A.2.2 Example of the Physical Partition Configuration Procedure
A.2.2 Example of the Physical Partition Configuration Procedure
- Log in to the master XSCF.
Execute the showbbstatus command to check that the XSCF to which you have logged in is the master XSCF.
If you have logged in to a standby XSCF, log out and then log in to the master XSCF again.
| XSCF> showbbstatus BB#00 (Master) |
- Set mirror mode for memory.
- a. Execute the showfru command to check mirror mode of the memory on the SPARC M10-4S system board (PSB<BB>).The following example displays the setting information for devices under PSB 00-0.
| XSCF> showfru sb 00-0 Device Location Memory Mirror Mode sb 00-0 cpu 00-0-0 no cpu 00-0-1 no cpu 00-0-2 no cpu 00-0-3 no |
- b. To use the memory mirror mode, execute the setupfru command to set that mode.If you are not using memory mirror mode, this step is not necessary.For details on memory mirror mode, see "14.1 Configuring Memory Mirroring" in the Fujitsu SPARC M12 and Fujitsu M10/SPARC M10 System Operation and Administration Guide.The following example places all the CPUs under PSB 00-0 in memory mirror mode.
| XSCF> setupfru -m y sb 00-0 |
| Note - When SPARC M12 is used, specify -c mirror=yes to create a mirroring configuration for the memory. Example: XSCF> setupfru -c mirror=yes sb 00-0 |
- Execute the showfru command to check the setting of memory mirror mode.
| XSCF> showfru sb 00-0 Device Location Memory Mirror Mode sb 00-0 cpu 00-0-0 yes cpu 00-0-1 yes cpu 00-0-2 yes cpu 00-0-3 yes |
- Create the physical partition configuration information.
- a. Execute the showpcl command to check the physical partition configuration information.
| XSCF> showpcl -p 0 PPAR-ID LSB PSB Status |
- b. Execute the setpcl command to register the system board in the physical partition configuration information.Execute the setpcl command to register the SPARC M10-4S system board in the physical partition configuration information for the built-in destination.In the following example, physical system boards (PSB 00-0 and PSB 01-0) are mapped to logical system boards (LSB 00 and LSB 01) of physical partition 0.
| XSCF> setpcl -p 0 -a 00=00-0 01=01-0 |
- c. Execute the showpcl command to check the physical partition configuration information.Check the physical partition configuration information.
| XSCF> showpcl -p 0 PPAR-ID LSB PSB Status 00 Running 00 00-0 01 01-0 |
- Specify the -v option to display details on the configuration policy, IO nullification option (No-IO), and memory nullification option (No-Mem).
| XSCF> showpcl -v -p 0 PPAR-ID LSB PSB Status No-Mem No-IO Cfg-policy 00 Running System 00 00-0 False False 01 01-0 False False |
- Use the setpcl command to change the setting of the configuration policy, IO nullification option (No-IO), and memory nullification option (No-Mem).For details on the setpcl command, see the Fujitsu SPARC M12 and Fujitsu M10/SPARC M10 XSCF Reference Manual.
- Assign a system board (PSB<BB>) to a physical partition.
- a. Execute the showboards-a command to check the PSB status.Execute the showboards -a command to check that each PSB status is "SP" (system board pool).
| XSCF> showboards -a PSB PPAR-ID(LSB) Assignment Pwr Conn Conf Test Fault ---- ------------ ----------- ---- ---- ---- ------- -------- 00-0 SP Available n n n Passed Normal 01-0 SP Available n n n Passed Normal |
- b. Execute the addboard -c assign command to assign the PSBs.
| XSCF> addboard -c assign -p 0 00-0 01-0 |
- c. Execute the showboards-p command to check the PSB status.Execute the showboards -p command to check the status of each PSB assigned to the physical partition.This example checks that the [Assignment] field of each PSB becomes "Assigned" since each PSB has been normally assigned to physical partition 0.
| XSCF> showboards -p 0 PSB PPAR-ID(LSB) Assignment Pwr Conn Conf Test Fault ---- ------------ ----------- ---- ---- ---- ------- -------- 00-0 00(00) Assigned n n n Passed Normal 01-0 00(01) Assigned n n n Passed Normal |
- Register the CPU Activation key to assign CPU core resources.
- a. Execute the showcodactivation command to check the information on the CPU Activation key.Execute the showcodactivation command to check whether the physical partition contains an assignable CPU Activation key.If only the header is displayed, the CPU Activation key is not registered in the XSCF.
| XSCF> showcodactivation Index Description Count ------- ----------- ------ |
| Note - If the number of registered CPU Activations is not enough for the number of CPUs to be used, purchase CPU Activations and add the CPU Activation keys. |
- b. Execute the addcodactivation command to register the CPU Activation key.For details on how to register a CPU Activation key, see "5.3 Adding CPU Core Resources" in the Fujitsu SPARC M12 and Fujitsu M10/SPARC M10 System Operation and Administration Guide.
| XSCF> addcodactivation "Product: SPARC M10-4S SequenceNumber:10005 Cpu: noExpiration 2 Text-Signature-SHA256-RSA2048: PSSrElBrse/r69AVSVFd38sT6AZm2bxeUDdPQHKbtxgvZPsrtYguqiNUieB+mTDC : : b1GCkFx1RH27FdVHiB2H0A==" AboveKeywillbeadded,Continue?[y|n]:y |
- c. Execute the showcodactivation command to check the information on the CPU Activation keys.Execute the showcodactivation command to check whether the physical partition contains an assignable CPU Activation key.
| XSCF> showcodactivation Index Description Count ------- ----------- ------ 0 PROC 2 1 PROC 2 2 PROC 2 3 PROC 2 --- Omitted --- 62 PROC 2 63 PROC 2 |
- d. Execute the setcod command to assign CPU core resources.Execute the setcod command to assign the CPU core resources to the physical partition.The following example assigns 128 CPU core resources to physical partition 0.
| XSCF> setcod -p 0 -s cpu -c set 128 PROC Permits assigned for PPAR 0 : 0 -> 128 PROC Permits assigned for PPAR will be changed. Continue? [y|n] :y Completed. |
| Note - XSCF firmware of versions XCP 2250 and earlier do not support the -c add, -c delete, and -c set options. Specify the options of the setcod command as shown below to interactively add and delete CPU core resources. XSCF> setcod -s cpu |
- Execute the showcod command to check information on the CPU core resources assigned to the physical partition.The following example confirms that 128 CPU core resources have been assigned to physical partition 0 with the setcod command that was just executed.
| XSCF> showcod -p 0 PROC Permits assigned for PPAR 0: 128 |
- Execute the resetdateoffset command to reset the difference with the time managed by the XSCF.
Execute the resetdateoffset command to reset the difference between the time managed by the XSCF and the time managed by the physical partitions.
| XSCF> resetdateoffset -p 0 |
- Execute the showpparmode command to check the setting of the detail level of the diagnosis messages and that of the PPAR DR mode.
Execute the showpparmode command to check that the detail level (Message Level) of the diagnosis message is "normal" (standard) and that Next of the PPAR DR mode is set to "on" (enabled).
| XSCF> showpparmode -p 0 Host-ID :9007002b Diagnostic Level :min Message Level :normal Alive Check :on Watchdog Reaction :reset Break Signal :on Autoboot(Guest Domain) :on Elastic Mode :off IOreconfigure :false PPAR DR(Current) :- PPAR DR(Next) :on |
- If the detail level of the diagnosis message is other than "normal", execute the setpparmode command to set it to "normal".
For details on the setpparmode command, see the Fujitsu SPARC M12 and Fujitsu M10/SPARC M10 XSCF Reference Manual.
| XSCF> setpparmode -p 0 -m message=normal |
- If PPAR DR mode is set to "off" (disabled), execute the setpparmode command to set it to "on".
| XSCF> setpparmode -p 0 -m ppar_dr=on |
- Execute the poweron command to power on the physical partitions.
| XSCF> poweron -p 0 |
- Execute the console command to connect the console to the physical partition.
| XSCF> console -p 0 |
- Install Oracle Solaris and Oracle VM Server for SPARC.
Install, on the control domain, Oracle Solaris and Oracle VM Server for SPARC.
- For details on the installation, see the following documents, presented on the Oracle Corporation web site (https://docs.oracle.com/en/).- Oracle Solaris 10Oracle Solaris 10 1/13 Installation Guide- Oracle Solaris 11Installing Oracle Solaris 11.x Systems- Oracle VM Server for SPARC"Installing and Enabling Software" in the Oracle VM Server for SPARC Administration Guide
- Configure a logical domain.
This item describes how to configure the logical domains defined in "Table A-1 Example of the 2BB Configuration With No Empty Hardware Resources."
- a. Reduce the number of control domain resources.For the factory-default configuration, all the CPU cores, memory, and the PCIe root complexes are assigned to the control domain (primary). To allow these resources to be assigned to other logical domains, first delete some of the resources from the control domain and configure the console service.
- The following shows command execution examples.Execute the ldm start-reconf command to switch to delayed reconfiguration mode.
| # ldm start-reconf primary Initiating a delayed reconfiguration operation on the primary domain. All configuration changes for other domains are disabled until the primary domain reboots, at which time the new configuration for the primary domain will also take effect. |
- Remove the root complex with the ldm remove-io command.The following example partially describes the command for removing PCIE1, PCIE2, PCIE3, PCIE5, PCIE6, PCIE7, PCIE9, PCIE10, PCIE11, PCIE13, PCIE14, and PCIE15 according to the configuration example.
| # ldm remove-io PCIE1 primary ------------------------------------------------------------------------------ Notice: The primary domain is in the process of a delayed reconfiguration. Any changes made to the primary domain will only take effect after it reboots. ------------------------------------------------------------------------------ .... # ldm remove-io PCIE15 primary ------------------------------------------------------------------------------ Notice: The primary domain is in the process of a delayed reconfiguration. Any changes made to the primary domain will only take effect after it reboots. ------------------------------------------------------------------------------ |
- Reduce the number of CPU cores and the size of memory assigned to the control domain by specifying a size smaller than the original size with the ldm set-core and ldm set-memory commands.The following gives an example of setting the number of CPU cores of the control domain to 32 and the memory size to 56 GB, according to the configuration example.
| # ldm set-core 32 primary ------------------------------------------------------------------------------ Notice: The primary domain is in the process of a delayed reconfiguration. Any changes made to the primary domain will only take effect after it reboots. ------------------------------------------------------------------------------ # ldm set-memory 56G primary ------------------------------------------------------------------------------ Notice: The primary domain is in the process of a delayed reconfiguration. Any changes made to the primary domain will only take effect after it reboots. ------------------------------------------------------------------------------ |
| Note - We recommend that you first configure the CPU cores with the ldm set-core command and then the memory with the ldm set-memory command. This facilitates the assignment of a continuous area of collective memory. This also makes it easy to easily satisfy the placement conditions for the CPU cores for which physical partition dynamic reconfiguration is possible and the memory placement conditions. For details on the CPU core and memory placement conditions, see "Placement of CPU cores and memory" in "2.5.2 Considerations in System Operation for Dynamic Reconfiguration." |
- Create a service of the virtual console terminal concentrator, called vcc0, with the ldm add-vconscon command and then start the daemon of the virtual network terminal server (vntsd) with the svcadm command. Configure a console connection to each logical domain via this vcc0.
| # ldm add-vconscon port-range=5000-5200 vcc0 primary ------------------------------------------------------------------------------ Notice: The primary domain is in the process of a delayed reconfiguration. Any changes made to the primary domain will only take effect after it reboots. ------------------------------------------------------------------------------ # svcadm enable vntsd |
- Save the configuration information and then restart Oracle Solaris.The following example checks the configuration information saved with the ldm list-spconfig command, and then saves the configuration as name ldm-set1 with the ldm add-spconfig command. Then, it again checks that the configuration has been saved with the ldm list-spconfig command and finally gives an example of restarting Oracle Solaris.
| # ldm list-spconfig factory-default [current] # ldm add-spconfig ldm-set1 # ldm list-spconfig factory-default ldm-set1 [current] # shutdown -i6 -g0 -y |
- b. Establish a redundant configuration for the system volume of the control domain.Establish a redundant configuration for the system volume of the control domain. This item describes an example of the commands for establishing a mirroring configuration with ZFS. For details, see the following documents presented on the Oracle Corporation web site (https://docs.oracle.com/en/).- For Oracle Solaris 10"How to Create a Mirrored Root Pool (Post Installation)" in the Oracle Solaris ZFS Administration Guide- For Oracle Solaris 11"How to Configure a Mirrored Root Pool (SPARC or x86/VTOC)" in the Oracle Solaris 11.x Administration: ZFS File Systems
- To use other redundant configuration software, see the manual for that software.
- Execute the zpool status command to check the status of the root pool.The following example indicates that c2t50000393E802CCE2d0s0 has been assigned to the default root pool (rpool).
| # zpool status rpool pool: rpool state: ONLINE scan: none requested config: NAME STATE READ WRITE CKSUM rpool ONLINE 0 0 0 c2t50000393E802CCE2d0s0 ONLINE 0 0 0 errors: No known data errors |
- Execute the format command to check which disks can be added.The following example indicates that c3t50000393A803B13Ed0 exists as the other disk.
| # format Searching for disks...done AVAILABLE DISK SELECTIONS: 0. c2t50000393E802CCE2d0 <TOSHIBA-MBF2300RC-3706 cyl 46873 alt 2 hd 20 sec 625> /pci@8000/pci@4/pci@0/pci@0/scsi@0/iport@f/disk@w50000393e802cce2,0 /dev/chassis/FUJITSU-BBEXP.500000e0e06d027f/0123_HDD00/disk 1. c3t50000393A803B13Ed0 <TOSHIBA-MBF2300RC-3706 cyl 46873 alt 2 hd 20 sec 625> /pci@8800/pci@4/pci@0/pci@0/scsi@0/iport@f/disk@w50000393a803b13e,0 /dev/chassis/FUJITSU-BBEXP.500000e0e06d243f/022U_HDD01/disk Specify disk (enter its number): ^C |
- Execute the zpool attach command to add the second disk to the rpool to provide a mirror configuration.The following example adds c3t50000393A803B13Ed0s0 with the zpool attach command and then checks the status of the synchronization processing (resilver) with the zpool status command. By referring to status and action, you can determine that the synchronization processing is in progress. Regularly execute the zpool status command to check the status of the synchronization processing until the processing ends.
| # zpool attach rpool c2t50000393E802CCE2d0s0 c3t50000393A803B13Ed0s0 # zpool status rpool pool: rpool state: DEGRADED status: One or more devices is currently being resilvered. The pool will continue to function in a degraded state. action: Wait for the resilver to complete. Run 'zpool status -v' to see device specific details. scan: resilver in progress since Wed Jan 29 21:35:39 2014 3.93G scanned out of 70.6G at 71.9M/s, 0h15m to go 3.90G resilvered, 5.56% done config: NAME STATE READ WRITE CKSUM rpool DEGRADED 0 0 0 mirror-0 DEGRADED 0 0 0 c2t50000393E802CCE2d0s0 ONLINE 0 0 0 c3t50000393A803B13Ed0s0 DEGRADED 0 0 0 (resilvering) |
- Upon the completion of the synchronization processing, [state] is set to "ONLINE", as shown below.
| # zpool status rpool pool: rpool state: ONLINE scan: resilvered 70.6G in 0h10m with 0 errors on Wed Jan 29 21:45:42 2014 config: NAME STATE READ WRITE CKSUM rpool ONLINE 0 0 0 mirror-0 ONLINE 0 0 0 c2t50000393E802CCE2d0s0 ONLINE 0 0 0 c3t50000393A803B13Ed0s0 ONLINE 0 0 0 errors: No known data errors |
- c. Configure a root domain.This item describes the procedure for configuring a root domain.
- Execute the ldm add-domain command to add a logical domain named root-dom0.
| # ldm add-domain root-dom0 |
- Execute the ldm set-variable command to change the OpenBoot PROM environment variable "auto-boot?", which is designed to automatically boot Oracle Solaris, to "false" (disabled). By default, this setting is "true" (enabled). So, OpenBoot PROM tries to start Oracle Solaris automatically when Oracle Solaris is not installed. Changing this setting to disabled facilitates the work to be performed before Oracle Solaris installation.
| # ldm set-variable auto-boot\?=false root-dom0 |
- First, assign the CPU cores with the ldm set-core command and then assign the memory with the ldm set-memory command.The following example assigns 16 CPU cores with the ldm set-core command and 32 GB of memory with the ldm set-memory command, according to the configuration example.
| # ldm set-core 16 root-dom0 # ldm set-memory 32G root-dom0 |
| Note - We recommend that you first configure the CPU cores with the ldm set-core command and then the memory with the ldm set-memory command. This facilitates the assignment of a continuous area of collective memory. This also makes it easy to easily satisfy the placement conditions for the CPU cores for which physical partition dynamic reconfiguration is possible and the memory placement conditions. For details on the CPU core and memory placement conditions, see "Placement of CPU cores and memory" in "2.5.2 Considerations in System Operation for Dynamic Reconfiguration." |
- Execute the ldm set-vconsole command to assign the virtual console (vcons).The following example executes the ldm set-vconsole command to assign port number 5000 of the service (vcc0) of the virtual console terminal concentrator in the control domain to the virtual console.
| # ldm set-vconsole service=vcc0 port=5000 root-dom0 |
- The following example executes the ldm list-io -l command to display the PCI assignment status.The line where NAME begins with "/BB0" and the [TYPE] column has "PCIE" indicates the PCle endpoint on the SPARC M10-4S0 (BB0). The line where the [DOMAIN] column is empty indicates an unassigned PCIe endpoint and the related root complex is displayed in the [BUS] column.Therefore, you can quickly understand that PCIE1, PCIE2, PCIE3, PCIE5, PCIE6, and PCIE7 are unassigned root complexes on BB0.
| # ldm list-io -l NAME TYPE BUS DOMAIN STATUS ---- ---- --- ------ ------ (Omitted) /BB0/CMUL/NET0 PCIE PCIE0 primary OCC [pci@8000/pci@4/pci@0/pci@9] network@0 network@0,1 /BB0/CMUL/SASHBA PCIE PCIE0 primary OCC [pci@8000/pci@4/pci@0/pci@0] scsi@0/iport@f/disk@w50000393e802cce2,0 scsi@0/iport@f/disk@w50000393d8285226,0 scsi@0/iport@f/smp@w500000e0e06d027f scsi@0/iport@f/enclosure@w500000e0e06d027d,0 scsi@0/iport@v0 /BB0/PCI0 PCIE PCIE1 EMP [pci@8100/pci@4/pci@0/pci@0] /BB0/PCI3 PCIE PCIE2 EMP [pci@8200/pci@4/pci@0/pci@0] /BB0/PCI4 PCIE PCIE2 EMP [pci@8200/pci@4/pci@0/pci@8] /BB0/PCI7 PCIE PCIE3 EMP [pci@8300/pci@4/pci@0/pci@0] /BB0/PCI8 PCIE PCIE3 EMP [pci@8300/pci@4/pci@0/pci@8] /BB0/CMUL/NET2 PCIE PCIE4 primary OCC [pci@8400/pci@4/pci@0/pci@a] network@0 network@0,1 /BB0/PCI1 PCIE PCIE5 EMP [pci@8500/pci@4/pci@0/pci@8] /BB0/PCI2 PCIE PCIE5 EMP [pci@8500/pci@4/pci@0/pci@9] /BB0/PCI5 PCIE PCIE6 EMP [pci@8600/pci@4/pci@0/pci@9] /BB0/PCI6 PCIE PCIE6 EMP [pci@8600/pci@4/pci@0/pci@11] /BB0/PCI9 PCIE PCIE7 EMP [pci@8700/pci@4/pci@0/pci@9] /BB0/PCI10 PCIE PCIE7 EMP [pci@8700/pci@4/pci@0/pci@11] (Omitted) |
- See the device path (string displayed as [pci@....]) displayed in the above result and "A.6 SPARC M10-4S Device Paths" in the Fujitsu SPARC M12 and Fujitsu M10/SPARC M10 System Operation and Administration Guide to determine which root complexes are to be assigned to the root domain.
- In the configuration example, all the unassigned root complexes (PCIE1, PCIE2, PCIE3, PCIE5, PCIE6, and PCIE7) on BB0 checked above are to be assigned. So, execute the ldm add-io command to assign them to root-dom0.The following example shows command execution.
| # ldm add-io PCIE1 root-dom0 # ldm add-io PCIE2 root-dom0 # ldm add-io PCIE3 root-dom0 # ldm add-io PCIE5 root-dom0 # ldm add-io PCIE6 root-dom0 # ldm add-io PCIE7 root-dom0 |
- Place the root domain in the bound status with the ldm bind-domain command, and then execute the ldm list-io command to check that the root complexes have been assigned.The following example checks that root-dom0 is bound with the ldm bind-domain command to check with the ldm list-io command that the root complexes have been assigned.The line where the [TYPE] column is "BUS" and the [DOMAIN] column is "root-dom0" indicates the root complex assigned to root-dom0. BUS in that line is the name of the assigned root complex.In the following example, you can check that PCIE1, PCIE2, PCIE3, PCIE5, PCIE6, and PCIE7 have been assigned to root-dom0.
| # ldm bind-domain root-dom0 # ldm list-io NAME TYPE BUS DOMAIN STATUS ---- ---- --- ------ ------ PCIE0 BUS PCIE0 primary IOV PCIE1 BUS PCIE1 root-dom0 IOV PCIE2 BUS PCIE2 root-dom0 IOV PCIE3 BUS PCIE3 root-dom0 IOV PCIE4 BUS PCIE4 primary IOV PCIE5 BUS PCIE5 root-dom0 IOV PCIE6 BUS PCIE6 root-dom0 IOV PCIE7 BUS PCIE7 root-dom0 IOV PCIE8 BUS PCIE8 primary IOV (Omitted) |
- According to step 11. c, configure the root domains (root-dom1 in the example) of the other SPARC M10-4Ss.
- d. Install Oracle Solaris in the root domain.This item describes the procedure for installing Oracle Solaris in the root domain.
- Execute the ldm start-domain command to start root domain root-dom0.
| # ldm start-domain root-dom0 LDom root-dom0 started |
- Execute the ldm set-variable command to change the OpenBoot PROM environment variable "auto-boot?", which is designed to automatically boot Oracle Solaris, to "true" (enabled). After Oracle Solaris is installed, execution of the ldm start-domain command also starts Oracle Solaris.
| # ldm set-variable auto-boot?=true root-dom0 |
- Execute the telnet command to connect to the console of the root domain.The following example checks that the port number of root-dom0 is 5000 by executing the ldm list-domain command. It can also check that root-dom0 is stopped in the OpenBoot PROM (OBP) status by connecting to localhost port number 5000 with the telnet command.
| # ldm list-domain NAME STATE FLAGS CONS VCPU MEMORY UTIL UPTIME primary active -n-cv- UART 64 56G 0.0% 7h 7m root-dom0 active -t---- 5000 32 32G 0.0% 20s root-dom1 bound ------ 5001 32 32G # telnet localhost 5000 .... {0} ok |
- The following example executes the command to start Oracle Solaris 11 installation through the network.
| {0} ok boot net:dhcp .... |
- For details on the installation, see the following documents, presented on the Oracle Corporation web site (https://docs.oracle.com/en/).- Oracle Solaris 10Oracle Solaris 10 1/13 Installation Guide- Oracle Solaris 11Installing Oracle Solaris 11.x Systems
- According to step 11. d, install the other root domains (root-dom1 in the example) in the same way.
- e. Assign the virtual I/O service to the root domain.This item describes an example of the procedure for setting the virtual I/O service to the root domain.The example shows that the entire physical disk of the root domain is lent as the virtual disk service (vds).This example shows that the Ethernet card port is assigned to the virtual switch network device (vsw).For details on the virtual I/O service, see the Oracle VM Server for SPARC Administration Guide published by Oracle Corporation.
- Log in to the root domain with the root privilege.
| root-dom0 console login: root Password: .... |
- To specify the back-end device (physical disk) to be assigned to the virtual disk service (vds), execute the format command to display the disk and its disk path.A string in the format of cXtXXXXXXXXXXXXXXXXdX indicates the disk, and a string beginning with "/pci@" indicates the device path.
| root-dom0# format Searching for disks...done AVAILABLE DISK SELECTIONS: 0. c2t50000393A802CCE2d0 <TOSHIBA-MBF2300RC-3706 cyl 46873 alt 2 hd 20 sec 625> /pci@8100/pci@4/pci@0/pci@0/scsi@0/iport@f/disk@w50000393e802cce2,0 1. c3t50000393D8285226d0 <TOSHIBA-MBF2300RC-3706 cyl 46873 alt 2 hd 20 sec 625> /pci@8200/pci@4/pci@0/pci@0/scsi@0/iport@f/disk@w50000393d8285226,0 2. c4t50000393A804B13Ed0 <TOSHIBA-MBF2300RC-3706 cyl 46873 alt 2 hd 20 sec 625> /pci@3200/pci@4/pci@0/pci@0/scsi@0/iport@f/disk@w50000393a803b13e,0 .... |
- See the device path displayed in the above result and "A.6 SPARC M10-4S Device Paths" in the Fujitsu SPARC M12 and Fujitsu M10/SPARC M10 System Operation and Administration Guide, and check the physical location of the disk. Then, determine the back-end disk (cXtXXXXXXXXXXXXXXXXdX) to be assigned to the virtual disk service.The configuration example assigns two disks (c3t50000393D8285226d0 and c4t50000393A804B13Ed0) as back-ends of the virtual disks of the guest domains (guest0, guest1).
- Next, to specify the network interface to be assigned to the virtual switch service (vsw), execute the dladm show-phys command to display the network interface name (LINK) associated with the location (LOC) of the PCIe slot.From the displayed results, check the location of the network interface and determine the network interface to be assigned to the virtual switch service.The configuration example assigns two network interfaces (net1 and net2) to the virtual switch associated with the virtual network of each guest domain.
| root-dom0# dladm show-phys -L LINK DEVICE LOC net0 igb0 BB#0-PCI#0 net1 igb1 BB#0-PCI#1 net2 igb2 BB#0-PCI#2 net3 igb3 BB#0-PCI#3 ... |
- Execute the ldm add-vdiskserver command to add the virtual disk service to the root domain.The following example adds the virtual disk service (vds0) to the root domain (root-dom0) with the ldm add-vdiskserver command.
| # ldm add-vdiskserver vds0 root-dom0 |
- Execute the ldm add-vdiskserverdevice command to export the back-end of the virtual disk from the root domain.The following example executes the ldm add-vdiskserverdevice command to assign two back-end disks to vds0 to export them.
| # ldm add-vdiskserverdevice /dev/dsk/c3t50000393D8285226d0s2 vol0@vds0 # ldm add-vdiskserverdevice /dev/dsk/c4t50000393A804B13Ed0s2 vol1@vds0 |
- Execute the ldm add-vswitch command to add the virtual switch to the root domain.The following example adds vsw0 and vsw1 to add a virtual switch for each guest domain and assigns a different physical network interface (net1, net2) to each.
| # ldm add-vswitch net-dev=net1 vsw0 root-dom0 # ldm add-vswitch net-dev=net2 vsw1 root-dom0 |
- Similarly, add the virtual I/O service to the root domain (root-dom1) to which the I/O of SPARC M10-4S on the BB1 side was assigned.As an example, assign the following virtual disk service name and virtual switch names to root-dom1. For details on the step, see 11. e.Virtual disk service name: vds1 (assign vol10 and vol11 as two back-end volume names.)Virtual switch name: vsw10, vsw11
- f. Configure a guest domain.This item describes the procedure for configuring a guest domain.
- Execute the ldm add-domain command to add the logical domain named guest0.
| # ldm add-domain guest0 |
- Execute the dm set-variable command to change OpenBoot PROM environment variable "auto-boot?", which is designed to automatically boot Oracle Solaris, to "false" (disabled). By default, this setting is "true" (enabled). So, OpenBoot PROM tries to start Oracle Solaris automatically when Oracle Solaris is not installed. Changing this setting to disabled facilitates the work to be performed before Oracle Solaris installation.
| # ldm set-variable auto-boot\?=false guest0 |
- First, assign the CPU cores with the ldm set-core command and then assign the memory with the ldm set-memory command.The following example assigns 32 CPU cores with the ldm set-core command and 64 GB of memory with the ldm set-memory command, according to the configuration example.
| # ldm set-core 32 root-dom0 # ldm set-memory 64G root-dom0 |
| Note - We recommend that you first configure the CPU cores with the ldm set-core command and then the memory with the ldm set-memory command. This facilitates the assignment of a continuous area of collective memory. This also makes it easy to easily satisfy the placement conditions for the CPU cores for which physical partition dynamic reconfiguration is possible and the memory placement conditions. For details on the CPU core and memory placement conditions, see "Placement of CPU cores and memory" in "2.5.2 Considerations in System Operation for Dynamic Reconfiguration." |
- Execute the ldm set-vconsole command to assign the virtual console (vcons).The following example executes the ldm set-vconsole command to assign port number 5100 of the service (vcc0) of the virtual console terminal concentrator in the control domain to the virtual console.
| # ldm set-vconsole service=vcc0 port=5100 guest0 |
- Execute the ldm add-vdisk command to assign the virtual disk (vdisk).The following example assigns two virtual disks (vdisk0, vdisk10) according to the configuration example. The back-end of each virtual disk specifies the back-ends (vds0 vol0, vds1 vol10) added to two root domains (root-dom0, root-dom1).
| # ldm add-vdisk vdisk0 vol0@vds0 guest0 # ldm add-vdisk vdisk10 vol10@vds1 guest0 |
- Execute the ldm add-vnet command to assign the virtual network device (vnet).The following example assigns two virtual network devices (vnet0, vnet10) according to the configuration example. The virtual switch connected to each virtual network device specifies virtual switches (vsw0, vsw10) added to two root domains (root-dom0, root-dom1).
| # ldm add-vnet vnet0 vsw0 guest0 # ldm add-vnet vnet10 vsw10 guest0 |
| Note - If multiple virtual disks or virtual network devices are assigned, record the value of the ID assigned to the virtual device, from the ldm list -l command execution result. To add a virtual device again after being dynamically deleted, this ID is required so that the virtual device path remains unchanged. |
- g. Install Oracle Solaris in the guest domain.This item describes the procedure for installing Oracle Solaris in the guest domain.
- Execute the ldm bind-domain command to bind the guest domain and then start it by executing the start-domain command.The following example shows the execution of the commands for starting guest0.
| # ldm bind-domain guest0 # ldm start-domain guest0 LDom guest0 started |
- Execute the ldm set-variable command to change the OpenBoot PROM environment variable "auto-boot?", which is designed to automatically boot Oracle Solaris, to "true" (enabled). After Oracle Solaris is installed, execution of the ldm start-domain command also starts Oracle Solaris.
| # ldm set-variable auto-boot\?=true guest0 |
- Execute the telnet command to connect to the console of the guest domain.The following example checks that the port number of the guest0 console is "5100" by executing the ldm list-domain command. It can also check that guest0 is stopped in the OpenBoot PROM status by connecting to localhost port number "5100" with the telnet command.
| # ldm list-domain NAME STATE FLAGS CONS VCPU MEMORY UTIL UPTIME primary active -n-cv- UART 64 56G 0.0% 8h 7m guest0 active -t---- 5100 64 64G 0.0% 20s root-dom0 active -n--v- 5000 32 32G 0.0% 43s root-dom1 active -n--v- 5001 32 32G 0.0% 20s guest1 inactive ------ 64 64G # telnet localhost 5100 .... {0} ok |
| {0} ok boot net:dhcp .... |
- For details on the installation, see the following documents, presented on the Oracle Corporation web site (https://docs.oracle.com/en/).- Oracle Solaris 10Oracle Solaris 10 1/13 Installation Guide- Oracle Solaris 11Installing Oracle Solaris 11.x Systems
- According to step 11. g, install the other guest domains (guest1 in the example) in the same way.
- h. Establish a redundant configuration for the virtual I/Os of a guest domain.The following describes an example of the procedure for establishing a redundant configuration for two virtual network interfaces (vnets) assigned to guest domain guest0, using IPMP. For details on the procedures for other redundant configurations, see the documentation about the software for the respective redundant configurations.
- Log in to guest domain guest0.In the example below, the ldm list-domain command is used to check the port number of the console of guest0, and the telnet command is used to connect to port number "5100."
| # ldm list-domain NAME STATE FLAGS CONS VCPU MEMORY UTIL UPTIME primary active -n-cv- UART 16 14G 0.0% 8h 7m guest0 active -n---- 5100 32 32G 0.0% 20s guest1 active -n---- 5101 32 32G 0.0% 19s root-dom0 active -n--v- 5000 24 24G 0.0% 43s root-dom1 active -n--v- 5001 24 24G 0.0% 20s # telnet localhost 5100 .... guest0 console login: root Password: ... guest0# |
- Execute the dladm command to check that the virtual network devices are visible.In the example below, it is possible to refer to virtual network devices as network interfaces net0 and net1.
| guest0# dladm show-phys LINK MEDIA STATE SPEED DUPLEX DEVICE net0 Ethernet up 0 unknown vnet0 net1 Ethernet up 0 unknown vnet1 |
- This section describes the procedure on Oracle Solaris 11. Execute the ipadm show-if command to check that net0 and net1 are not displayed.
| guest0# ipadm show-if IFNAME CLASS STATE ACTIVE OVER lo0 loopback ok yes -- |
- Execute the ipadm create-ip command to create IP interfaces net0 and net1, and then use the ipadm show-if command to check that they have been created normally.
| guest0# ipadm create-ip net0 guest0# ipadm create-ip net1 guest0# ipadm show-if IFNAME CLASS STATE ACTIVE OVER lo0 loopback ok yes -- net0 ip down no -- net1 ip down no -- |
- Execute the ipadm create-ipmp command to create IPMP interface ipmp0, and then execute the ipadm add-ipmp command to add IP interfaces net0 and net4 to the IPMP group.
| guest0# ipadm create-ipmp ipmp0 guest0# ipadm add-ipmp -i net0 -i net1 ipmp0 |
- Execute the ipadm create-addr command to assign an IP address to IPMP interface ipmp0, and then use the ipadm show-addr command to check the setting. In the example below, a fixed IP address is assigned.
| guest0# ipadm create-addr -T static -a local=xx.xx.xx.xx/24 ipmp0/v4 guest0# ipadm show-addr ADDROBJ TYPE STATE ADDR lo0/v4 static ok 127.0.0.1/8 ipmp0/v4 static ok xx.xx.xx.xx/24 lo0/v6 static ok ::1/128 |
- Execute the ipadm set-ifprop command to set a standby interface, and use the ipmpstat -i command to check the IPMP configuration.
| guest0# ipadm set-ifprop -p standby=on -m ip net1 guest0# ipmpstat -i INTERFACE ACTIVE GROUP FLAGS LINK PROBE STATE net1 no ipmp0 is----- up disabled ok net0 yes ipmp0 --mbM-- up disabled ok |
- Perform the same procedure for the other guest domains (guest1 in the example).
- i. Check the memory resource allocation status.This item describes the procedure for checking and adjusting the location of the memory assigned to each logical domain.
- To enable the SPARC M10-4S to be disconnected using the physical partition dynamic reconfiguration, the memory placement conditions described in "Placement of CPU cores and memory" in "2.5.2 Considerations in System Operation for Dynamic Reconfiguration" need to be met.On the other hand, the configuration example shows a configuration in which no empty resources are allocated. These empty resources are used in advance as movement destinations for the memory block. Therefore, it is important to adjust the memory block placement in advance according to the procedure described in "A.2.3 Example of the Active Replacement Procedure (for Oracle VM Server for SPARC 3.1.x)" to make it easy to take memory placement into consideration.The following describes the example procedure.
- Execute the ldm list-devices -a command to check the status of the continuous region (memory block) of the memory assigned to each logical domain. Then, check that there are not a large number of small memory blocks.As a guide, check that there is not a large number of memory blocks with a SIZE of about 256 MB to 512 GB.The following example shows the execution of the ldm list-devices -a memory command.
| # ldm list-devices -a memory MEMORY PA SIZE BOUND 0x700000000000 4G root-dom0 0x700100000000 4G root-dom1 0x700200000000 8G root-dom0 0x700400000000 4G root-dom1 0x700500000000 4G root-dom0 0x700600000000 8G root-dom1 0x720000000000 8G root-dom0 0x720200000000 8G root-dom0 0x720400000000 8G root-dom1 0x720600000000 8G root-dom1 0x740000000000 32G guest0 0x760000800000 1272M _sys_ 0x760050000000 31488M guest1 0x780000000000 16G guest0 0x780400000000 16G guest1 0x7a0000000000 16G guest0 0x7a0400000000 16G guest1 0x7c0000000000 28G primary 0x7c0700000000 256M guest1 0x7c0710000000 3840M 0x7e0000800000 1272M _sys_ 0x7e0050000000 512M _sys_ 0x7e0070000000 256M _sys_ 0x7e0080000000 8G primary 0x7e0280000000 20G primary 0x7e0780000000 2G guest1 |
- In the above example, the memory blocks of root-dom0 and root-dom1 are broken up into relatively small blocks. So, reassign the memory blocks as described in step 11. j.
- As shown below, if the memory blocks of each logical domain is not broken up, you need not perform reconfiguration. Go to step 11. k.
| # ldm list-devices -a memory MEMORY PA SIZE BOUND 0x700000000000 32G root-dom0 0x720000000000 32G root-dom1 0x740000000000 32G guest0 0x760000800000 1272M _sys_ 0x760050000000 2G primary 0x7600d0000000 29440M guest1 0x780000000000 32G guest0 0x7a0000000000 32G guest1 0x7c0000000000 32G primary 0x7e0000800000 1272M _sys_ 0x7e0050000000 512M _sys_ 0x7e0070000000 256M _sys_ 0x7e0080000000 3328M guest1 0x7e0150000000 4864M 0x7e0280000000 22G primary |
- j. Adjust the memory block placement of the logical domains.As a result of the checks performed in step 11. i, if the memory block of a logical domain is broken up, reallocate the resources assigned to the logical domains. The following is an example of the procedure.
- Stop each logical domain with the ldm stop-domain command and make it inactive with the ldm unbind-domain command.
| Note - Suppose that there are dependencies between the physical I/O assignment (root complex, PCIe endpoint, SR-IOV Virtual Function) or that there are virtual I/O dependencies. Stop and unbind the guest domain to which virtual I/O was assigned, the I/O domain to which the PCIe endpoint and SR-IOV VF were assigned, and the root domain, in that order. |
- The following example shows the commands for making guest0, guest1, root-dom0, and root-dom1 inactive.
| # ldm stop-domain guest0 LDom guest0 stopped # ldm stop-domain guest1 LDom guest1 stopped # ldm stop-domain root-dom0 LDom root-dom0 stopped # ldm stop-domain root-dom1 LDom root-dom1 stopped # ldm unbind-domain guest0 # ldm unbind-domain guest1 # ldm unbind-domain root-dom0 # ldm unbind-domain root-dom1 |
- Using the ldm list-domain command, check that they are inactive.The following example shows that guest0, guest1, root-dom0, and root-dom1 have been made inactive.
| # ldm list-domain NAME STATE FLAGS CONS VCPU MEMORY UTIL UPTIME primary active -n-cv- UART 64 56G 0.2% 5h 28m guest0 inactive ------ 64 64G guest1 inactive ------ 64 64G root-dom0 inactive ------ 32 32G root-dom1 inactive ------ 32 32G |
- Using the ldm start-reconf command, switch from the control domain to delay reconfiguration mode.
| # ldm start-reconf primary Initiating a delayed reconfiguration operation on the primary domain. All configuration changes for other domains are disabled until the primary domain reboots, at which time the new configuration for the primary domain will also take effect. |
- Again set the same values as the number of cores and memory size already assigned, in the order in which the ldm set-core and ldm set-memory commands are used. Then, restart Oracle Solaris of the control domain.The following example again sets the number of CPU cores to 32 with the ldm set-core command and again sets the memory size to 56 GB with the ldm set-memory command to restart Oracle Solaris. This is done with the shutdown command.
| # ldm set-core 32 primary ------------------------------------------------------------------------------ Notice: The primary domain is in the process of a delayed reconfiguration. Any changes made to the primary domain will only take effect after it reboots. ------------------------------------------------------------------------------ # ldm set-memory 56G primary ------------------------------------------------------------------------------ Notice: The primary domain is in the process of a delayed reconfiguration. Any changes made to the primary domain will only take effect after it reboots. ------------------------------------------------------------------------------ # shutdown -i6 -g0 -y |
- Bind and start each logical domain by executing the ldm bind-domain and ldm start-domain commands.Bind and start are performed in the reverse order to that applied for a stop.
- The following example shows that root-dom0, root-dom1, guest0, and guest1 are started (in that order) by executing the ldm bind-domain and ldm start-domain commands.
| # ldm bind-domain root-dom0 # ldm bind-domain root-dom1 # ldm bind-domain guest0 # ldm bind-domain guest1 # ldm start-domain root-dom0 LDom root-dom0 started # ldm start-domain root-dom1 LDom root-dom1 started # ldm start-domain guest0 LDom guest0 started # ldm start-domain guest1 LDom guest1 started |
- Execute the ldm list-domain command to check that all the logical domains are active.
| # ldm list-domain NAME STATE FLAGS CONS VCPU MEMORY UTIL UPTIME primary active -n-cv- UART 64 56G 0.4% 5h 55m guest0 active -n---- 5100 64 64G 1.7% 12s guest1 active -n---- 5101 64 64G 10% 9s root-dom0 active -n--v- 5000 32 32G 3.6% 15s root-dom1 active -n--v- 5001 32 32G 2.2% 11s |
- Execute the ldm list-devices command to check the memory assignment status.
| # ldm list-devices -a memory MEMORY PA SIZE BOUND 0x700000000000 32G root-dom0 0x720000000000 32G root-dom1 0x740000000000 32G guest0 0x760000800000 1272M _sys_ 0x760050000000 31488M guest1 0x780000000000 32G guest0 0x7a0000000000 32G guest1 0x7c0000000000 28G primary 0x7c0700000000 4G 0x7e0000800000 1272M _sys_ 0x7e0050000000 512M _sys_ 0x7e0070000000 256M _sys_ 0x7e0080000000 28G primary 0x7e0780000000 1280M guest1 0x7e07d0000000 768M |
- Save the configured local domain configuration information in the XSCF.
Execute the ldm set-spconfig command to save the configured information.
The following example checks the saved configuration information and then saves it with the same name as that of the existing configuration information.
- Execute the ldm list-spconfig command to check the current configuration information.
| # ldm list-spconfig factory-default ldm-set1 [next poweron] |
- Execute the ldm remove-spconfig command to delete the configuration information to be overwritten.
| # ldm remove-spconfig ldm-set1 |
- Execute the ldm add-spconfig command to re-save the configuration information.
| # ldm add-spconfig ldm-set1 |
- Execute the ldm list-spconfig command to check that the saved configuration information has become [current].
| # ldm list-spconfig factory-default ldm-set1 [current] |
- Save the configured logical domain configuration information to an XML file.
To guard against the configuration information saved to the XSCF being unusable, save the configuration information to an XML file. It is recommended that the XML file be saved to a different media.
The following describes the example procedure.
- Execute the ldm list-domain command to check that all the logical domains are active.
| # ldm list-domain NAME STATE FLAGS CONS VCPU MEMORY UTIL UPTIME primary active -n-cv- UART 64 56G 0.0% 6h 9m guest0 active -n---- 5100 64 64G 0.0% 15m guest1 active -n---- 5101 64 64G 0.0% 15m root-dom0 active -n--v- 5000 32 32G 0.0% 15m root-dom1 active -n--v- 5001 32 32G 0.0% 15m |
- Execute the ldm list-constraints command to save the configuration information to an XML file.
| # ldm list-constraints -x > /ldm-set1.xml |
< Previous Page | Next Page >