
•    Order allow,deny
•    Allow from all
•   </Directory>
•   AddHandler fastcgi-script fcgi
•   FastCgiServer /var/www/fcgi-bin/besim_fcgi.fcgi -processes 50
• Standard Linux NFS server is enabled to serve files to Web VM.
• /etc/exports file contains:
•   /support-download *(rw,no_subtree_check,crossmnt,no_root_squash)
• fcgi package was built from Fedora 14 source RPM, which includes patch to compile with GCC 4.4.
• The tuned-0.2.19 package was installed and activated with virtual-guest profile.
• root file-system options: noatime,barrier=0,data=writeback
• The following items were added in /etc/sysctl.conf:
•   fs.file-max = 500000
•   net.core.somaxconn = 81920
•   net.ipv4.tcp_max_tw_buckets = 1500000
•   net.ipv4.conf.default.arp_filter = 1
•   net.ipv4.conf.all.arp_filter = 1
•   net.ipv4.tcp_tw_reuse=1
•   net.ipv4.tcp_keepalive_time = 120
•   net.ipv4.tcp_keepalive_intvl = 30
•   net.ipv4.tcp_timestamps = 0
•   net.ipv4.tcp_syn_retries = 4
•   net.ipv4.tcp_retries2 = 7
•   vm.swappiness=0

Batch Server VM Notes

• The following boot options were used:
•   selinux=0 audit=0 elevator=deadline
• The following services are disabled:
•   abrt-ccpp, abrtd, atd, auditd, autofs, certmonger, cgconfig, cgred, crond, cups, haldaemon, ip6tables, iptables, irqbalance, kdump, mcelogd, messagebus, netconsole, nfs, nfslock, ntpd, ntpdate, numad, oddjobd, portreserve, postfix, psacct, quota_nld, rdisc, restorecond, rhnsd, rhsmcertd, rngd, rpcgssd, rpcsvcgssd, saslauthd, smartd, sssd, svnserve, winbind, ypbind
• /etc/rc.local was modified to include:
•   echo tsc >/sys/devices/system/clocksource/clocksource0/current_clocksource
• The tuned-0.2.19 package was installed and activated with virtual-guest profile.
• root file-system options: barrier=0
• The following items were added in /etc/sysctl.conf:
•   fs.file-max = 500000
•   net.core.somaxconn = 8192
•   net.ipv4.tcp_max_tw_buckets = 500000
•   net.ipv4.conf.default.arp_filter = 1
•   net.ipv4.conf.all.arp_filter = 1

Client Driver Notes

• The first four client systems (IBM System x3630 M4) each host prime client VMs using Red Hat Enterprise Linux 6.4. Two system host 5 client VMs. Two systems host 4 client VMs. The fifth client system (IBM System x3650 M2) is the prime controller.
• Each prime client VM has 4 VCPUs, 8192 MB memory, and a 32 GB disk. 1 virtio_net device and 1 virtio_blk device are used in all client VMs.

Other Notes

• 1 x Blade RackSwitch G8124E Ethernet switch was used to connect the clients to the SUT.

Performance Details: