LaoTsao's Weblog (老曹的網路記 ): 2011

Wednesday, December 14, 2011

Changes in This Release : Solaris 11

The following enhancements are included in the Oracle Solaris 11 release:

The configuration parameters that used to be set by editing the/etc/default/autofs and /etc/default/nfs can now be set in the SMF repository.
The NFS service provides support for mirror mounts. Mirror mounts enable an NFSv4 client to traverse shared file system mount points in the server namespace. For NFSv4 mounts, the automounter will perform a mount of the server namespace root and rely on mirror mounts to access its file systems. The main advantage that mirror mounts offer over the traditional automounter is that mounting a file system using mirror mounts does not require the overhead associated with administering automount maps. Mirror mounts provide these features:
- Namespace changes are immediately visible to all clients.
- New shared file systems are discovered instantly and mounted automatically.
- File systems unmount automatically after a designated inactivity period.
NFS referrals have been added to the NFS service. Referrals are server-based redirections that an NFSv4 client can follow to find a file system. The NFS server supports referrals created by the nfsref(1M)command, and the NFSv4 client will follow them to mount the file system from the actual location. This facility can be used to replace many uses of the automounter, with creation of referrals replacing the editing of automounter map. NFS referrals provide these features:
- All of the features of mirror mounts listed above
- Automounter-like functionality without any dependence on the automounter.
- No setup required at either the client or server.
The ability to mount the per-DNS-domain root of a Federated File System name space has been added. This mount point can be used with NFS referrals to bridge from one file server to another, building an arbitrarily large namespace.
The sharectl utility is included. This utility enables you to configure and manage file sharing protocols, such as NFS. For example, this utility allows you to set client and server operational properties, display property values for a specific protocol, and obtain the status of a protocol.
The NFS version 4 domain can be defined.

Significant Changes in Earlier Releases : Solaris 10

The Solaris 10 11/06 release provides support for a file system monitoring tool. See the following:

fsstat Command for a description and examples
fsstat(1M) man page for more information

Additionally, this guide provides a more detailed description of the nfsmapiddaemon. For information about nfsmapid, see the following:

nfsmapid Daemon
nfsmapid(1M) man page

Starting in the Solaris 10 release, NFS version 4 is the default. For information about features in NFS version 4 and other changes, refer to the following:

Additionally, the NFS service is managed by the Service Management Facility.

Administrative actions on this service, such as enabling, disabling, or restarting, can be performed by using the svcadm command.
The service's status can be queried by using the svcs command.

solaris10 zone to solaris10 brand in s11 recovery

there is interesting thread on zones-discuss the situation

start with s10u8 zone
zoneadm attach -u to s10u9 hosts
v2v to s11 express as solaris10 brand zone
update the host to solaris 11 from solaris 11 express
run /usr/lib/brand/shared/dsconvert
# zoneadm -z sandpit boot (failed)
zone 'sandpit': WARNING: vnic3:1: no matching subnet found in netmasks(4): 172.25.48.101; using default of 255.255.0.0.
zone 'sandpit': Error: The installed version of Solaris 10 is not supported.
zone 'sandpit': SPARC systems require patch 142909-17
zone 'sandpit': x86/x64 systems require patch 142910-17
zone 'sandpit': exec /usr/lib/brand/solaris10/s10_boot sandpit /zoneRoot/sandpit failed
zone 'sandpit': ERROR: unable to unmount /zoneRoot/sandpit/root.
to recover

1. Reboot to the Solaris 11 Express BE

   root@global# beadm activate <s11express-be-name>
   root@global# init 6

2. Partially revert the work done by dsconvert

   In this example, the zone's zonepath is /zones/s10.

   root@global# zfs list -r /zones/s10
   rpool/zones/s10                    3.18G  11.3G    51K  /zones/s10
   rpool/zones/s10/rpool              3.18G  11.3G    31K  /rpool
   rpool/zones/s10/rpool/ROOT         3.18G  11.3G    31K  legacy
   rpool/zones/s10/rpool/ROOT/zbe-0   3.18G  11.3G  3.18G  /
   rpool/zones/s10/rpool/export         62K  11.3G    31K  /export
   rpool/zones/s10/rpool/export/home    31K  11.3G    31K  /export/home

   The goal here is to move rpool/zones/s10/rpool/ROOT up one level.  We
   need to do a bit of a dance to get it there.  Do not reboot or issue
   'zfs mount -a' in the middle of this.  If something goes wrong and a
   reboot happens, it won't be disasterous - you will just need to
   complete the procedure when the next boot stops with
   svc:/filesystem/local problems.

   root@global# zfs set mountpoint=legacy rpool/zones/s10/rpool/ROOT/zbe-0
   root@global# zfs set zoned=off rpool/zones/s10/rpool
   root@global# zfs rename rpool/zones/s10/rpool/ROOT/zbe-0 \
      rpool/zones/s10/ROOT
   root@global# zfs set zoned=on rpool/zones/s10/rpool
   root@global# zfs set zoned=on rpool/zones/s10/ROOT

   Now the zone's dataset layout should look like:

   root@global# zfs list -r /zones/s10
   NAME                                USED  AVAIL  REFER  MOUNTPOINT
   rpool/zones/s10                    3.19G  11.3G    51K  /zones/s10
   rpool/zones/s10/ROOT               3.19G  11.3G    31K  legacy
   rpool/zones/s10/ROOT/zbe-0         3.19G  11.3G  3.19G  legacy
   rpool/zones/s10/rpool                93K  11.3G    31K  /rpool
   rpool/zones/s10/rpool/export         62K  11.3G    31K  /export
   rpool/zones/s10/rpool/export/home    31K  11.3G    31K  /export/home

3. Boot the zone and patch

   root@global# zoneadm -z s10 boot
   root@global# zlogin s10
   root@s10# ...  (apply required patches)



119254/119255 rev 75 (patch utils)
u9 kernale patch 14299/142910-17(SPARC/x86) 






4. Shutdown the zone

   root@s10# init 0

5. Revert the dataset layout to the way that dsconvert left it.

   Again, try to avoid reboots during this step.

   root@global# zfs set zoned=off rpool/zones/s10/ROOT
   root@global# zfs set zoned=off rpool/zones/s10/rpool
   root@global# zfs rename rpool/zones/s10/ROOT rpool/zones/s10/rpool/ROOT
   root@global# zfs set zoned=on rpool/zones/s10/rpool
   root@global# zfs inherit zoned rpool/zones/s10/rpool/ROOT

6. Reboot to Solaris 11

   root@global# beadm activate <solaris11-be-name>
   root@global# init 6

At this point, the zone should be bootable on Solaris 11.

Observations


since this is zoneadm attach -u from u8 to u9, only min update so it is not really a u9 zone
one should really do zoneadm attach -U from u8 to u9
RTFM,  to support SVR4 pkg and patching one need to install 119254-75 (SPARC), 119534-24, and 140914-02 or 119255-75, 119535-24 and 140915-02 (x86/x64), or later versions in solaris 10 before created archive

s10u9 introduce zoneadm attach -U option in addition to the -u option
s10u8 with patch 142910-17/142909-17 ( the update 9 Kernel patch, SPARC/x86 ), once that is installed -U is available, install latest rev of 119254/11955 first, if patching though, it's the patchutils patch
man page

attach [-u | -U] [-b patchid]... [-F] [-n path] [brand-specific options]

The attach subcommand takes a zone that has been detached from one system and attaches the zone onto a new system. Therefore, it is advised (though not required) that the detach subcommand should be run before the “attach” takes place. Once you have the new zone in the configured state, use the attach subcommand to set up the zone root instead of installing the zone as a new zone.

For native zones, zoneadm checks package and patch levels on the machine to which the zone is to be attached. If the packages/patches that the zone depends on from the global zone are different (have different revision numbers) from the dependent packages/patches on the source machine, zoneadm reports these conflicts and does not perform the attach. If the destination system has only newer dependent packages/patches (higher revision numbers) than those on the source system, you can use the -u or -U options. The -u option updates the minimum number of packages within the attached zone to match the higher-revision packages and patches that exist on the new system. The -U option updates all packages in the attached zone that are also installed in the global zone. With -u or -U, as in the default behavior, zoneadm does not perform an attach if outdated packages/patches are found on the target system.

For native zones, one or more -b options can be used to specify a patch ID for a patch installed in the zone. These patches will be backed out before the zone is attached or, if -u was also specified, updated.

The -F option can be used to force the zone into the “installed” state with no validation. This option should be used with care since it can leave the zone in an unsupportable state if it was moved from a source system to a target system that is unable to properly host the zone. The -n option can be used to perform a “dry run” of the attach subcommand. It uses the output of the “detach -n” subcommand as input and is useful to identify any conflicting issues, such as the network device being incompatible, and can also determine whether the host is capable of supporting the zone. The path can be “-”, to read the input from standard input.

It seems that to get to u9 from u8 with all the patches and update one should use -U option
one should not use -F at all
In any case always have a backup copy of the zone so one can always restore the zone to its original state

Wednesday, December 7, 2011

what's new in solaris 11

this link list the what's new

installation

Automated Installer

installation framework for automated system provisioning
network installation
manifest

system configuration
SW pkg
zone

bootable image

Jumpstart migration utility js2ai
interactive Text installations

server configuration
automatic or manual network configuration
no GUI desktop
audio or wireless drivers

Live Media Installation (x86)

automatic network configuration
full GUI desktop
GNU partition Edition

Distribution Constructor

CML tool for building pre-configured bootable customized s11 installation image
use manifest description

target disk
SW pkg
basic system configuration
gold image

packaging

Image Packing System (IPS0

framework for complete SW lifecycle mgmt

installation
upgrade
remove

integrated with ZFS

safe upgrade with ZFS clone FS

network based package repositores

with full automatic dependency checking

any SW that is required is sutomatically installed or update

boot to different boot env
can lock down individua pkg
fast boot feature

on by default in x86
off by default in SPARC

support SVR4 pkg

no legacy patching tool

System configuration

SMF
Name service

nscfg
/etc/nsswitch.conf svc:/system/name-service/switch
/etc/resolv.conf svc:/network/dns/client
/etc/nodename svc:/idenitty:node
/etc/defaultdomain svc:/system/identity:domain
/etc/default/init svc:/system/environment:init
/etc/driver/drv/driver.conf

sysconfig

replace sys-unconfig, sysidtool
unconfiguring
reconfiuring

SMF, FMA

SNMP trap
SMTP notification
ASR

zones are easier to create and manage
solaris10 zone

p2v
v2v
zonep2vchk
NFS server in zone
exclusive-IP zones by default

anet for exclusive-IP zone

administer network flow within NGZ

bandwidth
priority control based on IP address, subnet tramsport protocols and port
flowadm
flowstat

Delegated Administration

admin zone based on RBAC

zone boot env

ZFS boot env: ZBE
beadm inside zone

improved zones dataset layout

NGZ mimic GZ
NGZ support different ZFS dataset

immutable zones

read-only root for zones
Mandatory Write Access Control (MWAC

cleanly shutdown zones

zoneadm -z <z> shutdown

zonestat

observation of system resources
memory, CPU, resource control limit
exclusive-IP: network device utilization on data-links, vlink and zones
libzonestat

svc:/system/zones-monitoring.default

tecla CLi editing library for zonecfg

emacs mode :default
vi mode
tecla(5)

Security

Role Authentication

root is a role by default
1st user account is assigned root role
user assume root role

user or role passwd

Trusted Platform Module (TPM)

TPM chip is a HW device on MB
protected storage
protected capabilities on an inexpensive components with restricted resource
s11 provide drivers

TCG 12 spec
TSS SW to provide cryptographic openationd on secre device and adm toll for manageing the YPM and PKCS11 provider

labeled Ipsec

trusted extension

IPsec support AES FMAC Cryptographic Algorithm

data integrity of AES Galoris/Counter Mode (AES GCM) but without acturally encrypting the data

Kerberos Dtrace Providers

RFC4120

Trusted Extensions Enhancements

enables per-label and pe-user credentials to request a unique passwd for each label
tncfg :

create, modify and display networking properties
label network packets received from remote hosts

set security lables on ZFS dataset

Support ssh X.509 Certificate Extension
Solaris Cryptographic Framework

NSA Suite B algorithms
T4 support AES CFG mode used by table space encryption of oracle DB advanced Securiy option
support Intel Advanced Encrytion Stnadards (AES-NI0
Oracle key managemeny system now be used for AES key storage using the new pkcs11 kms plugin

In-kernel pfexec ZForced and Basic Privileges

Nwtrorking

re-architecture to unify, simplify and enhance observation and interoperability of NIC

GLDv3 driver framework

VLAN
link aggregation
MAC layer for Ethernet, Wi-Fi and IB
dladm

Network v12n and resource mgmt

V12N

VNIC
vswitch
VLANs
routing
firewall
tight integration with zone exclusive-ip

Resource Mgmt

bandwidth limits
CPU limit
interrupt-driven to polling

Manual and Automatic Networking

network profile svc:/network/physical:default

switch between automatic and manual networking by enabling Automatic or DefaultFixed profile through netadm and netcfg

Live Media install (LiveCD) use Automatic networking, useful for laptop

Default Names for Datalinks

net0, net1 etc
can be reverted

Changing MAC Address with dladm

persistent across reboots

IB Enabled and Optimized

improved support for Sockets Direct Protocol (SDP)

support RDMA; zero-copy data transfer
netstat, truss, pfiles mdb kmdb
NGZ for exclusive-IP and Shared-IP)

RDSv3 for Oracle RAC

Registration of VLANs

ability for broadcasting VLAN ID
VNIC support

Link Layer Discovery Protocol Support (LLDP)

one-way link layer protocol that allow an IEEE802 LAN station to advertise the capabilities and current status of the system
lldpadm: enable/disable LLDP agent on physical datalink

New Sockets Architecture

no longer use STREAMS
significant performance improvements
simplified developer interface for new socket types

Load Balancing

Integrated L3/L4 LB
stateless DSR and NAT modes
CLI
configuration API

Link Protection

prevent guest VM sending harmful packets to network
basic threats: IP, DHCP, MAC, L2 fram spoofing
use ipf for inbound filtering and customizable filter rules

Bridging and Tunneling

Bridging

Spanning Tree Protocol (STP, IEEE 802.ID-1998)
TRILL protocol

Tunneling

iptun
wireshark
snoop

IP observability

wireshark: packet sniffing tool and snoop
dlstat: runtme statistics for data link

IP Multipathing(IPMP)

re-architecture
ipadm
Transitive probe: new failure detection mode

without aditioning test IP address

svccfg -a svc:/network/ipmp setprop config/transitive-probing=true
svcadm refresh svc:/network/ipmp:default

in.mpathd

managed by SMF service svc:/network/ipmp

I/O Enhancements to netcat
new FTP server

proftpd replace WU-ftpd

Dtrace Networking Provider

tcp
udp
ipv4/IPv6

Storage

ZFS is root FS
easy upgrade with IPS
ZFS data Encrytion
ZFS deduplication :(need RAM, L2ARC with SSD)
ZFS Shadow Migration (local or NFS FS0
ZFS backup with NDMP with ZFS send/receive
Temporary ZFS mountpoint
ZFS snapshot Alias with zfs snap (snapshot)
Recursive ZFS send (dataset and descendents)
ZFS snapshot Diff
NFSv4 Client and Server Migration Support
SMB for Micosoft interoperability
Dtrace Storage Provider

SMB
iscsi

COMSTART SCSI target Frameworks

SCSI device type: disk, tape with FC
iSCSI Extensions for RDMA (iSER)
SCSI RDMA Protocol (SRP) for IB HCA
iSCSI
Fibre Channel over Ethernet (FCoE)
Dtrace Provider:

SCSI Target Mode Framework (STMF)
SCSI Block Device (SBD)

Kernel/Platform Support

SPARC T4

2GB page size
ISA cryptographic HW optimization
CPU and DRAM performance counter support
L3 cache support
20%-40% gain for various ciper and hash instruction
gain for SSL and direct cryptographic acceleration for DB 11.2.0.2
Critical Threads

dynamic allocation of HW resource to provide boots in performance
matching a thread's HW requirements with the amount of exclusive access to specific HW resources

Single-root I/O v12n (SR-IOV)

extension to PCIe to allow efficient sharing of PCIe devices among VMs both in HW and SW

NUMA I/O

allow kernel threads, interrupts and memory to be placed on physical resources according to the physical topology pf the machines
specific high-level affinity requirements of I/O frameworks, actual load, resource control and power mgmt policies

Intel Advanced Vector Extensions(AVX)

new instructions vector floating point operations

image, video, audio processing, 3D modeling, scientific simulation and financial analytics

Sandy Bridge and beyond

Dynamic Intimate Shared Memory (DISM) performance Improvements

for large memory system 8x oracle DB start up improvement for ISM and DISM creation, locking, destruction

Suspend and resume to RAM
Improved HW supported

FMA
generic topology enumeration
generic hotplug framework
latest Intel microprocessor
Intel's Latency TOP and Dtrace to measured latency

Dtrace cpc Provider

cycles executed
instructions executed
cache missed
TLB misses

user Environment

850 open source pkg in IPS

Java SE 6, 7
GCC 4.5.2
Python 2.7
Perl 5.1.2
Ruby 1.8.7
PHP 5.2.17
complete web stack

Desktop env

GNOME 2.30.3
Firefox 6
Thunderbird 6

in /usr/bin
in /usr/gnu/bin

Default shell:

user: bash
system: ksh93

Removable Media

HAL
D-Bus messaging passing system

new sound system
search for content in MAN pages

man -K searchstring

Virtual Console Terminals

svc:/system/vtdaemon:defaul
svc:/system/console-login:vt*
Alt-Ctrl-F#

Time Slider Snapshot Mgmt

use home
Gui

Common UNIX Printing System (CUPS) printing

Lp wrap CUPS functionality

libc Familiarity

improve familiarity with linux and BSD

paths.h Path Name Definitions

/usr/include/paths.h
/usr/include/sys/paths.h

locale and languages (200+)
TrueType Fonts

solaris branded zone in solaris 11

Solaris branded zone is the default zone in Solaris 11

whole-root type only
immutable (read-only zone root) zone with file-mac-profile (mandatory acccess control)

none: standard read-write
strict: read-only FS, no exceptions, only logged remotely
fixed-configuration: permits updates to /var except systome configuration
flexible-configuration:permit change

/etc
root home directory
/var

zonecfg add dataset: read-only dataset
zonecfg add fs, can mount read-only FS

IPS packing
install, detach, attach and P2V
NGZ root is a ZFS dataset
use boot env: beadm
All enabled IPS pkg repositories must be accessible while installing a zone
zone SW is minimized
default exclusive-IP with Automatic NET (anet) VNIC
support

ZFS encryption
Network V12n and QoS
SMB and NFS
can be NFS server

not supported

DHCP address assignment in a shared-IP zone
ndmpd
SMB server
SSL proxy server
ZFS pool administration through zpool cmd

zonestat: report CPU, memeory resource control, network bandwidth for exclusive-ip zone
admin resource

user
auths

solari.zone.login
solaris.zone.manage
solaris.zone.clonefrom

resources pool association

dedicated-cpu

ncpus
importance

capped-cpu
capped-memory

physical
swap
locked

zone network interface

shared-IP

shared a network interface with GZ
use ipadm
net resource properties

address
physical

exclusive-IP

must have dedicated network interface
anet resource, a dedicated VNIC is automatically created and assigned to zone
can use pre-configured VNIC
default
support

DHCP v4 and v6
IP filter
IPMP
IP routing
ipadm for setting TCP/UDP/SCTP and IP/ARP
IPsec and IKE
snoop
dtadm
sysconfig

hostid
disk format: uscsi
devices: /dev in zone
zone-wide resource

zone.cpu-cap
zone.cpu-shares
zone.max-locked-memeory
zone.max-lofi
zone.max-lwps
zne.max-msg-ids
zpne.max-processes
zone.max-sem-ids
zone.mac-shm-ids
zone.max-shm-memory
zone.max-swap

use attr for comment

solaris10 branded zone in solaris 11

Due to change in package system (SRV4 to IPS) there is no direct upgrade from S10 to S11 one can use

P2Vconverting s10 physical system to solaris10 branded zone in s11
V2V converting s10 native full root zone to solaris10 branded zone in s11

The following list some limitations of solaris10 zone in s11 from solaris10(5)

s10u9 or earily with patch 142909-17(SPARC) or 142910-17(x86) or later
32bit and 64-bit solaris 10 apps
zone must reside on its own zfs dataset
delegated ZFS dataset configuration is currently experimental and has not yet been tested
para-v xvm domain is experimental and know problem for 64-bit apps
/dev/sound device can not be conigured
file-mac-profile property used to create read-only zones is not available
quota(1M0 to retrieve UFS FS info is not available
the following ndd parms are not available

ip_squeue_fanout
ip_soft_rings_cnt
ip_ire_pathmtu_interval
tcp_mdt_max_pbufs

Networking features that are different

Mobile IP is not available in s11
/dev/net ;VNIC are not supported by libdlpi in s11 but support by libslpi(3LIB) in s10
IPMP output are not the same

mdb and dtrace are not fully functional when used in global zone to examine processes executing within solaris10 zone
zonep2vchk (can be copied for s11 to s10) is used to generate info needed for P2V
solaris10 zone does not supported statically linked binaries
to support SVR4 pkg and patching one need to install 119254-75 (SPARC), 119534-24, and 140914-02 or 119255-75, 119535-24 and 140915-02 (x86/x64), or later versions in solaris 10 before created archive
in s11 pkg:/system/zones/brand/solaris10 must be installed
zonecfg: use create -t SYSsolaris10 or set brand=solaris10
can set hostid
support sysidcfg
support migration between two s11 hosts

Tuesday, December 6, 2011

OSC4.0 what's new

the release note , we just summary the what's new part here

Support Solaris 11 (for s10 use OSC3.3.1)
IPS package format
Automated Installer Support
run only in a GZ and in a zone cluster. A zone cluster is now configured with the solaris brand NGZ with cluster attr. The solaris and solaris10 brands of NGZ are supported for configuration with the HA for Oracle Solaris Zones data service.
Support for Apache, Apache Tomcat, DHCP, DNS, NFS, Oracle Database 11.2.0.3 (single instance and Oracle RAC) and WebLogic Server.
DR (geo-cluster) support for replication solutions such as StorageTek
Availability Suite 4.0, Oracle Data Guard and a script-based plug-in.
New IPS pkg name for SPARC and x86

Previous Cluster Package Cluster New IPS Package Name
SUNWscapc                          ha-cluster/data-service/apache
SUNWscdhc                          ha-cluster/data-service/dhcp
SUNWscdns                          ha-cluster/data-service/dns
SUNWsczone    ha-cluster/data-service/ha-zones
SUNWscnfs     ha-cluster/data-service/nfs
SUNWscor                            ha-cluster/data-service/oracle-database
SUNWsctomcat ha-cluster/data-service/tomcat
SUNWscwls                         ha-cluster/data-service/weblogic
SUNWscdsbuilder                ha-cluster/developer/agent-builder
SUNWscdev    ha-cluster/developer/api
SUNWscderby                       ha-cluster/ha-service/derby
SUNWscgds                           ha-cluster/ha-service/gds
SUNWscrtlh                           ha-cluster/ha-service/logical-hostname
SUNWscsmf                           ha-cluster/ha-service/smf-proxy
SUNWsctelemetry                  ha-cluster/ha-service/telemetry
SUNWsccacao                       ha-cluster/library/cacao
SUNWscucm                          ha-cluster/library/ucmm
SUNWesc, SUNWfsc, SUNWjsc,
SUNWcsc
                                                  ha-cluster/locale
SUNWscnmr, SUNWscnmu ha-cluster/release/name
SUNWscmasar, SUNWscmasazu,
SUNWscmautil, SUNWscmautilr
                                                 ha-cluster/service/management
SUNWscmasasen                    ha-cluster/service/management/slm
SUNWscqsr, SUNWscqsu       ha-cluster/service/quorum-server
SUNWscqsman                          ha-cluster/service/quorum-server/manual
SUNWjscqsu, SUNWcscqsu     ha-cluster/service/quorum-server/locale
SUNWjscqsman                    ha-cluster/service/quorum-server/manual/locale
SUNWmdmr, SUNWmdmu               ha-cluster/storage/svm-mediator
SUNWscsckr, SUNWscscku             ha-cluster/system/cfgchk
SUNWsc, SUNWscu, SUNWscr,
SUNWsczr, SUNWsczu,
SUNWsccomu, SUNWsccomzu
                                                    ha-cluster/system/core
SUNWscmasa, SUNWscmasau ha-cluster/system/dsconfig-wizard
SUNWscman ha-cluster/system/manual
SUNWscdsman ha-cluster/system/manual/data-services
SUNWjscman ha-cluster/system/manual/locale

(SPARCOnly) SUNWscxvm ha-cluster/data-service/ha-ldom

TABLE 3 NewIPS PackageNames forGeographic Edition
Previous Geographic Edition Package Name New IPS Package Name
SUNWscgctl, SUNWscgctlr,
SUNWscghb, SUNWscghbr
                                                          ha-cluster/geo/framework
SUNWscgrepavs, SUNWscgrepavsu   ha-cluster/geo/replication/availability-suite
SUNWscgrepodg, SUNWscgrepodgu ha-cluster/geo/replication/data-guard
SUNWscgrepsbpu ha-cluster/geo/replication/sbp
SUNWscgman                                        ha-cluster/geo/manual

What's Not Included in the Oracle Solaris Cluster 4.0 Software
The following features were included in the Oracle Solaris Cluster 3.3 release but are not included in the Oracle Solaris Cluster 4.0 release:
■ Support for Veritas File System (VxFS) and Veritas VolumeManager (VxVM)
■ Support for the VxVM cluster feature for Oracle RAC in addition to VxVM with Oracle
Solaris Cluster
■ GUI and GUI wizards
■ Support for SunManagement Center
■ Support for SunQFS from Oracle
■ Support for non-global zones as resource-group node-list targets
■ Support for Oracle Solaris IP Security Architecture (IPsec)
■ Support for Oracle Solaris Trusted Extensions
■ The scsnapshot tool
■ The cconsole utility (the Oracle Solaris pconsole utility can be used instead)
■ Storage-based replication:
■ Support for EMC Symmetrix RemoteData Facility (SRDF)
■ Support for Hitachi True Copy andHitachi Universal Replicator storage-based
replication
■ Three-data-center (3DC) configuration
The following HA data services are not initially available with the 4.0 release but might become available at a later time:
■ Afga IMPAX
■ ASE
■ Informix
■ Kerberos
■ MySQL
■ Oracle Business Intelligence Enterprise Edition
■ Oracle eBusiness Suite
■ Oracle GlassFishMessage Queue
■ Oracle iPlanet Web Server
■ PeopleSoft Enterprise
■ PostgreSQL
■ Samba
■ SAP
■ SAP liveCache
■ SAP Web Application Server
■ Siebel, SWIFTAlliance Access and Gateway
What's Not Included in the Oracle Solaris Cluster 4.0 Software
Oracle Solaris Cluster 4.0 Release Notes 13
■ Sybase
■ TimesTen
■ WebSphereMessage Broker
■ WebSphereMessage Queue
The Grid Engine and Sun Java System Application Server EE (formerly calledHADB) data services have been removed from the Oracle Solaris Cluster software.

oracle DB/RAC 11.2.0.3

Oracle DB/RAc 11.2.0.3
11/10/11: Patch Set 11.2.0.3 for Linux, Solaris, Windows, AIX and HP-UX Itanium is now available on support.oracle.com. Note: it is a full installation (you do not need to download 11.2.0.1 first). See the README for more info (login to My Oracle Support required).

For solaris 11 you need 11.2.0.3

osc4.0 docs and download

dec/4 2011 oracle Solaris 4 is out

download
http://www.oracle.com/technetwork/server-storage/solaris-cluster/downloads/index.html

docs links
Oracle Solaris Cluster

Release Notes:
Oracle Solaris Cluster 4.0 | Oracle Solaris Cluster 3.3
Read about new features, new qualifications, and workarounds for known issues.
Installation:
Oracle Solaris Cluster 4.0 | Oracle Solaris Cluster 3.3
Learn everything you need to know to install the cluster.
Oracle Solaris Cluster Geographic Edition 4.0 | Oracle Solaris Cluster Geographic Edition 3.3
Learn everything you need to know to install the Geographic Edition software.
Administration:
Oracle Solaris Cluster 4.0 | Oracle Solaris Cluster 3.3
Learn how to keep your enterprise and data highly available.
How-To Guides and White Papers:
Access all the Oracle Solaris Cluster How-To Guides and White Papers.

How-To Guides
	How-To Install and Configure a Two-Node Cluster (Oracle Solaris 11) This article provides a step-by-step process for quickly and easily installing and configuring Oracle Solaris Cluster software for two nodes, including the configuration of a quorum device.
	How-To Create a Failover Zone in a Cluster (Oracle Solaris 11) This how-to-guide describes how to quickly and easily configure an Oracle Solaris Zone in failover mode using the Oracle Solaris Cluster High Availability (HA) agent for Oracle Solaris Zones, which supports both Oracle Solaris 10 and 11 Zones.

Other Resources

Oracle Solaris Cluster Features and Benefits

What's New in Oracle Solaris Cluster

Oracle Solaris Cluster Frequently Asked Questions

Oracle Solaris Cluster System Requiremen

full docs link

Observations: there are very few data service docs:

failover zone example in release note
SUNW.nfs agent

Oracle Solaris Cluster 4 Webcast Dec 06 2011 12noon EST webast

Oracle Solaris Cluster 4.0 Launch Webcast - Join the webcast on Tuesday, 12/6/11 at 9am PT.

Register Today and learn about Oracle Solaris Cluster 4.0, the first release providing high availability (HA) and disaster recovery (DR) capabilities for Oracle Solaris 11, the first cloud OS. Bill Nesheim, VP, Oracle Solaris Platform Engineering, will present how Oracle Solaris Cluster extends Oracle Solaris to provide the HA and DR infrastructure required for deploying mission critical workloads , in private, public and hybrid clouds deployments as well as enterprise data centers.

Register Now!

oracle solaris Summit at LISA 2011 (Dec/06/2011)

Dec 06 2011 the live stream of Solaris Day @LISA2011
http://psav.mediasite.com/mediasite/Viewer/?peid=b31d4d7b75fe4fcc8b3798a42d3d6b711d

Agenda
8:00 a.m.	Registration
9:00 a.m.	Oracle Solaris 11 Strategy Markus Flierl, VP Software Development, Oracle
9:30 a.m.	Next Generation OS Lifecycle Management with Oracle Solaris 11 Dave Miner, Principal Software Engineer, Oracle Bart Smaalders, Principal Software Engineer, Oracle
11:00 a.m.	Data Management with ZFS Mark Maybee, Principal Software Engineer, Oracle
12:00 noon	Lunch
1:00 p.m.	Oracle Solaris Virtualization and Oracle Solaris Networking Mike Gerdts, Senior Software Engineer, Oracle Sebastian Roy, Software Engineer, Oracle
2:30 p.m.	Security in your Oracle Solaris Cloud Environment Glen Faden, Sr. Principal Software Engineer, Oracle
3:15 p.m.	Break
3:30 p.m.	Oracle Solaris – The Best Platform to run your Oracle Applications David Brean, Principal Software Engineer, Oracle
4:15 p.m.	Oracle Solaris Cluster – HA in the Cloud Gia-Khanh Nguyen, Principal Software Engineer, Oracle
5:00 p.m.	Networking Reception sponsored by Oracle Solaris Cluster

oracle DB and RAC and v12N/partition and Solaris 11

this link list the Supported V12N and partitioning tech for Oracle DB and RAC
I just want to highlight the solaris part as of 12/06/2011

The Solaris 11 will support oracle 11.2.0.3 and above
Solaris zone will work with oracle clusterware only with DB and RAC

Platform	Virtualization Technology	Operating System	Certified Oracle Single Instance Database Releases	Certified Oracle RAC Database Releases
Oracle Solaris Sparc	Dynamic Domain	Solaris 10	10gR2 11gR1 Note 11gR2	10gR2 11gR1 11gR2
	Dynamic Domain	Solaris 9	10gR2 11gR1	10gR2 11gR1
	Oracle VM Server for SPARC	Solaris 10	10gR2 Note 11gR1 Note 11gR2 Note	10gR2 11gR1 Note 11gR2 Note
	Oracle VM Server for SPARC	Solaris 11	11gR2 (11.2.0.3 and above)	11gR2 (11.2.0.3 and above)
	Oracle Solaris Containers	Solaris 10	10gR1 Note 10gR2 Note 11gR1 Note 11gR2 Note	9iR2 (9.2.0.5 and above) Note 10gR2 Note 11gR1 Note 11gR2 Note
	Oracle Solaris Containers	Solaris 11	11gR2 (11.2.0.3 and above)	11gR2 (11.2.0.3 and above)
	Oracle Solaris 8 Branded Zone Oracle Solaris 9 Branded Zone	Solaris 10	10gR2 Note 11gR1 Note	N/A
Oracle Solaris x86-64	Oracle Solaris Containers	Solaris 10	10gR1 Note 10gR2 Note 11gR2 Note	10gR2 Note 11gR2 Note
	Oracle Solaris Containers	Solaris 11	11gR2 (11.2.0.3 and above)	11gR2 (11.2.0.3 and above)
	Oracle VM Server for x86-64	Solaris 10	10gR1 Note 10gR2 Note 11gR2 Note	10gR2 Note 11gR2 Note
	Oracle VM Server for x86-64	Solaris 11	11gR2 (11.2.0.3 and above)	11gR2 (11.2.0.3 and above)

Oracle Solaris Sparc Notes

For 11gR1, please apply Oracle patch 8799617 and OS patch 138888-07 on Solaris 10 10/08 (Update6) or later.
Oracle supports the single instance database in "Oracle Solaris 8 Containers" & "Oracle Solaris 9 Containers" (also known as Solaris 8 Branded Zones & Solaris 9 Branded Zones) on a host running Solaris 10. Supported versions are Solaris 8 Containers 1.0.1 and Solaris 9 Containers 1.0.1 running on Solaris 10 Update 10/8 and later. Please check the documentation for the appropriate Oracle Database version to ensure the corresponding Solaris versions are supported. See Documents for further information:
Using Oracle RAC on Oracle Solaris Containers within an Oracle VM Server for SPARC (LDoms) is not supported. Oracle Single Instance database on Oracle Solaris Containers within an Oracle VM Server for SPARC (LDoms) is supported
11gR1 (11.1.0.7) Solaris 10 Logical Domains on Sparc 64-bit requires patch 7535429
Oracle Solaris Logical Domains are supported with Oracle RAC 10gR2, 11gR1 and 11gR2 with Solaris version 10 Update 8 or later with Oracle Solaris Cluster 3.2 1/09 and later versions of 3.2 and Oracle Solaris Cluster 3.3
Oracle Solaris Logical Domains are supported with Oracle RAC 11gR2 with Solaris version 10 Update 6 or later (patches 142900-12, 141870-03) with Oracle Clusterware 11.2
Please reference My Oracle Support note 317257.1 for best practices document for deploying Oracle Single Instance database in a Solaris Container.
Please reference the RAC/Container Best Practices document for deploying Oracle RAC in Solaris Containers.
Oracle Solaris Containers are supported with Oracle RAC 9iR2 (9.2.0.5 and above), 10gR2 and 11gR1 (with Oracle Solaris Cluster on SPARC). Solaris version 10 Update 7 or later (patches 141444-09, 143055-01, 142900-06, 143137-04 "md patch") with Oracle Solaris Cluster 3.3 and 3.2u2 patched to 126106-39 or later.
Oracle Solaris Containers are supported with Oracle RAC 11.2.0.2 with patch 12419331 (for Oracle Solaris Cluster on SPARC). Solaris 10 9/10 (Update 9) or later (patch 142909-17) with Oracle Solaris Cluster 3.3.
Oracle Solaris Containers are supported with Oracle RAC 10gR2, 11gR1 and 11gR2 with Oracle Clusterware on SPARC. Solaris version 10 Update 8 or later (patches 142900-14, 143055-01) with
- Oracle Clusterware 10.2 (patch 9352164)
- Oracle Clusterware 11.1 (patch 9207257, 9352179)
- Oracle Clusterware 11.2.0.1 (patch 11840629)
- Oracle Clusterware 11.2.0.2 (patch 12419353)

Oracle Solaris x86-64 Notes

Please reference My Oracle Support note 317257.1 for best practices document for deploying Oracle Single Instance database in a Solaris Container.
Please reference the RAC/Container Best Practices document for deploying Oracle RAC in Solaris Containers.
Oracle Solaris Containers are supported with Oracle RAC 10gR2 and 11gR2 with Oracle Clusterware on Solaris x86-64. Solaris 10 10/09 (Update 8) or later (patches 142901-15, 142934-02) with
- Oracle Clusterware 10.2 (patch 7172531)
- Oracle Clusterware 11.2.0.1 (patch 11840629)
- Oracle Clusterware 11.2.0.2 (patch 12419353)
Oracle Solaris Containers are supported with Oracle RAC 10gR2 (patch 9654991) and 11gR2 (11.2.0.2 with patch 12419331) with Oracle Solaris Cluster on Solaris x86-64. Solaris 10 9/10 (Update 9) or later (patch 142910-17) with Oracle Solaris Cluster 3.3.
Please check the documentation for the appropriate Oracle Database version to ensure the corresponding Solaris versions are supported. See the following document for further information:

Monday, December 5, 2011

oracle solaris cluster 33 pricing

In light of dec/4/2011 OSC 4.0 announcement, I just want to keep a record on the OSC3.3.1 pricing
There is only one OSC Enterprise Edition per processor pricing

2 years: $1050
3 years: $1500
4 years: $1800
5 years: $2100
Perpetual:$3000

1st years support: $660

Thursday, December 1, 2011

Fujitsu/LSI 16 core SPARC64-IXfx

The register detail the new FJ SPARC64-IXfx chip and PrimeHPC FX10
and cpu-world provide some detail

SPARC64-IXfx

16 core

each core :

32KB L1 D$, 32KB L1 I$
two INT IU
two address calculaion unit,
four FP unit FMA allow fat SIMD span two FP unit (8 flops/core)
a Storage Unit SU (Ld/Store)

12MB L2$
integrated memory controller/DDR3

64GB
Bandwidth 86GB/sec

designed by FJ with LSI
Fabbed by TSMC @40nm
21.9mm x22.1mm
110W
1.85Ghz@128flops=236 gflops
4 Tofu Interconnect interface

handles collective operations
Tofu router: 10 Tofu links
6D mesh/torus
PCI_E2 controller
65nm@312.5Mhz
10 bi-diectional orts@5GB/sec peak of 100GB/sec switching capacity

PRIMEHPC FX10 SW

4 SPARC64-IXfs per blade
4 Tofu interconnect chips per blade
cooled with water blocks attached to rear door water jackets
base PrimeHPC FX10

64 racks@$650k
6144 process
385TB RAM
384 I/O nodes
1536 expansion slots
1.5 pflops peak
1.4 MegaW

Fujitsu Exabyte File System/variant of Lustre FS

Wednesday, November 23, 2011

top three china HPC systm in china hpc top100 and topc500 2011

the Top three sites from china hpc top100 2011 and top500 2011

Tianhe-1A:天河一号

top500 (2nd)

NUDT YH MPP, 7168 x2 xeon x5670 6C 2.93Ghz, 7168 NIVIDIA M2050
Total core= 7168x(2x6+14)=186368 (agree with top500 list)
Rpeak=7168x(2x6x2.93x4+14x32x1.15)=4701061 (top500=4701000)

top100(1st)

NUDT YH MPP, 7168 x2 xeon x5670 6C 2.93Ghz, 7168 NIVIDIA M2050 and 2048 FT-1000, 8C and 1GZ (it was list as Hex Core, a misprint)
Total core= 7168x(2x6+14)+ 2048x8=202750 (agree with top100)
Rpeak=470100, mean that it does not include FT-1000 in Rpeak and linpack
It seems that to highlight china own FT-1000 chip it need to be included in top100
we all know that it is not easy to run linpack in a mix-CPU environment so it make sense not to include FT-1000 in linpack

Nebulae:曙光星云

top500 (4th)

Dawning TC6300 Blade, 4640 x2 xeon x5650 6C 2.66Ghz, 4640 NIVIDIA C2050
Total core= 4640x(2x6+14)=120640 (agree with top500 list)
Rpeak=4640x(2x6x2.66x4+14x32x1.15)=2982964 (top500=2984300)

top100(4th)

Dawning TC6300 Blade,2560 x2 xeon x5650 6C 2.66Ghz, 2560 NIVIDIA C2050,but for calculation to match the reported total core cout the total blade is only 2016
Total core= 2016x(2x6+14)=52416(agree with top100 list)
Rpeak=2016x(2x6x2.66x4+14x32x1.15)=1296046 (top100=1296320)

so between top100 and top500 report Nebulae system almost double in size

if one look at the Top500 site Nebulae system did not change between 06/2010 and 11/2011
One just wondering why Nebulae NEED to report few system for China HPC100

SUNWAY Blue light:神威蓝光

top500 (14th)

SUNWAY bluelight MPP shenwei SW1600 16 core 975Mhz
Total core= 8575x(16)=137200 (agree with top500 list)
Rpeak=8575x(16x.975x8)=1070160 (agree with top500 list)

top100(2nd)

SUNWAY bluelight MPP shenwei SW1600 16 core 975Mhz
Total core= 8575x(16)=137200 (agree with top100 list)
Rpeak=8575x(16x.975x8)=1070160 (agree with top100 list)

SW1600 is 16 core CPU that use 5x5 crossbar to connect 4 (4core) group and IO system that support PCI-E and ge NIC on chip and mgmt channel
each 4 core group support DDR3/1333 memory channel/bank
PCI-E support 8x5 gbps
each system bd has two CPU, it is not know how these two CPU connected
1U system has 4 system bd, I ASSUME that each system bd has its own QDR connection
SUNWAY use 324 port and 256 port IB switch

324 is 3 layer COS network of 3x9 of 36 port chip
256 is two layer fat tree, I could not understand the network diagram

Due to the high density of CPU in 1U, system use water cooling
overall very impressive system with only 9 rack

May be by next year all these systems will use China's own chip

godson
shenwei

To replace these Xeon chip and NIVIDIA with GODSON or SHENWEI will cost money , but Today China does not have shortage of MONEY

Thursday, November 17, 2011

china hpc in top500 2011

some observations on the recent top500 2011 list from china

74 entry in top500
most are xeon cpu, some amd and shenwei sw1600
FT-1000 is list in china top100 but not in top500
there is no GODSON, loongson cpu

IMHO, these list donot provide the whole picture of the system

no detail list of the cpu spec
no detail of the server arch

hdd
ram
pci slots
io chip

it is about time for top500 to provide more detail info or require the org to provide more detail info

Wednesday, November 9, 2011

Oracle Solaris 11 launch

Today Oracle announce the Solaris 11 ,1st Cloud OS
It could/should also take the name Solaris 12c
press release highlight some keep features
the download page for s11 (11/11/2011)
Oracle Solaris 11 page that include how to and whitpaper
Solaris 11 documents

One can watch the webcast later
There will be 100 city Solaris 11 tour
These are the some interesting points during the launch

There is no change in open source policy for Oracle Solaris
S10 will be supported for a long time, T5 and M4 come out next year will support s10
S11 zone as NFS server feature will not be backport to s10
Solaris is looking at Ksplice technology
Oracle Solaris does not have any plan for KVM
Ops-center for S11 will come out next year
Oracle Solaris Cluster will support s11 (I did not see any docs or announcement)
SPARC super cluster will be GA in Dec/2011, SPARC will run S11 and Cell will run OEL, Oracle Solaris Cluster is optional addon
Exalogic run Solaris 11 will support zone
There will be many White paper/How To on S11
There is no tool for in place "Defragment" of zpool for s10 , if one send/receive zpool it will get "defragement", S11 handle this better
There is no in place upgrade from s10 to s11, but one can move s10 to s10 zone in s11; p2v or v2v

some pictures