blog'o thnet

The Oracle Solaris 11 Early Adopter release is available for some days by now. This EA release provides access to the final (complete) functionality which will be delivered in Oracle Solaris 11 GA. Although I only played with it for a few days, here are my very, very first notes about things I found interesting to mention, in no particular order.

• I noticed that the Oracle Solaris 11/11 release (and not EA) was mentioned in one of the subsections of the provided draft for the documentation. Was this inadvertently forgotten... on purpose? ;)
• The support for Flash Archives seems to have finally disappeared. I know about the Distribution Constructor argument, but a flar(1M) (as an mksysb(1) for AIX) definitely has a special place in the Solaris ecosystem (particularly for crash recovery scenario).
• The -x option has been removed from the vi(1) command (among others), and is now replaced by the use of the encrypt(1) command. I know a place where she will be missed: you know who you are :)
• It seems that the network-boot-arguments command is now supported to be able to set IP configuration directly from the OBP, just in case a DHCP server is not an option to get this information at installation time (as we can do on IBM AIX using the IPL configuration from the SMS menu).
• Automated Installer is now able to install Zones along with the main system.
• New utilities are provided to help migrating JumpStart configuration files to AI manifests (I did not use them yet though).
• RBAC things have changed a little, for example the provided profiles are now defined under different files under the /etc/security/prof_attr.d directory instead of a single file (/etc/security/prof_attr) before that (even in Solaris 11 Express). More, there is no Primary Administrator profile anymore, but a new System Administrator profile which doesn't have some security privileges the old profile has (can not read the /etc/shadow file for example).
• The useradd(1m) command has well evolved. This utility is now able to automatically create a dedicated ZFS dataset as the home directory (which is not a directory anymore :)) if the -d flag is given, to populate the /etc/auto_home file, and to enable to autofs service to serve the /home content automatically as needed.
• Although the default shell is now bash(1) (why not the newly integrated ksh93(1)?), the default PATH seen in OpenSolaris releases and Solaris 11 Express, which used to set GNU tools in front of SYSV commands, is reverted back to a more classical and fully functional paths: /usr/bin:/usr/sbin. At least the ls -v is OK again by default. Nonetheless, the path /usr/gnu/bin is here for whoever is interested.
• An interesting change is the motivation to put out some old and well known configuration file. For better or for worse, the /etc/nodename is dead in Solaris 11. It is replaced by a property of a new SMF. So in order to change the nodename of a host, you must now do:

  # svccfg -s node setprop config/nodename = "mynewnodename"
  # svcadm refresh node
  

• In the same vein, the /etc/default/init is replaced by a SMF too. The SMF is named system/environment:init, and the corresponding environment properties are environment/LANG, environment/LC_*, and environment/TZ.
• If you want to be able to manually configure the network, you have to disable NWAM, to change the active Network Configuration Profiles (NCP) and enable traditional configuration:

  # netadm enable -p ncp DefaultFixed
  

• The old sys-unconfig(1m) command is now replaced by a more powerful sysconfig(1m) utility which can unconfigure or reconfigure a Solaris instance, and generate a configuration profile which can be used to configure a system, or a Zone (exit the sysidcfg file).
• Shares (NFS, SMB) are now supported inside a non-global zone.
• The default networking mode is switched to exclusive-IP.
• Similarly as can be found for SRM and privileges configuration settings with automatic Resources Pools, a VNIC can now be automatically instantiated for the time the Zone is booted, and automatically removed when she shuts down.
• A new mode for the Zone known as Read-Only permits to create some instance which may be more or less writable, i.e. some parts may not be changed (configuration, file systems, etc.).
• The IPS packages are now automagically updated in each Zones using Boot Environments.
• Last point in this quick entry, the default locale positioned is en_US.UTF-8, and not just the old C. Well, not a big deal, but I found some tools which have issued some warnings against this locale such as expect for example.

So, I think that Solaris 11 is getting better, even from a Solaris 11 Express experience standpoint. Some choices are surprising, but the overall seems coherent and works as expected. A more longer experience in real user cases will be necessary to judge of this (very big) release, but I am mostly pleased with the direction taken by Solaris, and I am exited to put all of this new stuff in production!

Here is a little press review around Oracle technologies, and Solaris in particular:

Sun ZFS 7000 Storage Appliance 2010.Q3.4.0 Release

• http://blogs.oracle.com/fishworks/entry/sun_zfs_7000_storage_appliance

A new minor update to the 2010.Q3 software has been posted. Note that the release has over 80 bug fixes and includes a Disk Shelf SIM firmware upgrade, ZFS resilvering performance improvements and Update Healthchecks.

What's New in Oracle Solaris 11: Oracle University session at Oracle Open World for System administrators and Developers New

• http://www.oracle.com/openworld/learn/other/oracle-university/index.html#solaris

Course Description: This 1-day seminar provides a detailed look at the newest key features of Oracle Solaris 11 and how to use these features within your deployments. You will learn how to implement the new packaging system, how using the default file system, ZFS, will improve your data management capabilities, how to deliver fully virtualized networking, and how to use the advanced user, application, and device security.

Developing Enterprise Applications for Oracle Solaris

• http://www.oracle.com/technetwork/server-storage/solaris11/technologies/developer-isv-422892.html

Oracle Solaris delivers a highly robust, scalable and secure platform for developing and delivering mission-critical enterprise and ISV applications:
. Run Oracle Solaris 10 applications unmodified in Oracle Solaris 10 Zones on Oracle Solaris 11
. Protect your investment with the industry's first and most extensive binary compatibility and source code guarantees
. Leverage Oracle Solaris 11 cloud-ready application deployment technologies such as IPS, SMF and Zones
. Maximize application performance, increase application observability and enhance developer productivity with Oracle Solaris Studio
. [...]

IBM POWER Roadmap... 7+ now late and only an almost 3 years projection for 8?

• http://netmgt.blogspot.com/2011/08/ibm-power-roadmap-now-late-and-almost-3.html

Thank you for the IBM August 2011 POWER Roadmap tha the public marketplace has been begging for... did we miss the POWER7+ release??? A POWER 7 February 2010 launch would have POWER 7+ August 2011 launch (and today is August 31, so unless there is a launch in the next 23 hours, it looks late to me.) Sketchy details on something possibly 3 years out??? No commitment beyond (almost) 3 years for POWER???

Solaris 10 Update 10 Eminent and Imminent

• http://netmgt.blogspot.com/2011/08/solaris-10-update-10-eminent.html

Oracle's Solaris and SPARC public road map is pretty clear - Solaris 10 Update 10 release is expected 2H 2011 with Solaris 10 Update 11 scheduled for 2H 2012.

In a little more than 1 month away, Oracle OpenWorld 2011 is scheduled (October 2-6, 2011) to occur, which means significant announcements!

Disable Large Segment Offload (LSO) in Solaris 10

• http://nilesh-joshi.blogspot.com/2011/08/disable-large-segment-offload-lso-in.html

LSO saves valuable CPU cycles by allowing the network protocol stack to handle large segments instead of the traditional model of MSS (TCP Maximum Segment Size) sized segments. In the traditional network stack, the TCP layer segments the outgoing data into the MSS sized segments and passes them down to the driver. This becomes computationally expensive with 10 GigE networking because of the large number of kernel functional calls required for every MSS segment. With LSO, a large segment is passed by TCP to the driver, and the driver or NIC hardware does the job of TCP segmentation (LSO offload the segmentation job on Layer 4 to the NIC driver). An LSO segment may be as large as 64 KByte. The larger the LSO segment, better the CPU efficiency since the network stack has to work with smaller number of segments for the same throughput.

So in simple words, use LSO for better network performance while reducing processor (CPU) utilization.

100% of Solaris users use RBAC

• http://blogs.oracle.com/darren/entry/100_of_solaris_users_use

Some of us in the Solaris Security Engineering group been asked a few times recently questions like "so how many customers actually use Solaris RBAC ?" The answer we give is usually variant of "For Solaris 10 onwards 100% of users use RBAC". Surely that is wrong and we can't guarantee 100% of users of Solaris 10 and Solaris 11 are or will be using RBAC can we ? We don't have data to back that up because we don't even know who all the users of Solaris actually are. It actually is correct we don't need data on usage to back it up. The reason being you can't turn RBAC off in Solaris 10 onwards it is always in use in parts of the system that 100% of users of Solaris always use.

The kernel always checks Solaris's fine grained privileges (82 distinct privileges in Solaris 11 Express), even if the process is running "as root". So 100% of Solaris systems make RBAC privilege checks.

Oracle Unveils Oracle VM 3.0

• http://www.oracle.com/us/corporate/press/459406

News Facts:
. Today at an event for customers, partners and industry experts, Oracle announced Oracle VM 3.0, the latest release of Oracle’s server virtualization and management solution.
. Oracle VM 3.0 is suitable for all datacenter workloads and features new policy-based management capabilities, advanced storage management via the Oracle VM Storage Connect plug-in API; centralized network configuration management, improved ease-of-use and Open Virtualization Format (OVF) support.
. With the centralization of storage management alongside of logical network configuration and management, Oracle VM 3.0 allows administrators to streamline and automate end-to-end virtual machine provisioning for a significant reduction in time and overhead, simplifying IT processes and helping to reduce costs.
. Oracle VM 3.0 helps customers deploy enterprise software in a rapid, repeatable and error-free manner with immediate availability of over 90 Oracle VM Templates for Oracle applications, middleware and databases.
. Oracle VM 3.0 is four times more scalable than the latest VMware offering, supporting up to 128 virtual CPUs per virtual machine, at a fraction of the cost. Oracle VM 3.0 demonstrated support for up to 160 physical CPUs and 2TB memory using Oracle’s Sun Fire X4800 M2 servers.
. When compared to VMware vSphere5 running Red Hat Enterprise Linux guest VMs, Oracle VM 3.0 running Oracle Linux guest VMs is four times less expensive.
Oracle VM is free to download, has zero license cost, and affordable, enterprise-quality support is offered through a simple subscription model per server. Terms, conditions and restrictions apply.
[...]

Using smpatch to apply Solaris Cluster patches and other enhancements

• http://blogs.oracle.com/patch/entry/using_smpatch_to_apply_solaris

It is now possible again to use the in-built Solaris 10 patch automation utility, 'smpatch' / Update Manager, to download patches for products such as Oracle Solaris Cluster and Oracle Solaris Studio, as well as Oracle Solaris Operating System patches. It is now also possible again to use 'smpatch' / Update Manager on 3rd party hardware.

These steps effectively switch 'smpatch' / Update Manager from using hardware serial number based access entitlement to User based access entitlement, similar to the access entitlement mechanism used when downloading patches via 'wget' or manually via My Oracle Support (MOS).

Oracle Solaris 11 Express Available on Oracle Exadata Database Machines

• http://www.oracle.com/us/corporate/press/454114

News Facts:
. Oracle Solaris 11 Express is available on Oracle Exadata Database Machines X2-2 and X2-8, Oracle today announced.
. Customers can take advantage of the mission-critical reliability, scalability, and security of Oracle Solaris to run their online transaction processing (OLTP), data warehousing and consolidated workloads on the x86-based Oracle Exadata systems.
. With Oracle Exadata, Oracle Solaris customers can rapidly deploy an engineered system to manage the largest and most critical database applications, enabling them to run up to 10x faster with the rock solid stability that Oracle Solaris consistently delivers.

Oracle's SPARC T4 chip: Will you pay Larry's premium?

• http://www.theregister.co.uk/2011/08/22/oracle_sparc_t4_hot_chips/

The SPARC T4 chips are presumably timed to hit the market with the impending Solaris 11, which has been in the making for more than six years and which presumably has been tuned to take every advantage of the SPARC T4 chips. The original Sun roadmap had a eight-core, eight-threaded SPARC T series chip coming out in the second half of 2011 for machines with one to four sockets implemented in a 40 nanometer process from Taiwan Semiconductor Manufacturing Corp. This Yosemite Falls processor was supposed to run at 2.5GHz and be based on a new SPARC T core code-named "VT," presumably short for "Virtual Threads" but neither Sun or Oracle have said what VT is short for (probably not Vermont).

Oracle has been hinting that this new VT core, which is now being called S3 we learn from the Hot Chips presentation, has a feature called the critical thread API. This feature allows a high priority application to grab one thread on a core and hog all of the resources on that core to significantly boost performance of that single thread; the other seven threads on the chip get told to sit tight. In the prior S1 and S2 cores, used in the prior SPARC T1, T2, T2+, and T3 processors, the threads were hard coded and their sharing algorithms were set in stone--etched in silicon, to be more precise.

Nyt om næste generation SPARC T-series

• http://blogs.oracle.com/hwpartner/entry/nyt_om_n%C3%A6ste_generations_sparc

Excerpts from the article:
"The new T4 processor, running at 3GHz or more, has features that will also allow T4-based systems to take on some workloads that today are going to Intel Xeon processors, which today perform faster on single-thread workloads than do the T3- series of SPARC processors."

"At the Hot Chips 2011 conference, an IEEE technical conference held at Stanford University from August 17-19, 2011, Oracle systems engineers described the top features of the new T4 processors, including a 16-stage integer instruction pipeline and enhanced cryptographic performance. Among the business benefits associated with the new design will be: double the amount of per-thread throughput performance, compared to T3 – and a range of 2 to 7 times more single-thread performance for business workloads than T3 processors. Given the binary-compatibility of T3 and T4, this means that the same Oracle Solaris applications that have been running on T3 will see considerable speedup on T4, without recompilation."

Why and How to Use Cluster Check in Oracle Solaris Cluster 3.3 5/11

• http://www.oracle.com/technetwork/articles/servers-storage-admin/o11-068-cluster-check-solaris-485278.html

The latest set of enhancements to cluster check is focused on validating clusters during installation and initial configuration. The intent of these enhancements is to enable administrators to perform the most important of the Enterprise Installation Services (EIS) checks themselves. If EIS personnel are involved, they, too, will benefit from using cluster check, as described here. As a matter of fact, the EIS team as well as the Oracle Support Services team played a vital part in defining and implementing these enhancements.

In addition to a new focus on installation-time and configuration-time checking, a seemingly small but quite important change was made to the checks themselves: the way the results are titled. I always recommend that cluster check be run with the -v (verbose) flag to turn on verbose progress reporting. In the past, the checks were titled with "problem statements" that described a problem. Many people would miss the fact that most or all the checks were passing, so they were alarmed by the titles and thought lots of problems were being discovered on their cluster. Now, all existing checks are titled with "check titles" instead of "problem statements." Most of the titles are actually a question and, typically, a "yes" answer means the cluster passed the check.

Now for the big stuff: There are over forty new checks, many of which apply both before clustering is installed (recall that scinstall(1M) runs cluster check before configuring a node) and right after the initial configuration of services. And even though the focus is on initial installation and configuration, these checks are still useful over the entire life of the cluster.

Core Factor for T4 published

• http://www.oracle.com/us/corporate/contracts/processor-core-factor-table-070634.pdf

Oracle has published an update to the Processor Core Factor Table that lists the (yet to be released) T4 CPU with a factor of 0.5. This leaves the license cost per socket the same compared to T3 and puts T4 in the same league as SPARC64 VII+ and all current x86 CPUs. We will have to wait for the announcement of the CPU until we can actually speak about performance. But this core factor (which is by no means a measure of CPU performance!) seems to confirm what the few other available bits of information seem to be hinting at: T4 will deliver on Oracle's performance claims.

Installing Oracle Solaris 10 Using JumpStart on an Oracle Solaris 11 Express Server

• http://www.oracle.com/technetwork/articles/servers-storage-admin/installjumpstartons11ex-219229.html

If you are familiar with using JumpStart to install the Oracle Solaris 10 operating system on networked SPARC and x86 platforms, then you probably know that JumpStart can be used to install only the Oracle Solaris 10 OS, not the Oracle Solaris 11 Express OS. However, the JumpStart install server can be an Oracle Solaris 11 Express system.

Your Oracle Solaris 11 Express server can do two different jobs:
. Serve Oracle Solaris 11 Express OS installations using Automated Installer. For more information, see Oracle Solaris 11 Express Automated Installer Guide.
. Serve Oracle Solaris 10 OS installations using JumpStart. This article describes how to set up a JumpStart install server on an Oracle Solaris 11 Express system.

Some months ago, I had the great opportunity to be profiled for the Peer-to-Peer column of the Oracle Magazine, which is an international publication published by Oracle on a monthly basis.

Although the interview is now available in the September/October 2011 issue of the Oracle Magazine, it has also been truncated because of some space concerns, and other things related to the format of the publication. Nonetheless, and for those who are interested, here is the complete interview.

Basic information

Brief description of your responsibilities

Multi-platform UNIX systems consultant and administrator in mutualized and virtualized environments. Architecture and expertise building Solaris and UNIX experience in large enterprise such as banking and financial services, IT services, Telecommunication and multimedia companies.

Length of time using Oracle products

For over eleven years I worked with Solaris as a system administrator, but it was Sun Microsystems at this time, not yet Oracle.

Published author?

I am not a published author, but I participated in writing an article in GLMF, and blogging publicly on the Internet for many years.

If so, title/publication/year

. Title: GNU Solaris : Introduction
Publication: GNU/Linux Magazine France (GLMF)
Year: June 2008
Authors: Yves Mettier, Julien Gabel
Participation in the writing of an article on OpenSolaris: brief history, major distributions, use with a concrete case, and evolutions to come.

. I blogged for more than six years at blog'o thnet: http://blog.thilelli.net/

Which version and release of the Oracle database [or for Applications ACEs, which systems] do you currently use?

I work on Solaris systems for more than ten years, from Solaris 2.6 on SPARC systems to Solaris 11 on both x86 and SPARC platforms.

Interview questions

How did you get started in IT?

I had the chance to discover IT and UNIX system administration when I was yet at school preparing a Master's degree in Engineering, but this was very very late in fact, since I was twenty years old at this time! Since then I always found myself passionate by sysadmin on UNIX platforms without interruption, greatly helped by the always evolving technologies, such as virtualization for example.

What is your favorite tool or technique on the job?

In general, I tend to encourage work on topics around virtualization, in particular when based on Zones and ZFS which both completely revolutionized and extends the possibilities of such configurations.

Secondly, I am interested by performance and pinpointing the root cause of a performance penalty problem, or observation. If DTrace is definitely a invaluable new tool introduced in Solaris 10, better understand how things works at the operating system level will help better understand the metric reported by more classical (historical) tools which can always gives us useful information if we are able to interpret them properly.

If you are exploring the new features in Oracle Database or Oracle Fusion Middleware, please tell us which features you're finding valuable, and why. If you are exploring any new features in an Oracle application, please tell use which features you're finding valuable, and why.

Since Solaris 10, the integration of Zones, Solaris Resource Manager and ZFS opened new opportunities with very light overhead, and interesting capabilities for virtualization: dynamic resource capping for CPU and memory at the non-global zone level, and very flexible use for data management through the use of ZFS.

With the upcoming release of Solaris 11, virtual networks will greatly extends the Zones experience, while some other new features will help us at least just as much: for example the new IPS packaging technology in conjunction with boot environments which will make old patching paths a thing of the past, and automates some new capability such as instantly cloning an environment when a package require a reboot (a sort of intelligent automated Live Upgrade).

What advice do you have about how to get into Web, database, and/or application development or software architecture?

Well, some essential points came in mind here: be voluntary, be passionate and interested by what you are doing, educate yourself (helped by appropriate trainings, experienced co-workers, etc.), and keep up-to-date as much as possible!

What would you like to see Oracle, as a company or as a technology, do differently, better, or more of?

Well, good question. Two points here.

As a company, I really liked Oracle to better integrate and co-operate with OpenSolaris, Solaris and OpenSource projects and communities. In fact, I particularly regrets the fact that Oracle will not deliver more regularly development releases of the operating system and its ecosystems (such as Sun did in the past since early 2005) because it gave the interested system administrators and specialists some insight of what will arrive next, and is a good opportunity to test new functionality before they arrived as a finalized product. I miss this opportunity where it was possible to test, to report and to follow next system technologies beforehand, followed by a large community of enthusiasts participating in the Solaris evolution. I am convinced this will greatly help to improve Solaris as a target enterprise OS for Oracle, and get a new and vibrant community as can be found in the Linux world.

As a technology, I think Solaris miss a tool such as nmon on IBM AIX. This nifty tool has the ability to provide lots of interesting and important information live on the systems from one single tool (such as sar but with much more information) but, more interestingly, it has the capability to collect these information to be used by different tools to generate graphs afterward. We can always do this ourselves (self made solutions) or let some expensive enterprise tools do this job, but I really think this may be a very valuable add-on to Oracle Solaris.

What's your favorite Oracle reference book and why? (Put another way, if you were going to the International Space Station for six months and could only take one Oracle reference book, what would it be?)

Without hesitation, my answer is the second book of the latest edition of Solaris Internals "Solaris Performance and Tools, DTrace and MDB Techniques for Solaris 10 and OpenSolaris". This book is a must have since it is really complete, covering both tools and methodologies for performance observability and debugging. It provides a better understanding and metrics interpretation of the output of some already known tools up to the latest utilities found in Solaris 10 and Solaris 11.

What's your favorite thing to do that doesn't involve work?

Although I am involved in system administration and IT communities even when I am not at work, I really like driving my motorcycle... at least when the weather permit ;)

Do you have a favorite vacation spot?

Not a favorite vacation spot in itself, but I like winter sports, especially snowboarding.

Here is a little press review around Oracle technologies, and Solaris in particular:

SPARC64 VIIIfx: New HPC #1

• http://www.top500.org/lists/2011/06/press-release
• http://www.opensparc.net/

The 37th edition of the closely watched list was released Monday, June 20, at the 2011 International Supercomputing Conference in Hamburg. The ranking of all systems is based on how fast they run Linpack, a benchmark application developed to solve a dense system of linear equations.

For the first time, all of the top 10 systems achieved petaflop/s performance – and those are also the only petaflop/s systems on the list. The U.S. is tops in petaflop/s with five systems performing at that level; Japan and China have two each, and France has one.

The K Computer, built by Fujitsu, currently combines 68544 SPARC64 VIIIfx CPUs, each with eight cores, for a total of 548,352 cores—almost twice as many as any other system in the TOP500. The K Computer is also more powerful than the next five systems on the list combined.

Oracle Solaris Cluster 3.3 5/11 announced

• http://blogs.oracle.com/SC/en_US/entry/oracle_solaris_cluster_3_31

Oracle announces Oracle Solaris Cluster 3.3 5/11, delivering unrivaled High Availability (HA) on the Oracle Solaris OS for the largest selection of enterprise applications and databases. Oracle Solaris Cluster unique HA solution for Oracle Solaris' virtual environments lowers the inherent risk of consolidated infrastructures by leveraging redundancy to protect from outages. Oracle Solaris Cluster Disaster Recovery (DR) option extends these benefits to multi-site, multi-cluster architectures to protect against planned and unplanned downtime, in physical and virtual environments.

Oracle Solaris Virtual Machine (VM) Templates

• http://www.oracle.com/technetwork/server-storage/solaris/solaris-vm-405695.html

This page offers a variety of links to Oracle Solaris Virtual Machine Templates. These Oracle VM Templates speed the creation of Proof of Concept environments and other evaluation/development tasks by dramatically simplifying the installation process. Oracle Solaris VM Templates are available for both SPARC and x86-based systems.

Pre-Built Developer VMs (for Oracle VM VirtualBox)

• http://www.oracle.com/technetwork/community/developer-vm/index.html

Learning your way around a new software stack is challenging enough without having to spend multiple cycles on the install process. Instead, we have packaged such stacks into pre-built Oracle VM VirtualBox appliances that you can download, install, and experience as a single unit. Just assemble the downloaded files (if needed), import into VirtualBox (available for free), import, and go!

A Solaris Recommended Patchset to bind them all

• http://blogs.oracle.com/patch/entry/a_solaris_recommended_patchset_to

A collaborative effort between the Software Patch Services, Enterprise Installation Standards (EIS), Sun Risk Analysis System (SRAS) - now renamed Oracle Risk Analysis Services (ORAS) - and the Explominer team in the Oracle Solaris Technical Center (TSC), has achieved this goal with the creation of the Recommended Patchset for Solaris.

Up until now, while the Solaris OS Recommended Patch Cluster was the core basis for Solaris patch recommendations, various teams tended to recommend their own favorite patches on top of this core set. This wasn't just by whim. Each team was looking at patching from a slightly different angle - for example various angles of proactive patching (issue prevention) versus reactive patching (issue correction).

The Recommended Patchset for Solaris is the result of the combined wisdom of the various teams. It is designed for proactive patching (issue prevention). The contents are generic and should be suitable for most customer configurations.

Accelerating SPARC Solaris Servers

• http://www.oracle.com/us/products/servers-storage/servers/sparc-enterprise/public-sparc-roadmap-421264.pdf
• http://digitaldaily.allthingsd.com/files/2009/09/oraclead.jpg

Integrated Load Balancer

• http://blogs.oracle.com/observatory/en_US/entry/integrated_load_balancer

I'm not sure how well known it is that Solaris 11 contains a load balancer. The official documentation, starting with the Integrated Load Balancer Overview, does a great job of explaining this feature. In this blog entry my goal is to provide an implementation example.

For starters, I will be using the HALF-NAT operation mode. Basically, HALF-NAT means that the client's IP address is not mapped so that the servers know the real client address. This is usually preferred for server logging (see ILB Operation Modes for more).

Effective Resource Management Using Oracle Solaris Resource Manager

• http://www.oracle.com/technetwork/articles/servers-storage-admin/o11-055-solaris-rm-419384.pdf

Oracle Solaris Resource Manager provides specific software components and utilities that are used to manage hardware resources. It is integrated into the operating system, and it is available on SPARC and x86/x64 platforms running Solaris 9 or later.

Oracle Solaris Resource Manager is a key enabler for server and workload consolidation and increased resource utilization. It provides the ability to allocate and control major system resources, such as CPU, virtual memory, physical memory, I/O bandwidth, and number of processes. It also implements administrative policies that govern which resources different users can access and, more specifically, how much of a particular resource each user is permitted to use. Based on the implemented policies, all users can receive resources commensurate with their service levels and the relative importance of their work.

Here is a simple procedure switching from a RDAC/fcparray management mode to a MPIO multipath mode for SAN disks presented from an IBM DSXX00 array.

Verification of the current monopath configuration:

# manage_disk_drivers
1: DS4100: currently RDAC; supported: RDAC/fcparray, MPIO
2: DS4300: currently RDAC; supported: RDAC/fcparray, MPIO
3: DS4500: currently RDAC; supported: RDAC/fcparray, MPIO
4: DS4700/DS4200: currently RDAC; supported: RDAC/fcparray, MPIO
5: DS4800: currently RDAC; supported: RDAC/fcparray, MPIO

Listing of the disks from the array:

# fget_config -vA
---dar0---

User array name = 'CUSTOMERSOFT'
dac0 ACTIVE dac5 ACTIVE

Disk     DAC   LUN Logical Drive
utm            127
hdisk7   dac5    6 beastie1_oracle
hdisk14  dac5   13 beastie2_datavg
hdisk15  dac0   14 beastie3_datavg
hdisk2   dac0    1 beastie3_rootvg
hdisk3   dac0    2 beastie4_rootvg
hdisk4   dac5    3 beastie5_rootvg
hdisk5   dac5    4 beastie2_rootvg
hdisk6   dac0    5 bakup
hdisk8   dac0    7 customer1
hdisk9   dac0    8 customer3
hdisk10  dac5    9 customer6
hdisk11  dac0   10 customer14
hdisk12  dac5   11 beastie2_db2
hdisk13  dac0   12 beastie3_scheduler
hdisk16  dac0   15 customer9
hdisk17  dac0   16 customer8

Listing of the disks as seen from the operating system:

# lsdev -Cc disk | grep DS
hdisk2  Available 00-08-02 1814     DS4700 Disk Array Device
hdisk3  Available 00-08-02 1814     DS4700 Disk Array Device
hdisk4  Available 02-00-02 1814     DS4700 Disk Array Device
hdisk5  Available 02-00-02 1814     DS4700 Disk Array Device
hdisk6  Available 00-08-02 1814     DS4700 Disk Array Device
hdisk7  Available 02-00-02 1814     DS4700 Disk Array Device
hdisk8  Available 00-08-02 1814     DS4700 Disk Array Device
hdisk9  Available 00-08-02 1814     DS4700 Disk Array Device
hdisk10 Available 02-00-02 1814     DS4700 Disk Array Device
hdisk11 Available 00-08-02 1814     DS4700 Disk Array Device
hdisk12 Available 02-00-02 1814     DS4700 Disk Array Device
hdisk13 Available 00-08-02 1814     DS4700 Disk Array Device
hdisk14 Available 02-00-02 1814     DS4700 Disk Array Device
hdisk15 Available 00-08-02 1814     DS4700 Disk Array Device
hdisk16 Available 00-08-02 1814     DS4700 Disk Array Device
hdisk17 Available 00-08-02 1814     DS4700 Disk Array Device

Switch to a multipath management for the SAN volumes presented from the DS4700/DS4200 arrays:

# manage_disk_drivers -c 4
DS4700/DS4200 currently RDAC/fcparray
Change to alternate driver? [Y/N] Y
DS4700/DS4200 now managed by MPIO

It is necessary to perform a bosboot before rebooting the system in
order to incorporate this change into the boot image.

In order to change to the new driver, either a reboot or a full
unconfigure and reconfigure of all devices of the type changed
must be performed.

Reboot the system:

# bosboot –a
bosboot: Boot image is 39636 512 byte blocks.

# shutdown –Fr
[...]

Verification of the new multipath configuration:

# manage_disk_drivers
1: DS4100: currently RDAC; supported: RDAC/fcparray, MPIO
2: DS4300: currently RDAC; supported: RDAC/fcparray, MPIO
3: DS4500: currently RDAC; supported: RDAC/fcparray, MPIO
4: DS4700/DS4200: currently MPIO; supported: RDAC/fcparray, MPIO
5: DS4800: currently RDAC; supported: RDAC/fcparray, MPIO

Listing of the disks from the array:

# mpio_get_config -vA
Frame id 0:
    Storage Subsystem worldwide name: 60ab80016253400009786efca
    Controller count: 2
    Partition count: 1
    Partition 0:
    Storage Subsystem Name = 'CUSTOMERSOFT'
        hdisk      LUN #   Ownership          User Label
        hdisk0         1   A (preferred)      beastie3_rootvg
        hdisk1         2   A (preferred)      beastie4_rootvg
        hdisk2         3   B (preferred)      beastie5_rootvg
        hdisk3         4   B (preferred)      beastie2_rootvg
        hdisk4         5   A (preferred)      bakup
        hdisk5         6   B (preferred)      beastie1_oracle
        hdisk6        16   A (preferred)      customer8
        hdisk7         7   A (preferred)      customer1
        hdisk8         8   A (preferred)      customer3
        hdisk9         9   B (preferred)      customer6
        hdisk10       10   A (preferred)      customer14
        hdisk11       11   B (preferred)      beastie2_db2
        hdisk12       12   A (preferred)      beastie3_scheduler
        hdisk13       13   B (preferred)      beastie2_datavg
        hdisk14       14   A (preferred)      beastie3_datavg
        hdisk15       15   A (preferred)      customer9

Listing of the disks as seen from the operating system:

# lsdev -Cc disk | grep DS
hdisk0  Available 06-08-02 MPIO Other DS4K Array Disk
hdisk1  Available 06-08-02 MPIO Other DS4K Array Disk
hdisk2  Available 06-08-02 MPIO Other DS4K Array Disk
hdisk3  Available 06-08-02 MPIO Other DS4K Array Disk
hdisk4  Available 06-08-02 MPIO Other DS4K Array Disk
hdisk5  Available 06-08-02 MPIO Other DS4K Array Disk
hdisk6  Available 06-08-02 MPIO Other DS4K Array Disk
hdisk7  Available 06-08-02 MPIO Other DS4K Array Disk
hdisk8  Available 06-08-02 MPIO Other DS4K Array Disk
hdisk9  Available 06-08-02 MPIO Other DS4K Array Disk
hdisk10 Available 06-08-02 MPIO Other DS4K Array Disk
hdisk11 Available 06-08-02 MPIO Other DS4K Array Disk
hdisk12 Available 06-08-02 MPIO Other DS4K Array Disk
hdisk13 Available 06-08-02 MPIO Other DS4K Array Disk
hdisk14 Available 06-08-02 MPIO Other DS4K Array Disk
hdisk15 Available 06-08-02 MPIO Other DS4K Array Disk
hdisk16 Available 06-08-02 MPIO Other DS4K Array Disk

That's it.