Search Results

Search found 16971 results on 679 pages for 'blogs'.

Page 176/679 | < Previous Page | 172 173 174 175 176 177 178 179 180 181 182 183  | Next Page >

  • Get The Most From MySQL Database With MySQL Performance Tuning Training

    - by Antoinette O'Sullivan
    Get the most from MySQL Server's top-level performance by improving your understanding of perforamnce tuning techniques. MySQL Performance Tuning Class In this 4 day class, you'll learn practical, safe, highly efficient ways to optimize performance for the MySQL Server. You can take this class as: Training-on-Demand: Start training within 24 hours of registering and follow the instructor-led lecture material through streaming video at your own pace. Schedule time lab-time to perform the hands-on exercises at your convenience. Live-Virtual Class: Follow the live instructor led class from your own desk - no travel required. There are already a range of events on the schedule to suit different timezones and with delivery in languages including English and German. In-Class Event: Travel to a training center to follow this class. For more information on this class, to see the schedule or register interest in additional events, go to http://oracle.com/education/mysql Troubleshooting MySQL Performance with Sveta Smirnova  During this one-day, live-virtual event, you get a unique opportunity to hear Sveta Smirnova, author of MySQL Troubleshooting, share her indepth experience of identifying and solving performance problems with a MySQL Database. And you can benefit from this opportunity without incurring any travel costs! Dimitri's Blog If MySQL Performance is a topic that interests you, then you should be following Dimitri Kravtchuk's blog. For more information on any aspect of the Authentic MySQL Curriculum, go to http://oracle.com/education/mysql.

    Read the article

  • How to Achieve OC4J RMI Load Balancing

    - by fip
    This is an old, Oracle SOA and OC4J 10G topic. In fact this is not even a SOA topic per se. Questions of RMI load balancing arise when you developed custom web applications accessing human tasks running off a remote SOA 10G cluster. Having returned from a customer who faced challenges with OC4J RMI load balancing, I felt there is still some confusions in the field how OC4J RMI load balancing work. Hence I decide to dust off an old tech note that I wrote a few years back and share it with the general public. Here is the tech note: Overview A typical use case in Oracle SOA is that you are building web based, custom human tasks UI that will interact with the task services housed in a remote BPEL 10G cluster. Or, in a more generic way, you are just building a web based application in Java that needs to interact with the EJBs in a remote OC4J cluster. In either case, you are talking to an OC4J cluster as RMI client. Then immediately you must ask yourself the following questions: 1. How do I make sure that the web application, as an RMI client, even distribute its load against all the nodes in the remote OC4J cluster? 2. How do I make sure that the web application, as an RMI client, is resilient to the node failures in the remote OC4J cluster, so that in the unlikely case when one of the remote OC4J nodes fail, my web application will continue to function? That is the topic of how to achieve load balancing with OC4J RMI client. Solutions You need to configure and code RMI load balancing in two places: 1. Provider URL can be specified with a comma separated list of URLs, so that the initial lookup will land to one of the available URLs. 2. Choose a proper value for the oracle.j2ee.rmi.loadBalance property, which, along side with the PROVIDER_URL property, is one of the JNDI properties passed to the JNDI lookup.(http://docs.oracle.com/cd/B31017_01/web.1013/b28958/rmi.htm#BABDGFBI) More details below: About the PROVIDER_URL The JNDI property java.name.provider.url's job is, when the client looks up for a new context at the very first time in the client session, to provide a list of RMI context The value of the JNDI property java.name.provider.url goes by the format of a single URL, or a comma separate list of URLs. A single URL. For example: opmn:ormi://host1:6003:oc4j_instance1/appName1 A comma separated list of multiple URLs. For examples:  opmn:ormi://host1:6003:oc4j_instanc1/appName, opmn:ormi://host2:6003:oc4j_instance1/appName, opmn:ormi://host3:6003:oc4j_instance1/appName When the client looks up for a new Context the very first time in the client session, it sends a query against the OPMN referenced by the provider URL. The OPMN host and port specifies the destination of such query, and the OC4J instance name and appName are actually the “where clause” of the query. When the PROVIDER URL reference a single OPMN server Let's consider the case when the provider url only reference a single OPMN server of the destination cluster. In this case, that single OPMN server receives the query and returns a list of the qualified Contexts from all OC4Js within the cluster, even though there is a single OPMN server in the provider URL. A context represent a particular starting point at a particular server for subsequent object lookup. For example, if the URL is opmn:ormi://host1:6003:oc4j_instance1/appName, then, OPMN will return the following contexts: appName on oc4j_instance1 on host1 appName on oc4j_instance1 on host2, appName on oc4j_instance1 on host3,  (provided that host1, host2, host3 are all in the same cluster) Please note that One OPMN will be sufficient to find the list of all contexts from the entire cluster that satisfy the JNDI lookup query. You can do an experiment by shutting down appName on host1, and observe that OPMN on host1 will still be able to return you appname on host2 and appName on host3. When the PROVIDER URL reference a comma separated list of multiple OPMN servers When the JNDI propery java.naming.provider.url references a comma separated list of multiple URLs, the lookup will return the exact same things as with the single OPMN server: a list of qualified Contexts from the cluster. The purpose of having multiple OPMN servers is to provide high availability in the initial context creation, such that if OPMN at host1 is unavailable, client will try the lookup via OPMN on host2, and so on. After the initial lookup returns and cache a list of contexts, the JNDI URL(s) are no longer used in the same client session. That explains why removing the 3rd URL from the list of JNDI URLs will not stop the client from getting the EJB on the 3rd server. About the oracle.j2ee.rmi.loadBalance Property After the client acquires the list of contexts, it will cache it at the client side as “list of available RMI contexts”.  This list includes all the servers in the destination cluster. This list will stay in the cache until the client session (JVM) ends. The RMI load balancing against the destination cluster is happening at the client side, as the client is switching between the members of the list. Whether and how often the client will fresh the Context from the list of Context is based on the value of the  oracle.j2ee.rmi.loadBalance. The documentation at http://docs.oracle.com/cd/B31017_01/web.1013/b28958/rmi.htm#BABDGFBI list all the available values for the oracle.j2ee.rmi.loadBalance. Value Description client If specified, the client interacts with the OC4J process that was initially chosen at the first lookup for the entire conversation. context Used for a Web client (servlet or JSP) that will access EJBs in a clustered OC4J environment. If specified, a new Context object for a randomly-selected OC4J instance will be returned each time InitialContext() is invoked. lookup Used for a standalone client that will access EJBs in a clustered OC4J environment. If specified, a new Context object for a randomly-selected OC4J instance will be created each time the client calls Context.lookup(). Please note the regardless of the setting of oracle.j2ee.rmi.loadBalance property, the “refresh” only occurs at the client. The client can only choose from the "list of available context" that was returned and cached from the very first lookup. That is, the client will merely get a new Context object from the “list of available RMI contexts” from the cache at the client side. The client will NOT go to the OPMN server again to get the list. That also implies that if you are adding a node to the server cluster AFTER the client’s initial lookup, the client would not know it because neither the server nor the client will initiate a refresh of the “list of available servers” to reflect the new node. About High Availability (i.e. Resilience Against Node Failure of Remote OC4J Cluster) What we have discussed above is about load balancing. Let's also discuss high availability. This is how the High Availability works in RMI: when the client use the context but get an exception such as socket is closed, it knows that the server referenced by that Context is problematic and will try to get another unused Context from the “list of available contexts”. Again, this list is the list that was returned and cached at the very first lookup in the entire client session.

    Read the article

  • The most challenging part of blogging about OpenWorld is…

    - by Irem Radzik
    Normal 0 false false false EN-US X-NONE X-NONE MicrosoftInternetExplorer4 /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-qformat:yes; mso-style-parent:""; mso-padding-alt:0in 5.4pt 0in 5.4pt; mso-para-margin:0in; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:11.0pt; font-family:"Calibri","sans-serif"; mso-ascii-font-family:Calibri; mso-ascii-theme-font:minor-latin; mso-fareast-font-family:"Times New Roman"; mso-fareast-theme-font:minor-fareast; mso-hansi-font-family:Calibri; mso-hansi-theme-font:minor-latin; mso-bidi-font-family:"Times New Roman"; mso-bidi-theme-font:minor-bidi;} ...not knowing where to start. Do I talk about the great presentations from our partners and executives in our keynote sessions; do I write about the music festival, or many great sessions we had in the Data integration track? A short blog can never do justice. For now I will stick to our data integration sessions for those who could not attend with so many other sessions running concurrently. And in the coming weeks we will be writing more about what we talked in our sessions and what we learned from our customers and partners. For today, I will give some of the key highlights from Data Integration sessions that took place on Wednesday and Thursday of last week  On Wednesday, GoldenGate was highlighted in multiple Database and Data Integration sessions. I found particularly the session about Oracle’s own use of GoldenGate for its large E-Business Suite implementation for supply chain management and service contract management very interesting. In 2011, Oracle implemented a new operational reporting system using GoldenGate real-time data replication to an operational data store that leverages data from E-Business Suite.The results are very impressive. Data freshness improved by 2,210X while report run performance improved by 60X. For more information on this implementation and its results please see the white paper: Real-Time Operational Reporting for E-Business Suite via GoldenGate Replication to an Operational Data Store Other sessions that provided very rich content were: "Best Practices for Conflict Detection and Resolution in Oracle GoldenGate for Active/Active", "Tuning and Troubleshooting Oracle GoldenGate on Oracle Database", "Next-Generation Data Integration on Oracle Exadata" and "Accelerate Oracle Data Integrator with Advanced Features, SOA, Groovy, SDK, and XML". Below is a slide presented by Stephan Haisley in the Tuning and Troubleshooting Oracle GoldenGate session. If you missed them during OpenWorld, I highly recommend downloading the slides. We will continue to blog about these topics and related resources. .

    Read the article

  • O the Agony - Merging Scrum and Waterfall

    - by John K. Hines
    If there's nothing else to know about Scrum (and Agile in general), it's this: You can't force a team to adopt Agile methods.  In all cases, the team must want to change. Well, sure, you could force a team.  But it's going to be a horrible, painful process with a huge learning curve made even steeper by the lack of training and motivation on behalf of the team.  On a completely unrelated note, I've spent the past three months working on a team that was formed by merging three separate teams.  One of these teams has been adopting and using Agile practices like Scrum since 2007, the other was in continuous bug fix mode, releasing on average one new piece of software per year using semi-Waterfall methods.  In particular, one senior developer on the Waterfall team didn't see anything in Agile but overhead. Fast forward through three months of tension, passive resistance, process pushback, and you have seven people who want to change and one who explicitly doesn't.  It took two things to make Scrum happen: The team manager took a class called "Agile Software Development using Scrum". The team lead explained the point of Agile was to reduce the workload of the senior developer, with another senior developer and the manager present. It's incredible to me how a single person can strongly influence the direction of an entire team.  Let alone if Scrum comes down as some managerial decree onto a functioning team who have no idea what it is.  Pity the fool. On the bright side, I am now an expert at drawing Visio process flows.  And I have some gentle advice for any first-level managers: If you preside over a team process change, it's beneficial to start the discussion on how the team will work as early as possible.  You should have a vision for this and guide the discussion, even if decisions are weeks away.  Don't always root for the underdog.  It's been my experience that managers who see themselves as compassionate and caring spend a great deal of time understanding and advocating for the one person on the team who feels left out.  Remember that by focusing on this one person you risk alienating the rest of the team, allow tension to build, and delay the resolution of the problem. My way would have been to decree Scrum, force all of my processes on everyone else, and use the past three months ironing out the kinks.  Which takes us all the way back to point number one. Technorati tags: Scrum Scrum Process Scrum and Waterfall

    Read the article

  • Jerome has written a nice article on integrating SceneBuilder with several IDEs

    - by daniel
    My colleague Jerome Cambon has written a very nice article about how to get SceneBuilder working with several IDEs. The JavaFX SceneBuilder is at the root a stand-alone tool - but there are various tweaks and tricks that you can use to make its use in conjunction with your favorite IDE a more enjoyable experience. In his article - Jerome shows how this can be done with NetBeans (7.3), Eclipse, with Tom's excellent e(fx)clipse plugin, and IntelliJ IDEA. Good work Jerome!

    Read the article

  • Oracle ATG Web Commerce 10 Implementation Developer Boot Camp - Reading (UK) - October 1-12, 2012

    - by Richard Lefebvre
    REGISTER NOW: Oracle ATG Web Commerce 10 Implementation Developer Boot Camp Reading, UK, October 1-12, 2012! OPN invites you to join us for a 10-day implementation bootcamp on Oracle ATG Web Commerce in Reading, UK from October 1-12, 2012.This 10-day boot camp is designed to provide partners with hands-on experience and technical training to successfully build and deploy Oracle ATG Web Commerce 10 Applications. This particular boot camp is focused on helping partners develop the essential skills needed to implement every aspect of an ATG Commerce Application from scratch, (not CRS-based), with a specific goal of enabling experienced Java/J2EE developers with a path towards becoming functional, effective, and contributing members of an ATG implementation team. Built for both new and experienced ATG developers alike, the collaborative nature of this program and its exercises, have proven to be highly effective and extremely valuable in learning the best practices for implementing ATG solutions. Though not required, this bootcamp provides a structured path to earning a Certified Oracle ATG Web Commerce 10 Specialization! What Is Covered: This boot camp is for Application Developers and Software Architects wanting to gain valuable insight into ATG application development best practices, as well as relevant and applicable implementation experience on projects modeled after four of the most common types of applications built on the ATG platform. The following learning objectives are all critical, and are of equal priority in enabling this role to succeed. This learning boot camp will help with: Building a basic functional transaction-ready ATG Web Commerce 10 Application. Utilizing ATG’s platform features such as scenarios, slots, targeters, user profiles and segments, to create a personalized user experience. Building Nucleus components to support and/or extend application functionality. Understanding the intricacies of ATG order checkout and fulfillment. Specifying, designing and implementing new commerce features in ATG 10. Building a functional commerce application modeled after four of the most common types of applications built on the ATG platform, within an agile-based project team environment and under simulated real-world project conditions. Duration: The Oracle ATG Web Commerce 10 Implementation Developer Boot Camp is an instructor-led workshop spanning 10 days. Audience: Application Developers Software Architects Prerequisite Training and Environment Requirements: Programming and Markup Experience with Java J2EE, JavaScript, XML, HTML and CSS Completion of Oracle ATG Web Commerce 10 Implementation Specialist Development Guided Learning Path modules Participants will be required to bring their own laptop that meets the minimum specifications:   64-bit PC and OS (e.g. Windows 7 64-bit) 4GB RAM or more 40GB Hard Disk Space Laptops will require access to the Internet through Remote Desktop via Windows. Agenda Topics: Week 1 – Day 1 through 5 Build a Basic Commerce Application In week one of the boot camp training, we will apply knowledge learned from the ATG Web Commerce 10 Implementation Developer Guided Learning Path modules, towards building a basic transaction-ready commerce application. There will be little to no lectures delivered in this boot camp, as developers will be fully engaged in ATG Application Development activities and best practices. Developers will work independently on the following lab assignments from day's 1 through 5: Lab Assignments  1 Environment Setup 2 Build a dynamic Home Page 3 Site Authentication 4 Build Customer Registration 5 Display Top Level Categories 6 Display Product Sub-Categories 7 Display Product List Page 8 Display Product Detail Page 9 ATG Inventory 10 Build “Add to Cart” Functionality 11 Build Shopping Cart 12 Build Checkout Page  13 Build Checkout Review Page 14 Create an Order and Build Order Confirmation Page 15 Implement Slots and Targeters for Personalization 16 Implement Pricing and Promotions 17 Order Fulfillment Back to top Week 2 – Day 6 through 10 Team-based Case Project In the second week of the boot camp training, participants will be asked to join a project team that will select a case project for the team to implement. Teams will be able to choose from four of the most common application types developed and deployed on the ATG platform. They are as follows: Hard goods with physical fulfillment, Soft goods with electronic fulfillment, a Service or subscription case example, a Course/Event registration case example. Team projects will have approximately 160 hours of use cases/stories for each team to build (40 hours per developer). Each day's Use Cases/Stories will build upon the prior day's work, and therefore must be fully completed at the end of each day. Please note that this boot camp intends to simulate real-world project conditions, and as such will likely require the need for project teams to possibly work beyond normal business hours. To promote further collaboration and group learning, each team will be asked to present their work and share the methodologies and solutions that they've applied to their cases at the end of each day. Location: Oracle Reading CVC TPC510 Room: Wraysbury Reading, UK 9:00 AM – 5:00 PM  Registration Fee (10 Days): US $3,375 Please click on the following link to REGISTER or  visit the Oracle ATG Web Commerce 10 Implementation Developer Boot Camp page for more information. Questions: Patrick Ty Partner Enablement, Oracle Commerce Phone: 310.343.7687 Mobile: 310.633.1013 Email: [email protected]

    Read the article

  • JDeveloper and ADF at UKOUG

    - by Grant Ronald
    This year, Oracle ADF and JDeveloper has a big showing at the UKOUG (about 22 hours worth!!)- Europe's largest Oracle User Group.  There are three days packed with awesome ADF content delivered by some of the leading lights in ADF Developement including Duncan Mills, Frank Nimphius, Shay Shmeltzer, Susan Duncan, Lucas Jellema, Steven Davelaar, Sten Vesterli (and I'll be there as well!). Please make sure you refer to the official agenda for timings but an outline is here (if you think there are any sessions I have missed let me know and I will add them) Monday 10:00 - 10:45 - Deepdive into logical and physical data modeling with JDeveloper 10:00 - 12:15 - Debugging ADF Applications 12:15 - 13:15 - Learn ADF Task Flows in 60 Minutes 14:30 - 15:15 - ADF's Hidden Gem - the Groovy scripting language in Oracle ADF 15:25 - 16:10 - ADF Patterns for Forms Conversions 16:35 - 17:35 - Dummies Guide to Oracle ADF 16:35 - 17:35 - ADF Security Overview - Strategies and Best Practices 17:45 - 18:30 - A Methodology for Enterprise Applications with Oracle ADF Tuesday 09:00 - 10:00 - Real World Performance Tuning for Oracle ADF 11:15 - 12:15 - Keynote: Modern Development, Mobility and Rich Internet Applications 11:15 - 12:15 - Migration to Fusion Middleware 11g: Real world cases of Forms, ADF and Identity Management upgrades 14:40 - 15:20 - What's new in JDeveloper 11gR2 14:40 - 15:20 - Development Tools Roundtable 15:35 - 16:20 - ALM in Jdeveloper is exciting! 16:40 - 17:40 - Moving Oracle Forms to Oracle ADF: Case Studies Wednesday 09:00 - 10:00 - Building a Multi-Tasking ADF Application with Dynamic Regions and Dynamic Tabs 10:10 - 10:55 - Building Highly Reusable ADF Taskflows 12:30 - 13:30 - Design Patterns, Customization and Extensibility of Fusion Applications 14:25 - 15:10 - Continuous Integration with Hudson: What a year! 14:00 - 17:00 - Wednesday Wizardry with Fusion Middleware - Live application development demonstration with ADF, SOA Suite 15:20 - 16:05 - Adding Mobile and Web 2.0 UIs to Existing Applications - The Fusion Way  16:15 - 17:00 - Leveraging ADF for Building Complex Custom Applications

    Read the article

  • Improving Manageability of Virtual Environments

    - by Jeff Victor
    Boot Environments for Solaris 10 Branded Zones Until recently, Solaris 10 Branded Zones on Solaris 11 suffered one notable regression: Live Upgrade did not work. The individual packaging and patching tools work correctly, but the ability to upgrade Solaris while the production workload continued running did not exist. A recent Solaris 11 SRU (Solaris 11.1 SRU 6.4) restored most of that functionality, although with a slightly different concept, different commands, and without all of the feature details. This new method gives you the ability to create and manage multiple boot environments (BEs) for a Solaris 10 Branded Zone, and modify the active or any inactive BE, and to do so while the production workload continues to run. Background In case you are new to Solaris: Solaris includes a set of features that enables you to create a bootable Solaris image, called a Boot Environment (BE). This newly created image can be modified while the original BE is still running your workload(s). There are many benefits, including improved uptime and the ability to reboot into (or downgrade to) an older BE if a newer one has a problem. In Solaris 10 this set of features was named Live Upgrade. Solaris 11 applies the same basic concepts to the new packaging system (IPS) but there isn't a specific name for the feature set. The features are simply part of IPS. Solaris 11 Boot Environments are not discussed in this blog entry. Although a Solaris 10 system can have multiple BEs, until recently a Solaris 10 Branded Zone (BZ) in a Solaris 11 system did not have this ability. This limitation was addressed recently, and that enhancement is the subject of this blog entry. This new implementation uses two concepts. The first is the use of a ZFS clone for each BE. This makes it very easy to create a BE, or many BEs. This is a distinct advantage over the Live Upgrade feature set in Solaris 10, which had a practical limitation of two BEs on a system, when using UFS. The second new concept is a very simple mechanism to indicate the BE that should be booted: a ZFS property. The new ZFS property is named com.oracle.zones.solaris10:activebe (isn't that creative? ). It's important to note that the property is inherited from the original BE's file system to any BEs you create. In other words, all BEs in one zone have the same value for that property. When the (Solaris 11) global zone boots the Solaris 10 BZ, it boots the BE that has the name that is stored in the activebe property. Here is a quick summary of the actions you can use to manage these BEs: To create a BE: Create a ZFS clone of the zone's root dataset To activate a BE: Set the ZFS property of the root dataset to indicate the BE To add a package or patch to an inactive BE: Mount the inactive BE Add packages or patches to it Unmount the inactive BE To list the available BEs: Use the "zfs list" command. To destroy a BE: Use the "zfs destroy" command. Preparation Before you can use the new features, you will need a Solaris 10 BZ on a Solaris 11 system. You can use these three steps - on a real Solaris 11.1 server or in a VirtualBox guest running Solaris 11.1 - to create a Solaris 10 BZ. The Solaris 11.1 environment must be at SRU 6.4 or newer. Create a flash archive on the Solaris 10 system s10# flarcreate -n s10-system /net/zones/archives/s10-system.flar Configure the Solaris 10 BZ on the Solaris 11 system s11# zonecfg -z s10z Use 'create' to begin configuring a new zone. zonecfg:s10z create -t SYSsolaris10 zonecfg:s10z set zonepath=/zones/s10z zonecfg:s10z exit s11# zoneadm list -cv ID NAME STATUS PATH BRAND IP 0 global running / solaris shared - s10z configured /zones/s10z solaris10 excl Install the zone from the flash archive s11# zoneadm -z s10z install -a /net/zones/archives/s10-system.flar -p You can find more information about the migration of Solaris 10 environments to Solaris 10 Branded Zones in the documentation. The rest of this blog entry demonstrates the commands you can use to accomplish the aforementioned actions related to BEs. New features in action Note that the demonstration of the commands occurs in the Solaris 10 BZ, as indicated by the shell prompt "s10z# ". Many of these commands can be performed in the global zone instead, if you prefer. If you perform them in the global zone, you must change the ZFS file system names. Create The only complicated action is the creation of a BE. In the Solaris 10 BZ, create a new "boot environment" - a ZFS clone. You can assign any name to the final portion of the clone's name, as long as it meets the requirements for a ZFS file system name. s10z# zfs snapshot rpool/ROOT/zbe-0@snap s10z# zfs clone -o mountpoint=/ -o canmount=noauto rpool/ROOT/zbe-0@snap rpool/ROOT/newBE cannot mount 'rpool/ROOT/newBE' on '/': directory is not empty filesystem successfully created, but not mounted You can safely ignore that message: we already know that / is not empty! We have merely told ZFS that the default mountpoint for the clone is the root directory. List the available BEs and active BE Because each BE is represented by a clone of the rpool/ROOT dataset, listing the BEs is as simple as listing the clones. s10z# zfs list -r rpool/ROOT NAME USED AVAIL REFER MOUNTPOINT rpool/ROOT 3.55G 42.9G 31K legacy rpool/ROOT/zbe-0 1K 42.9G 3.55G / rpool/ROOT/newBE 3.55G 42.9G 3.55G / The output shows that two BEs exist. Their names are "zbe-0" and "newBE". You can tell Solaris that one particular BE should be used when the zone next boots by using a ZFS property. Its name is com.oracle.zones.solaris10:activebe. The value of that property is the name of the clone that contains the BE that should be booted. s10z# zfs get com.oracle.zones.solaris10:activebe rpool/ROOT NAME PROPERTY VALUE SOURCE rpool/ROOT com.oracle.zones.solaris10:activebe zbe-0 local Change the active BE When you want to change the BE that will be booted next time, you can just change the activebe property on the rpool/ROOT dataset. s10z# zfs get com.oracle.zones.solaris10:activebe rpool/ROOT NAME PROPERTY VALUE SOURCE rpool/ROOT com.oracle.zones.solaris10:activebe zbe-0 local s10z# zfs set com.oracle.zones.solaris10:activebe=newBE rpool/ROOT s10z# zfs get com.oracle.zones.solaris10:activebe rpool/ROOT NAME PROPERTY VALUE SOURCE rpool/ROOT com.oracle.zones.solaris10:activebe newBE local s10z# shutdown -y -g0 -i6 After the zone has rebooted: s10z# zfs get com.oracle.zones.solaris10:activebe rpool/ROOT rpool/ROOT com.oracle.zones.solaris10:activebe newBE local s10z# zfs mount rpool/ROOT/newBE / rpool/export /export rpool/export/home /export/home rpool /rpool Mount the original BE to see that it's still there. s10z# zfs mount -o mountpoint=/mnt rpool/ROOT/zbe-0 s10z# ls /mnt Desktop export platform Documents export.backup.20130607T214951Z proc S10Flar home rpool TT_DB kernel sbin bin lib system boot lost+found tmp cdrom mnt usr dev net var etc opt Patch an inactive BE At this point, you can modify the original BE. If you would prefer to modify the new BE, you can restore the original value to the activebe property and reboot, and then mount the new BE to /mnt (or another empty directory) and modify it. Let's mount the original BE so we can modify it. (The first command is only needed if you haven't already mounted that BE.) s10z# zfs mount -o mountpoint=/mnt rpool/ROOT/zbe-0 s10z# patchadd -R /mnt -M /var/sadm/spool 104945-02 Note that the typical usage will be: Create a BE Mount the new (inactive) BE Use the package and patch tools to update the new BE Unmount the new BE Reboot Delete an inactive BE ZFS clones are children of their parent file systems. In order to destroy the parent, you must first "promote" the child. This reverses the parent-child relationship. (For more information on this, see the documentation.) The original rpool/ROOT file system is the parent of the clones that you create as BEs. In order to destroy an earlier BE that is that parent of other BEs, you must first promote one of the child BEs to be the ZFS parent. Only then can you destroy the original BE. Fortunately, this is easier to do than to explain: s10z# zfs promote rpool/ROOT/newBE s10z# zfs destroy rpool/ROOT/zbe-0 s10z# zfs list -r rpool/ROOT NAME USED AVAIL REFER MOUNTPOINT rpool/ROOT 3.56G 269G 31K legacy rpool/ROOT/newBE 3.56G 269G 3.55G / Documentation This feature is so new, it is not yet described in the Solaris 11 documentation. However, MOS note 1558773.1 offers some details. Conclusion With this new feature, you can add and patch packages to boot environments of a Solaris 10 Branded Zone. This ability improves the manageability of these zones, and makes their use more practical. It also means that you can use the existing P2V tools with earlier Solaris 10 updates, and modify the environments after they become Solaris 10 Branded Zones.

    Read the article

  • Frank Ludolph's Last Day at Work

    - by mprove
    Hi Frank, today is your last day at Oracle. I cannot belief that retirement is an alternative to designing software and improving products for decades. I might figure it out myself in a couple of years. Our ways have crossed several times. And I am extremely thankful for that. I still remember my first session on an Apple Lisa. It must have been around 1985. I was still in school, and we were visiting the University of Hamburg to get some orientation on the departments. When I started I chose Informatics. And I suppose the Apple Lisa played a significant role in my decision. Is it fate that I later wrote about Apple Lisa? I’ve attended your presentation and public demo of the Lisa System at CHI ’98 in Los Angeles. Maybe a video still exists. I should look it up and publish it somewhere. You had also booth duty for Sun Microsystems – presenting HotJava Views, a user interface for a network computer. And you were handing out VHS tapes (!) of Starfire. I still have mine – but no player anymore. Then I joined Sun in 2002, and I guess I popped up in your office each time when I came to Santa Clara. The SEED mentoring program finally made it possible that we exchanged and discussed many ideas on the past and future of HCI. Dueling Interaction Models of Personal-Computing and Web-Computing at MEDICI 2007 is one of the results. But do you remember for instance also our jam session with Phil Clevenger on Hello World? Marvelous! I will miss you at Oracle. Enjoy your life and let’s stay in touch.Matthias

    Read the article

  • Brainless Backups

    - by Jesse
    I’m a software developer by trade which means to my friends and family I’m just a “computer guy”. It’s assumed that I know everything about every facet of computing from removing spyware to replacing hardware. I also can do all of this blindly over the phone or after hearing a five to ten word description of the problem over dinner ;-) In my position as CIO of my friends and families I’ve been in the unfortunate position of trying to recover music, pictures, or documents off of failed hard drives on more than one occasion. It’s not a great situation for anyone, and it’s always at these times that the importance of backups becomes so clear. Several months back a friend of mine found himself in this situation. The hard drive on his 8 year old laptop failed and took a good number of his digital photos with it. I think most folks can deal with losing some of their music and even some of their documents, but it really stings to lose pictures of past events and loved ones. After ordering a new laptop, my friend went out and bought an external hard drive so that he could start keeping a backup of his data. As fate would have it, several months later the drive in his new laptop failed and he learned the hard way that simply buying the external hard drive isn’t enough… you actually have to copy your stuff over every once in awhile! The importance of backup and recovery plans is (hopefully) well known in IT organizations. Well executed backup plans are in place, and hopefully the backup and recovery process is tested regularly. When you’re talking about users at home, however, the need for these backups is often understood far too late. Most typical users can’t be expected to remember to backup their data regularly and also don’t always have the know-how to setup automated backups. For my friends and family members in this situation I recommend tools like Dropbox, Carbonite, and Mozy. Here’s why I like them: They’re affordable: Dropbox and Mozy both have free offerings, though most people with lots of music and/or photos to backup will probably exceed the storage limitations of those free plans pretty quickly. Still, all three offer pretty affordable monthly or yearly plans. In my opinion, Carbonite’s unlimited storage plan for $50-$60 per year is the best value around. They’re easy to setup: Both Dropbox and Carbonite are very easy to get setup and start using. I’ve never used Mozy, but I imagine it’s similarly painless to get up and running. Backups are automatically “off-site”: A backup that is sitting on an external hard drive right next to your computer is great, but might not protect against flood damage, a power surge, or other disasters in that single location. These services exist “in the cloud” so to speak, helping mitigate those concerns. Granted, this kind of backup scheme requires some trust in the 3rd party to protect your data from both malicious people and disastrous events. This truly is a bit of a double edged sword, but I sleep well at night knowing that my data is being backed up and secured by a company made up of engineers that focus on the business of doing backups right. Backups are “brainless”: What I like most about services like these is that they work “automagically” in the background, watching for files to be updated and automatically backing up those changes. There’s no need to remember to plug in that external drive and copy your data over. Since starting to recommend these services to my friends and family I find myself wearing my “data recovery” hat far less often. The only way backups are effective for your standard computer user is if they’re completely automatic. Backups need to be brainless, or they just won’t work.

    Read the article

  • Ghost team foundation build controllers

    - by Martin Hinshelwood
    Quite often after an upgrade there are things left over. Most of the time they are easy to delete, but sometimes it takes a little effort. Even rarer are those times when something just will not go away no matter how much you try. We have had a ghost team build controller hanging around for a while now, and it had defeated my best efforts to get rid of it. The build controller was from our old TFS server from before our TFS 2010 beta 2 upgrade and was really starting to annoy me. Every time I try to delete it I get the message: Controller cannot be deleted because there are build in progress -Manage Build Controller dialog   Figure: Deleting a ghost controller does not always work. I ended up checking all of our 172 Team Projects for the build that was queued, but did not find anything. Jim Lamb pointed me to the “tbl_BuildQueue” table in the team Project Collection database and sure enough there was the nasty little beggar. Figure: The ghost build was easily spotted Adam Cogan asked me: “Why did you suspect this one?” Well, there are a number of things that led me to suspect it: QueueId is very low: Look at the other items, they are in the thousands not single digits ControllerId: I know there is only one legitimate controller, and I am assuming that 6 relates to “zzUnicorn” DefinitionId: This is a very low number and I looked it up in “tbl_BuildDefinition” and it did not exist QueueTime: As we did not upgrade to TFS 2010 until late 2009 a date of 2008 for a queued build is very suspect Status: A status of 2 means that it is still queued This build must have been queued long ago when we were using TFS 2008, probably a beta, and it never got cleaned up. As controllers are new in TFS 2010 it would have created the “zzUnicorn” controller to handle any build servers that already exist. I had previously deleted the Agent, but leaving the controller just looks untidy. Now that the ghost build has been identified there are two options: Delete the row I would not recommend ever deleting anything from the database to achieve something in TFS. It is really not supported. Set the Status to cancelled (Recommended) This is the best option as TFS will then clean it up itself So I set the Status of this build to 2 (cancelled) and sure enough it disappeared after a couple of minutes and I was then able to then delete the “zzUnicorn” controller. Figure: Almost completely clean Now all I have to do is get rid of that untidy “zzBunyip” agent, but that will require rewriting one of our build scripts which will have to wait for now.   Technorati Tags: ALM,TFBS,TFS 2010

    Read the article

  • OPN PartnerCasts com Stein Surlin

    - by Claudia Costa
      Desde de Abril 2009  Stein Surlin - Senior VP Alliances& Channels EMEA, tem realizado diversas  entrevistas com executivos da Oracle EMEA, em que se têm discutido temas acerca da organização Oracle EMEA, de como está alinhada em trabalhar em conjunto com os seus  parceiros e em apoiá-los na obtenção de novas oportunidades de negócio . Neste link poderá assistir aos OPN PartnerCasts. Aqui poderá encontrar as notícias mais actualizadas sobre a estratégia de parceiros da Oracle, oportunidades de negócio, estratégias de produtos e muito mais! Fique atento ao Portal OPN Partner Casts, mais podcasts estão a caminho!        

    Read the article

  • Repeat row headers after Page Break

    - by klaus.fabian
    The lead developer of the FO engine send me by chance an email about a REALLY nice feature I did not know about. Did you ever encounter a long table with merged cells, where the merged cell went on to the next page? While column headers are by default repeated on the next page, row headers are not. Tables with group-left column and pivot tables are prime examples where this problem occurs. I have seen reports where merged cells could go over multiple pages and you would need to back to find the row header on previous pages. The BI Publisher RTF templates have a special tag you can added to a merged cell to repeat the contents after each page break. You just need to add the following (wordy) tag to the next merged table cell: true Example: 2nd page of report before adding the tag 2nd page of report after adding the tag. Thought you might want to know. Klaus

    Read the article

  • SBUG Session: The Enterprise Cache

    - by EltonStoneman
    [Source: http://geekswithblogs.net/EltonStoneman] I did a session on "The Enterprise Cache" at the UK SOA/BPM User Group yesterday which generated some useful discussion. The proposal was for a dedicated caching layer which all app servers and service providers can hook into, sharing resources and common data. The architecture might end up like this: I'll update this post with a link to the slide deck once it's available. The next session will have Udi Dahan walking through nServiceBus, register on EventBrite if you want to come along. Synopsis Looked at the benefits and drawbacks of app-centric isolated caches, compared to an enterprise-wide shared cache running on dedicated nodes; Suggested issues and risks around caching including staleness of data, resource usage, performance and testing; Walked through a generic service cache implemented as a WCF behaviour – suitable for IIS- or BizTalk-hosted services - which I'll be releasing on CodePlex shortly; Listed common options for cache providers and their offerings. Discussion Cache usage. Different value propositions for utilising the cache: improved performance, isolation from underlying systems (e.g. service output caching can have a TTL large enough to cover downtime), reduced resource impact – CPU, memory, SQL and cost (e.g. caching results of paid-for services). Dedicated cache nodes. Preferred over in-host caching provided latency is acceptable. Depending on cache provider, can offer easy scalability and global replication so cache clients always use local nodes. Restriction of AppFabric Caching to Windows Server 2008 not viewed as a concern. Security. Limited security model in most cache providers. Options for securing cache content suggested as custom implementations. Obfuscating keys and serialized values may mean additional security is not needed. Depending on security requirements and architecture, can ensure cache servers only accessible to cache clients via IPsec. Staleness. Generally thought to be an overrated problem. Thinking in line with eventual consistency, that serving up stale data may not be a significant issue. Good technical arguments support this, although I suspect business users will be harder to persuade. Providers. Positive feedback for AppFabric Caching – speed, configurability and richness of the distributed model making it a good enterprise choice. .NET port of memcached well thought of for performance but lack of replication makes it less suitable for these shared scenarios. Replicated fork – repcached – untried and less active than memcached. NCache also well thought of, but Express version too limited for enterprise scenarios, and commercial versions look costly compared to AppFabric.

    Read the article

  • BizTalk Envelopes explained

    - by Robert Kokuti
    Recently I've been trying to get some order into an ESB-BizTalk pub/sub scenario, and decided to wrap the payload into standardized envelopes. I have used envelopes before in a 'light weight' fashion, and I found that they can be quite useful and powerful if used systematically. Here is what I learned. The Theory In my experience, Envelopes are often underutilised in a BizTalk solution, and quite often their full potential is not well understood. Here I try to simplify the theory behind the Envelopes within BizTalk.   Envelopes can be used to attach additional data to the ‘real’ data (payload). This additional data can contain all routing and processing information, and allows treating the business data as a ‘black box’, possibly compressed and/or encrypted etc. The point here is that the infrastructure does not need to know anything about the business data content, just as a post man does not need to know the letter within the envelope. BizTalk has built-in support for envelopes through the XMLDisassembler and XMLAssembler pipeline components (these are part of the XMLReceive and XMLSend default pipelines). These components, among other things, perform the following: XMLDisassembler Extracts the payload from the envelope into the Message Body Copies data from the envelope into the message context, as specified by the property schema(s) associated by the envelope schema. Typically, once the envelope is through the XMLDisassembler, the payload is submitted into the Messagebox, and the rest of the envelope data are copied into the context of the submitted message. The XMLDisassembler uses the Property Schemas, referenced by the Envelope Schema, to determine the name of the promoted Message Context element.   XMLAssembler Wraps the Message Body inside the specified envelope schema Populates the envelope values from the message context, as specified by the property schema(s) associated by the envelope schema. Notice that there are no requirements to use the receiving envelope schema when sending. The sent message can be wrapped within any suitable envelope, regardless whether the message was originally received within an envelope or not. However, by sharing Property Schemas between Envelopes, it is possible to pass values from the incoming envelope to the outgoing envelope via the Message Context. The Practice Creating the Envelope Add a new Schema to the BizTalk project:   Envelopes are defined as schemas, with the <Schema> Envelope property set to Yes, and the root node’s Body XPath property pointing to the node which contains the payload. Typically, you’d create an envelope structure similar to this: Click on the <Schema> node and set the Envelope property to Yes. Then, click on the Envelope node, and set the Body XPath property pointing to the ‘Body’ node:   The ‘Body’ node is a Child Element, and its Data Structure Type is set to xs:anyType.  This allows the Body node to carry any payload data. The XMLReceive pipeline will submit the data found in the Body node, while the XMLSend pipeline will copy the message into the Body node, before sending to the destination. Promoting Properties Once you defined the envelope, you may want to promote the envelope data (anything other than the Body) as Property Fields, in order to preserve their value in the message context. Anything not promoted will be lost when the XMLDisassembler extracts the payload from the Body. Typically, this means you promote everything in the Header node. Property promotion uses associated Property Schemas. These are special BizTalk schemas which have a flat field structure. Property Schemas define the name of the promoted values in the Message Context, by combining the Property Schema’s Namespace and the individual Field names. It is worth being systematic when it comes to naming your schemas, their namespace and type name. A coherent method will make your life easier when it comes to referencing the schemas during development, and managing subscriptions (filters) in BizTalk Administration. I developed a fairly coherent naming convention which I’ll probably share in another article. Because the property schema must be flat, I recommend creating one for each level in the envelope header hierarchy. Property schemas are very useful in passing data between incoming as outgoing envelopes. As I mentioned earlier, in/out envelopes do not have to be the same, but you can use the same property schema when you promote the outgoing envelope fields as you used for the incoming schema.  As you can reference many property schemas for field promotion, you can pick data from a variety of sources when you define your outgoing envelope. For example, the outgoing envelope can carry some of the incoming envelope’s promoted values, plus some values from the standard BizTalk message context, like the AdapterReceiveCompleteTime property from the BizTalk message-tracking properties. The values you promote for the outgoing envelope will be automatically populated from the Message Context by the XMLAssembler pipeline component. Using the Envelope Receiving Enveloped messages are automatically recognized by the XMLReceive pipeline, or any other custom pipeline which includes the XMLDisassembler component. The Body Path node will become the Message Body, while the rest of the envelope values will be added to the Message context, as defined by the Property Shemas referenced by the Envelope Schema. Sending The Send Port’s filter expression can use the promoted properties from the incoming envelope. If you want to enclose the sent message within an envelope, the Send Port XMLAssembler component must be configured with the fully qualified envelope name:   One way of obtaining the fully qualified envelope name is copy it off from the envelope schema property page: The full envelope schema name is constructed as <Name>, <Assembly> The outgoing envelope is populated by the XMLAssembler pipeline component. The Message Body is copied to the specified envelope’s Body Path node, while the rest of the envelope fields are populated from the Message Context, according to the Property Schemas associated with the Envelope Schema. That’s all for now, happy enveloping!

    Read the article

  • Gilda Garretón, a Java Developer and Parallelism Computing Researcher

    - by Yolande
    In a new interview titled “Gilda Garretón, a Java Developer and Parallelism Computing Research,” Garretón shares her first-hand experience developing with Java and Java 7 for very large-scale integration (VLSI) of computer-aided design (CAD). Garretón gives an insightful overview of how Java is contributing to the parallelism development and to the Electric VLSI Design Systems, an open source VLSI CAD application used as a research platform for new CAD algorithms as well as the research flow for hardware test chips.  Garretón considers that parallelism programming is hard and complex, yet important developments are taking place.  "With the addition of the concurrent package in Java SE 6 and the Fork/Join feature in Java SE 7, developers have a chance to rely more on existing frameworks and dedicate more time to the essence of their parallel algorithms." Read the full article here  

    Read the article

  • Cannot resolve the collation conflict between "Latin1_General_CI_AS" and "SQL_Latin1_General_CP1_CI_

    - by Patrick Olurotimi Ige
    I was writing a store proc for a report and i needed some data from another server so i added a linked server to connect to this new db server. when i do a select like below its all fine select a,b,c from Server.DatabaseName.dbo.table But when i use the table in a join i get the error "Cannot resolve the collation conflict between "Latin1_General_CI_AS" and "SQL_Latin1_General_CP1_CI_AS" in the equal to operation." I did check the collation set on the 2 databases and it was actually the same and had mo idea why i'm getting the error. I later found out that you could specifically tell it to use a COLLATE Just rewrite your join like this on a.name COLLATE Latin1_General_CI_AS = eaobjname Hope that helps and saves your precious time Patrick

    Read the article

  • New ACS Resell Portfolio for OPN Members

    - by rituchhibber
    Oracle Advanced Customer Support (ACS) Services is pleased to announce availability of the ACS Resell Portfolio to Oracle PartnerNetwork (OPN) members on June 28, 2012. The ACS Resell Portfolio is available to Gold level OPN members and above selling to end users with valid Oracle Premier Support/End User agreements, and in countries where ACS has a local in-country presence to support the partner business. ACS provides mission critical support services for complex IT environments to help maximize performance, achieve higher availability, and reduce risk. The ACS Resell Portfolio can be leveraged to reduce time to market and drive improved end user satisfaction. Including ACS services at point of license sale can maximize your success as an Oracle partner.       On July 10, 2012, Oracle ACS is hosting a 60-minute resell portfolio training session. Topics include: ACS Resell Portfolio objectives   Partner participation requirements ACS portfolio services enabled for partner resell ACS sales engagement and transaction processes Contracting requirements Attend the following session to hear how you can maximize your profit opportunities by including ACS services, which compliment your solutions with integrated Oracle advanced support technologies.      DIAL-IN INFORMATION Webconference July 10, 2012 4:00 PM CEST Webconference Session Number: 591 988 820 Session Password: ebh12345 International: 706.501.7506 US: 866.589.6202 Call ID: 95867658 Click here for a list of toll-free international numbers. Please contact [email protected] with any questions or visit the ACS website.

    Read the article

  • Silverlight beyond the basics

    - by Braulio Díez Botella
    Once I have learned the basics of Silverlight, I realized that there was still a lot to learn, architecture, patterns & practices, data access technologies… BUT… there’s plenty of material out there and few available time. I have compiled a set of articles/web casts / posts I found pretty useful for me, and defined a “learning roadmap”.   About the learning road map:    About the links: Basics MVVM Pattern: MSDN Magazine Basics RIA Services: RIA Services Intro RIA Services and Visual Studio 2010 MVVM + PRISM: MVVM + PRISM  MEF: MSDN Magazine RIA Services + MVVM: Mix10 RIA Services + MVVM + MEF: Shawn Wildermuth Series (1) Shawn Wildermuth Series (2) Shawn Wildermuth Series (3) Shawn Wildermuth Series (4) Some of them are based on Beta version of the products, but the core concepts are there and quite well explained. Please if you have other superb references add it on the comments section, hope to build a “version 2” of this post including or your feeback, thanks.

    Read the article

  • Enterprise Architecture IS (should not be) Arbitrary

    - by pat.shepherd
    I took a look at a blog entry today by Jordan Braunstein where he comments on another blog entry titled “Yes, “Enterprise Architecture is Relative BUT it is not Arbitrary.”  The blog makes some good points such as the following: Lock 10 architects in 10 separate rooms; provide them all an identical copy of the same business, technical, process, and system requirements; have them design an architecture under the same rules and perspectives; and I guarantee your result will be 10 different architectures of varying degrees. SOA Today: Enterprise Architecture IS Arbitrary Agreed, …to a degree….but less so if all 10 truly followed one of the widely accepted EA frameworks. My thinking is that EA frameworks all focus on getting the business goals/vision locked down first as the primary drivers for decisions made lower down the architecture stack.  Many people I talk to, know about frameworks such as TOGAF, FEA, etc. but seldom apply the tenants to the architecture at hand.  We all seem to want to get right into the Visio diagrams and boxes and arrows and connecting protocols and implementation details and lions and tigers and bears (Oh, my!) too early. If done properly the Business, Application and Information architectures are nailed down BEFORE any technological direction (SOA or otherwise) is set.  Those 3 layers and Governance (people and processes), IMHO, are layers that should not vary much as they have everything to do with understanding the business -- from which technological conclusions can later be drawn. I really like what he went on to say later in the post about the fact that architecture attempts to remove the amount of variance between the 10 different architect’s work.  That is the real heart of what EA is about; REMOVING THE ARBRITRARITY.

    Read the article

  • Slides and Links from SQL Azure session at BizSpark Azure Day in London

    - by Eric Nelson
    A big thanks to all who attended my two sessions on SQL Azure yesterday (29th March 2010). As promised, my slides and links from the session. SQL Azure Overview for Bizspark day View more presentations from Eric Nelson. Related Links: UK Azure Online Community – join today. UK Windows Azure Site Start working with Windows Azure SQL Azure maximum database size rises from 10GB to 50GB in June TCO and ROI calculator for Windows Azure SQL Azure Migration Wizard

    Read the article

  • Oracle OpenWorld Healthcare Integration Session Highlights Challenges & Solutions

    - by Bruce Tierney
    In today’s session co-presented by Steve Schenks, Integration Architect from Ascension Health and Oracle’s Sundar Shenbagam and Suresh Sharma (apparently your initials must be SS to present during this session), interesting insights in many different areas including Steve’s descriptions of the challenges with their previous environment: Disparate hardware and software is an issue common across healthcare and most other industries…Larry Ellison spoke on this topic during Sundays’ keynote address. In the last part of session, Suresh is planning to go over some of the best practices and lesson learned to implement successful healthcare applications and will discuss the different options to model Sequencing (FIF0) use cases (one of most common use cases in the provider market). The session was “Implementing Successful Healthcare Applications with Oracle SOA Suite” – Session # CON8546. For more information about this session, please contact Senior Principal Product Manager Suresh Sharma

    Read the article

  • Microsoft TechDays Canada 2010

    - by David Nudelman
    John Oxley announced that TechDays is returning to Canada in more cities then ever in 2010. Vancouver – September 14/15 at the Vancouver Convention Centre Edmonton – October 5/6 at the Shaw Conference Centre Toronto – October 27/28 at the Metro Toronto Convention Centre Halifax – November 2/3 at the World Trade & Convention Centre Ottawa – November 9/10 at the Hampton Inn & Conference Centre Montreal – November 23/24 at the Palais de Congres Winnipeg – December 7/8 at the Winnipeg Convention Centre Calgary – December 14/15 at the Calgary Stampede Get all the info here Regards, David Nudelman

    Read the article

  • Thread placement policies on NUMA systems - update

    - by Dave
    In a prior blog entry I noted that Solaris used a "maximum dispersal" placement policy to assign nascent threads to their initial processors. The general idea is that threads should be placed as far away from each other as possible in the resource topology in order to reduce resource contention between concurrently running threads. This policy assumes that resource contention -- pipelines, memory channel contention, destructive interference in the shared caches, etc -- will likely outweigh (a) any potential communication benefits we might achieve by packing our threads more densely onto a subset of the NUMA nodes, and (b) benefits of NUMA affinity between memory allocated by one thread and accessed by other threads. We want our threads spread widely over the system and not packed together. Conceptually, when placing a new thread, the kernel picks the least loaded node NUMA node (the node with lowest aggregate load average), and then the least loaded core on that node, etc. Furthermore, the kernel places threads onto resources -- sockets, cores, pipelines, etc -- without regard to the thread's process membership. That is, initial placement is process-agnostic. Keep reading, though. This description is incorrect. On Solaris 10 on a SPARC T5440 with 4 x T2+ NUMA nodes, if the system is otherwise unloaded and we launch a process that creates 20 compute-bound concurrent threads, then typically we'll see a perfect balance with 5 threads on each node. We see similar behavior on an 8-node x86 x4800 system, where each node has 8 cores and each core is 2-way hyperthreaded. So far so good; this behavior seems in agreement with the policy I described in the 1st paragraph. I recently tried the same experiment on a 4-node T4-4 running Solaris 11. Both the T5440 and T4-4 are 4-node systems that expose 256 logical thread contexts. To my surprise, all 20 threads were placed onto just one NUMA node while the other 3 nodes remained completely idle. I checked the usual suspects such as processor sets inadvertently left around by colleagues, processors left offline, and power management policies, but the system was configured normally. I then launched multiple concurrent instances of the process, and, interestingly, all the threads from the 1st process landed on one node, all the threads from the 2nd process landed on another node, and so on. This happened even if I interleaved thread creating between the processes, so I was relatively sure the effect didn't related to thread creation time, but rather that placement was a function of process membership. I this point I consulted the Solaris sources and talked with folks in the Solaris group. The new Solaris 11 behavior is intentional. The kernel is no longer using a simple maximum dispersal policy, and thread placement is process membership-aware. Now, even if other nodes are completely unloaded, the kernel will still try to pack new threads onto the home lgroup (socket) of the primordial thread until the load average of that node reaches 50%, after which it will pick the next least loaded node as the process's new favorite node for placement. On the T4-4 we have 64 logical thread contexts (strands) per socket (lgroup), so if we launch 48 concurrent threads we will find 32 placed on one node and 16 on some other node. If we launch 64 threads we'll find 32 and 32. That means we can end up with our threads clustered on a small subset of the nodes in a way that's quite different that what we've seen on Solaris 10. So we have a policy that allows process-aware packing but reverts to spreading threads onto other nodes if a node becomes too saturated. It turns out this policy was enabled in Solaris 10, but certain bugs suppressed the mixed packing/spreading behavior. There are configuration variables in /etc/system that allow us to dial the affinity between nascent threads and their primordial thread up and down: see lgrp_expand_proc_thresh, specifically. In the OpenSolaris source code the key routine is mpo_update_tunables(). This method reads the /etc/system variables and sets up some global variables that will subsequently be used by the dispatcher, which calls lgrp_choose() in lgrp.c to place nascent threads. Lgrp_expand_proc_thresh controls how loaded an lgroup must be before we'll consider homing a process's threads to another lgroup. Tune this value lower to have it spread your process's threads out more. To recap, the 'new' policy is as follows. Threads from the same process are packed onto a subset of the strands of a socket (50% for T-series). Once that socket reaches the 50% threshold the kernel then picks another preferred socket for that process. Threads from unrelated processes are spread across sockets. More precisely, different processes may have different preferred sockets (lgroups). Beware that I've simplified and elided details for the purposes of explication. The truth is in the code. Remarks: It's worth noting that initial thread placement is just that. If there's a gross imbalance between the load on different nodes then the kernel will migrate threads to achieve a better and more even distribution over the set of available nodes. Once a thread runs and gains some affinity for a node, however, it becomes "stickier" under the assumption that the thread has residual cache residency on that node, and that memory allocated by that thread resides on that node given the default "first-touch" page-level NUMA allocation policy. Exactly how the various policies interact and which have precedence under what circumstances could the topic of a future blog entry. The scheduler is work-conserving. The x4800 mentioned above is an interesting system. Each of the 8 sockets houses an Intel 7500-series processor. Each processor has 3 coherent QPI links and the system is arranged as a glueless 8-socket twisted ladder "mobius" topology. Nodes are either 1 or 2 hops distant over the QPI links. As an aside the mapping of logical CPUIDs to physical resources is rather interesting on Solaris/x4800. On SPARC/Solaris the CPUID layout is strictly geographic, with the highest order bits identifying the socket, the next lower bits identifying the core within that socket, following by the pipeline (if present) and finally the logical thread context ("strand") on the core. But on Solaris on the x4800 the CPUID layout is as follows. [6:6] identifies the hyperthread on a core; bits [5:3] identify the socket, or package in Intel terminology; bits [2:0] identify the core within a socket. Such low-level details should be of interest only if you're binding threads -- a bad idea, the kernel typically handles placement best -- or if you're writing NUMA-aware code that's aware of the ambient placement and makes decisions accordingly. Solaris introduced the so-called critical-threads mechanism, which is expressed by putting a thread into the FX scheduling class at priority 60. The critical-threads mechanism applies to placement on cores, not on sockets, however. That is, it's an intra-socket policy, not an inter-socket policy. Solaris 11 introduces the Power Aware Dispatcher (PAD) which packs threads instead of spreading them out in an attempt to be able to keep sockets or cores at lower power levels. Maximum dispersal may be good for performance but is anathema to power management. PAD is off by default, but power management polices constitute yet another confounding factor with respect to scheduling and dispatching. If your threads communicate heavily -- one thread reads cache lines last written by some other thread -- then the new dense packing policy may improve performance by reducing traffic on the coherent interconnect. On the other hand if your threads in your process communicate rarely, then it's possible the new packing policy might result on contention on shared computing resources. Unfortunately there's no simple litmus test that says whether packing or spreading is optimal in a given situation. The answer varies by system load, application, number of threads, and platform hardware characteristics. Currently we don't have the necessary tools and sensoria to decide at runtime, so we're reduced to an empirical approach where we run trials and try to decide on a placement policy. The situation is quite frustrating. Relatedly, it's often hard to determine just the right level of concurrency to optimize throughput. (Understanding constructive vs destructive interference in the shared caches would be a good start. We could augment the lines with a small tag field indicating which strand last installed or accessed a line. Given that, we could augment the CPU with performance counters for misses where a thread evicts a line it installed vs misses where a thread displaces a line installed by some other thread.)

    Read the article

< Previous Page | 172 173 174 175 176 177 178 179 180 181 182 183  | Next Page >