Platform LSF Version 9 Release 1.2 Release Notes GI13-3413-03 Platform LSF Version 9 Release 1.2 Release Notes GI13-3413-03 Note Before using this information and the product it supports, read the information in “Notices” on page 31. First edition This edition applies to version 9, release 1 of IBM Platform LSF (product number 5725G82) and to all subsequent releases and modifications until otherwise indicated in new editions. Significant changes or additions to the text and illustrations are indicated by a vertical line (|) to the left of the change. If you find an error in any Platform Computing documentation, or you have a suggestion for improving it, please let us know. Send your suggestions, comments and questions to the following email address: [email protected] Be sure include the publication title and order number, and, if applicable, the specific location of the information about which you have comments (for example, a page number or a browser URL). When you send information to IBM, you grant IBM a nonexclusive right to use or distribute the information in any way it believes appropriate without incurring any obligation to you. © Copyright IBM Corporation 1992, 2013. US Government Users Restricted Rights – Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM Corp. Contents Chapter 1. Release Notes for IBM Platform LSF . . . . . . . . . . . . 1 Learn more about IBM Platform LSF . . . . . . 1 We’d like to hear from you . . . . . . . . . 3 Requirements and compatibility . . . . . . . . 3 Installation and migration notes . . . . . . . . 6 Platform LSF editions . . . . . . . . . . . 7 What’s new in Platform LSF Version 9.1.2 . . . . 12 Known issues. . . . . . . . . . . . . . 27 Limitations . . . . . . . . . . . . . . 28 © Copyright IBM Corp. 1992, 2013 Bugs fixed . . . . . . . . . . . . . . . 28 Chapter 2. Platform LSF product packages. . . . . . . . . . . . . . 29 Downloading the Platform LSF product packages . 30 Notices . . . . . . . . . . . . . . 31 Trademarks . . . . . . Privacy policy considerations . . . . . . . . . . . . . . . 33 . 33 iii iv Release Notes for Platform LSF Chapter 1. Release Notes for IBM Platform LSF Version: 9.1.2 Release date: December 2013 Last modified: January 10, 2014 Support: www.ibm.com/support Learn more about IBM Platform LSF Information about IBM Platform LSF (Platform LSF or LSF) is available from the following sources: v IBM Platform Computing web site: www.ibm.com/systems/ technicalcomputing/platformcomputing v The LSF area of the IBM Support Portal: www.ibm.com/platformcomputing/ support.html v IBM Technical Computing Community on IBM Service Management Connect: www.dw.rtp.raleigh.ibm.com/developerworks/servicemanagement/tc/ v Platform LSF documentation Platform LSF documentation The Platform LSF documentation is contained in the LSF documentation packages: v lsf9.1.2_documentation.tar.Z v lsf9.1.2_documentation.zip You can extract and install these packages to any server on your system. Open the LSF Documentation Center by navigating to the location where you extracted the files and open index.html in any Web browser.The Documentation Center provides an overview of the organization of the LSF documentation. It also provides easy access to each document and quick links to frequently used commands and tasks. In addition to links to all documents, the Documentation Center provides full search capabilities within the documentation. You can perform keyword searches within a document or across the full documentation set. If you have installed IBM Platform Application Center (PAC), you can access and search the LSF documentation through the Help link in the user interface. Platform LSF documentation is also available in PDF format on the IBM Web site: v By searching the IBM Publications Center: www.ibm.com/e-business/linkweb/ publications/servlet/pbi.wss v By searching the IBM Support Portal: www.ibm.com/support v On the IBM Cluster Products Information Center: http:// publib.boulder.ibm.com/infocenter/clresctr/vxrx/index.jsp The documentation set for Platform LSF 9.1.2 includes the following PDF documents: v Administering IBM Platform LSF - SC27530202 © Copyright IBM Corp. 1992, 2013 1 v v v v v IBM Platform LSF IBM Platform LSF IBM Platform LSF Running Jobs with IBM Platform LSF v v v v v v Using IBM Platform LSF Advanced Edition - SC27532102 Using IBM Platform LSF on Windows - SC27531102 Using IBM Platform MultiCluster - SC27531002 Installing IBM Platform LSF on UNIX and Linux - SC27531402 Upgrading IBM Platform LSF on UNIX and Linux - SC27531502 Migrating IBM Platform LSF Version 7 to IBM Platform LSF Version 9.1.2 on UNIX and Linux - SC27531802 Installing IBM Platform LSF on Windows - SC27531602 Migrating IBM Platform LSF Version 7 to IBM Platform LSF Version 9.1.2 on Windows - SC27531702 IBM Platform LSF Security - SC27530302 Using IBM Platform LSF with IBM Rational ClearCase - SC27537700 Using the IBM Platform LSF blaunch Framework - SC27531302 v v v v v Foundations - SC27530402 Command Reference - SC27530502 Configuration Reference - SC27530602 IBM Platform LSF - SC27530702 Quick Reference - GC27530902 v IBM Platform LSF Programmer's Guide - SC27531202 Related documentation can be found in the following 9.1.2 documents: v v v v Using IBM Platform License Scheduler - SC27530802 Release Notes for IBM Platform License Scheduler - GI13341401 Using IBM Platform Dynamic Cluster - SC27532002 Release Notes for IBM Platform Dynamic Cluster - GI13341702 v IBM Platform MPI User's Guide - SC27475801 v Release Notes for IBM Platform MPI: Linux - GI13189601 v Release Notes for IBM Platform MPI: Windows - GI13189701 IBM Technical Computing community Connect. Learn. Share. Collaborate and network with the IBM Platform Computing experts at the IBM Technical Computing community. Access the Technical Computing community on IBM Service Management Connect at http://www.ibm.com/developerworks/servicemanagement/tc/. Join today! Service Management Connect is a group of technical communities for Integrated Service Management (ISM) professionals. Use Service Management Connect the following ways: v Connect to become involved with an ongoing, open engagement among other users, system professionals, and IBM developers of Platform Computing products. v Learn about IBM Technical Computing products on blogs and wikis, and benefit from the expertise and experience of others. v Share your experience in wikis and forums to collaborate with the broader Technical Computing user community. 2 Release Notes for Platform LSF We’d like to hear from you Contact IBM or your LSF vendor for technical support. Or go to the IBM Support Portal: www.ibm.com/support If you find an error in any Platform Computing documentation, or you have a suggestion for improving it, please let us know. Send your suggestions, comments and questions to the following email address: [email protected] Be sure include the publication title and order number, and, if applicable, the specific location of the information about which you have comments (for example, a page number or a browser URL). When you send information to IBM, you grant IBM a nonexclusive right to use or distribute the information in any way it believes appropriate without incurring any obligation to you. Requirements and compatibility The following sections detail requirements and compatibility for version 9.1.2 of Platform LSF. System requirements v IBM AIX 6.x and 7.x on IBM Power 6/7 v Linux x64 Kernel 2.6 and 3.x on x86_64 v Linux Kernel 2.6 and 3.x on IBM Power 6/7 v HP UX B.11.31 (64-bit) on HP 9000 Servers (PA-RISC) v HP UX B.11.31 (IA64) on HP Integrity Servers (Itanium2) v Solaris 10 and 11 on Sparc Solaris 10 and 11 on x86-64 zLinux Kernel 2.6, glibc 2.4 SLES 10 on S390x-64 Cray XE6, XT6, XC-30, Linux Kernel 2.6, glibc 2.3 on x86_64 ARMv7 Kernel 3.6 glibc 2.15 (LSF slave host only) Apple Mac OS 10.x (LSF slave host only) Windows 2003 SP1 and 2, XP SP1 and 2, 2008 x86, 7 X86, and 8 x86 on x86/x86_64 v Windows XP SP1 x64, 2003 SP1/2, 2003 CCE SP1/SP2, 2008 X64, 7 X64, 2008 R2 X64, HPC server 2008, 2012 X64 on x86_64 v v v v v v For detailed LSF system support information, refer to the Compatibility Table at: http://www.ibm.com/systems/technicalcomputing/platformcomputing/products/ lsf/ Master host selection To achieve the highest degree of performance and scalability, use a powerful master host. There is no minimum CPU requirement. For the platforms on which LSF is supported, any host with sufficient physical memory can run LSF as master host. Chapter 1. Release Notes for IBM Platform LSF 3 Swap space is normally configured as twice the physical memory. LSF daemons use about 40 MB of memory when no jobs are running. Active jobs consume most of the memory LSF requires. Recommended server CPU Cluster size Active jobs Minimum recommended memory Small (<100 hosts) 1,000 1 GB any server CPU 10,000 2 GB recent server CPU 10,000 4 GB multi-core CPU (2 cores) 50,000 8 GB multi-core CPU (4 cores) 50,000 16 GB multi-core CPU (4 cores) 500,000 32 GB multi-core CPU (8 cores) Medium (100-1000 hosts) Large (>1000 hosts) (Intel, AMD, or equivalent) Server host compatibility Important: To take full advantage of all new features introduced in the latest release of Platform LSF, you must upgrade all hosts in your cluster. Platform LSF 7.x, 8.0.x, 8.3, and 9.1.x, servers are compatible with Platform LSF 9.1.2 master hosts. All LSF 7.x, 8.0.x, 8.3, and 9.1.x features are supported by Platform LSF 9.1.2 master hosts. LSF Family product compatibility IBM Platform RTM Customers can use IBM Platform RTM (RTM) 8.3 or 9.1.x to collect data from Platform LSF 9.1.2 clusters. When adding the cluster, select 'Poller for LSF 8' or 'Poller for LSF 9.1'. IBM Platform License Scheduler IBM Platform License Scheduler (License Scheduler) 8.3 and 9.1.x are compatible with Platform LSF 9.1.2. IBM Platform Analytics IBM Platform Analytics (Analytics) 8.3 and 9.1.x are compatible with Platform LSF 9.1.2 after the following manual configuration: To have Analytics 8.3 or 9.1.x collect data from Platform LSF 9.1.2 clusters: 1. Set the following parameters in lsb.params: v ENABLE_EVENT_STREAM=Y v ALLOW_EVENT_TYPE="JOB_NEW JOB_FINISH2 JOB_STARTLIMIT JOB_STATUS2 JOB_PENDING_REASONS" v RUNTIME_LOG_INTERVAL=10 2. Copy elim.coreutil to LSF: cp ANALYTICS_TOP/elim/os_type/elim.coreutil $LSF_SERVERDIR 3. In lsf.shared, create the following: 4 Release Notes for Platform LSF Begin Resource RESOURCENAME TYPE CORE_UTIL String End Resource INTERVAL INCREASING DESCRIPTION 300 () (Core Utilization) 4. In lsf.cluster.cluster_name, create the following: Begin ResourceMap RESOURCENAME LOCATION CORE_UTIL [default] End ResourceMap 5. 6. 7. 8. 9. 10. 11. 12. Restart all LSF daemons. Configure user group and host group. Run lsid and check the output. Install Platform Analytics 8.3 with COLLECTED_DATA_TYPE=LSF. Check perf.conf to see LSF_VERSION. Restart the Platform loader controller (plc). Check the log files and table data to make sure there are no errors. Change all the LSF related data loader intervals to 120 seconds, and run for one day. Check the plc and data loader log files to make sure there are no errors. IBM Platform Application Center IBM Platform Application Center (PAC) 8.3 and higher versions are compatible with Platform LSF 9.1.x after the following manual configuration. If you are using PAC 8.3 with LSF 9.1.x, $PAC_TOP/perf/lsf/8.3 must be renamed to $PAC_TOP/perf/lsf/9.1 For example: mv /opt/pac/perf/lsf/8.3 /opt/pac/perf/lsf/9.1 API compatibility To take full advantage of new Platform LSF 9.1.2 features, recompile your existing Platform LSF applications with Platform LSF 9.1.2. Applications need to be rebuilt if they use APIs that have changed in Platform LSF 9.1.2. New and changed Platform LSF APIs The following APIs or data structures have changed or are new for LSF Version 9.1.2: v struct queueInfoEnt v struct submit_ext v struct hRusage v struct jobInfoEnt v struct jobNewLog v struct jobModLog v struct jobFinishLog v struct jobFinish2Log v struct jobStatusLog Chapter 1. Release Notes for IBM Platform LSF 5 v struct sbdJobStatusLog v struct sbdUnreportedStatusLog For detailed information about APIs changed or created for LSF 9.1.2, refer to the IBM Platform LSF 9.1.2 API Reference. Third party APIs The following third party APIs have been tested and supported for this release: v DRMAA LSF API v 1.1.1 v PERL LSF API v1.0 v Python LSF API v1.0 with LSF 9 Packages will be available at www.github.com. For more information on using third party APIs with LSF 9.1.2 look on the Technical Computing community on IBM Service Management Connect at http://www.ibm.com/developerworks/servicemanagement/tc/. Installation and migration notes Upgrade Platform LSF on UNIX and Linux Follow the steps in Upgrading IBM Platform LSF on UNIX and Linux (lsf_upgrade_unix.pdf) to run lsfinstall to upgrade LSF: v Upgrade a pre-LSF Version 7 UNIX or Linux cluster to Platform LSF 9.1.x v Upgrade an LSF Version 7 Update 2 or higher to Platform LSF 9.1.x Important: DO NOT use the UNIX and Linux upgrade steps to migrate an existing LSF Version 7 or LSF 7 Update 1 cluster to LSF 9.1.x. Follow the manual steps in the document Migrating to Platform LSF Version 9.1.x on UNIX and Linux to migrate an existing LSF Version 7 or LSF 7 Update 1 cluster to LSF 9.1.x on UNIX and Linux. Migrate LSF Version 7 and Version 7 Update 1 cluster to LSF 9.1.x on UNIX and Linux Follow the steps in Migrating to Platform LSF Version 9.1.2 on UNIX and Linux (lsf_migrate_unix.pdf) to migrate an existing LSF 7 or LSF 7 Update 1 cluster: v Migrate an existing LSF Version 7 cluster to LSF 9.1.2 on UNIX and Linux v Migrate an existing LSF Version 7 Update 1 cluster to LSF 9.1.x on UNIX and Linux Note: To migrate an LSF 7 Update 2 or higher cluster to LSF 9.1.x follow the steps in Upgrading IBM Platform LSF on UNIX and Linux. Migrate an LSF Version 7 or higher cluster to LSF 9.1.x on Windows To migrate an existing LSF 7 Windows cluster to Platform LSF 9.1.2 on Windows, follow the steps in Migrating IBM Platform LSF Version 7 to IBM Platform LSF Version 9.1.2 on Windows. 6 Release Notes for Platform LSF Note: If you want to migrate a pre-version 7 cluster to LSF 9.1.2, you must first migrate the cluster to LSF Version 7. Platform LSF editions Platform LSF comes in three editions: Advanced, Standard, and Express. LSF Advanced Edition Platform LSF Advanced Edition is architected to support extreme scalability and throughput. LSF Advanced Edition may provide greater than three times more scalability than earlier versions of LSF, enabling you to consolidate your compute resources to achieve maximum flexibility and utilization. LSF Advanced Edition has been tested on configurations up to 18,000 nodes and 160,000 cores running high-throughput workloads of 160,000 concurrent short jobs with 2,000,000 pending jobs. These are not hard scalability or performance limits. LSF/XL feature LSF Advanced Edition includes advanced features such as LSF/XL and multithreaded resource requirement matching. IBM Platform LSF Session Scheduler (Session Scheduler) is also included with LSF Advanced Edition. The LSF/XL feature of LSF Advanced Edition is a new architecture to address long-term scalability and performance demands for extreme workloads. It is based on IBM Platform MultiCluster (Platform MultiCluster) technology, but LSF/XL is not a Platform MultiCluster mode. It is designed for a single data center. LSF Standard Edition LSF Standard Edition is designed for grid environments and optimized for complex workloads and user grouping structures. LSF Standard Edition contains full functionality of LSF including functionality for Platform MultiCluster, floating clients and Platform LSF Make. Session Scheduler is available as an add-on component. There are no performance or scalability restrictions. LSF Standard Edition is subject to the following usage constraints: v LSF Standard Edition has been tested on clusters up to 6,000 nodes and 60,000 cores running high-throughput workloads of 60,000 concurrent short jobs with 250,000 pending jobs. These are not hard scalability or performance limits. Higher node or core counts can be achieved with a lower volume of jobs such as parallel HPC workloads where cluster sizes of 75,000 to 100,000 cores are supported. Higher core counts are achievable with additional tuning and configuration. v For high-throughput workloads, the overall system performance depends on the processing power, I/O capacity, and memory of the scheduling node. For very large clusters, seek configuration assistance from IBM. LSF Express Edition (Linux only) LSF Express Edition is a solution for Linux customers with very simple scheduling requirements and simple fairshare setup. Smaller clusters typically have a mix of sequential and parallel work as opposed to huge volumes of jobs. For this reason, several performance enhancements and complex scheduling policies designed for large-scale clusters are not applicable to LSF Express Edition clusters. Session Scheduler is available as an add-on component. Chapter 1. Release Notes for IBM Platform LSF 7 Out of box configuration for LSF Express Edition is optimized for smaller clusters with basic scheduling requirements. An LSF Express cluster can have a maximum of 200 server hosts and 200 static client hosts. LSF Express Edition is subject to the following restrictions: v LSF Express Edition is supported only on x86_64 Linux. v LSF Express Edition, LSF Standard Edition, or LSF Advanced Edition cannot coexist in the same cluster. v Master candidate hosts defined in the LSF_MASTER_LIST in lsf.conf must exist within the first 200 server hosts in the configuration file. Additional hosts over the limit will not be loaded. The following LSF Standard Edition features are supported when you upgrade from LSF Express Edition to LSF Standard Edition: v v v v v v Job groups Live reconfiguration Delegation of administrator rights Resizable jobs Chunk jobs Floating clients v LSF Make v Resource preemption v Cross-queue user-based fairshare v v v v Goal-oriented SLA-driven scheduling Guaranteed resource pools Slot-based scheduling Fairshare scheduling (queue and user fairshare are enabled with a fairshare tree depth of 2 levels in Express and more than two levels in Standard) v Parallel job scheduling (PAM/PJL is supported. blaunch is supported with IBM Platform MPI by default) v Parallel mbatchd restart (badmin mbdrestart -p) v Rapid detection of host failure v v v v v v v Fast dispatching Alternative resource requirement bjobs shows all levels of resource requirement Interaction of select[] and rusage[] Process tracking/short jobs Platform MultiCluster features Multithreaded batch query LSF integration with IBM Parallel Environment Runtime Edition Memory / CPU enforcement for Linux Cgroup Job information security (access control level) Energy aware scheduling (CPU frequency control, automatic CPU frequency selection, power state management) v Global same for compound resource requirements v Affinity resource preemption v Host based pre- and post-execution processing v v v v 8 Release Notes for Platform LSF Platform product support with LSF Express Edition The following IBM Platform products are supported in LSF Express Edition: v IBM Platform RTM v IBM Platform Application Center v IBM Platform License Scheduler The following IBM Platform products are not supported in LSF Express Edition: v IBM Platform Analytics v IBM Platform Process Manager Default configuration for LSF Express Edition The following table lists the configuration enforced in LSF Express Edition: Parameter Setting Description RESIZABLE_JOBS in lsb.applications N If enabled, all jobs belonging to the application will be auto resizable. EXIT_RATE in lsb.hosts Not defined Specifies a threshold for exited jobs. BJOBS_RES_REQ_DISPLAY in lsb.params None Controls how many levels of resource requirements bjobs –l will display. CONDENSE_PENDING_REASONS in lsb.params N Condenses all host-based pending reasons into one generic pending reason. DEFAULT_JOBGROUP in lsb.params Disabled The name of the default job group. EADMIN_TRIGGER_DURATION in lsb.params 1 minute Defines how often LSF_SERVERDIR/eadmin is invoked once a job exception is detected. Used in conjunction with job exception handling parameters JOB_IDLE, JOB_OVERRUN, and JOB_UNDERRUN in lsb.queues. ENABLE_DEFAULT_EGO_SLA in lsb.params Not defined The name of the default service class or EGO consumer name for EGO-enabled SLA scheduling. EVALUATE_JOB_DEPENDENCY in lsb.params Unlimited Sets the maximum number of job dependencies mbatchd evaluates in one scheduling cycle. GLOBAL_EXIT_RATE in lsb.params 2147483647 Specifies a cluster-wide threshold for exited jobs JOB_POSITION_CONTROL_BY_ADMIN in lsb.params Disabled Allows LSF administrators to control whether users can use btop and bbot to move jobs to the top and bottom of queues. LSB_SYNC_HOST_STAT_FROM_LIM in lsb.params N Improves the speed with which mbatchd obtains host status, and therefore the speed with which LSF reschedules rerunnable jobs. This parameter is most useful for a large clusters, so it is disabled for LSF Express Edition. Chapter 1. Release Notes for IBM Platform LSF 9 Parameter Setting Description MAX_CONCURRENT_QUERY in lsb.params 100 Controls the maximum number of concurrent query commands. MAX_INFO_DIRS in lsb.params Disabled The number of subdirectories under the LSB_SHAREDIR/cluster_name/ logdir/info directory. MAX_JOBID in lsb.params 999999 The job ID limit. The job ID limit is the highest job ID that LSF will ever assign, and also the maximum number of jobs in the system. MAX_JOB_NUM in lsb.params 1000 The maximum number of finished jobs whose events are to be stored in lsb.events. MIN_SWITCH_PERIOD in lsb.params Disabled The minimum period in seconds between event log switches. MBD_QUERY_CPUS in lsb.params Disabled Specifies the master host CPUs on which mbatchd child query processes can run (hard CPU affinity). NO_PREEMPT_INTERVAL in lsb.params 0 Prevents preemption of jobs for the specified number of minutes of uninterrupted run time, where minutes is wall-clock time, not normalized time. NO_PREEMPT_RUN_TIME in lsb.params -1 (not defined) Prevents preemption of jobs that have been running for the specified number of minutes or the specified percentage of the estimated run time or run limit. PREEMPTABLE_RESOURCES in lsb.params Not defined Enables preemption for resources (in addition to slots) when preemptive scheduling is enabled (has no effect if queue preemption is not enabled) and specifies the resources that will be preemptable. PREEMPT_FOR in lsb.params 0 If preemptive scheduling is enabled, this parameter is used to disregard suspended jobs when determining if a job slot limit is exceeded, to preempt jobs with the shortest running time, and to optimize preemption of parallel jobs. SCHED_METRIC_ENABLE in lsb.params N Enables scheduler performance metric collection. SCHED_METRIC_SAMPLE_PERIOD in lsb.params Disabled Performance metric sampling period. SCHEDULER_THREADS in lsb.params 0 Sets the number of threads the scheduler uses to evaluate resource requirements. 10 Release Notes for Platform LSF Parameter Setting Description DISPATCH_BY_QUEUE in lsb.queues N Increases queue responsiveness. The scheduling decision for the specified queue will be published without waiting for the whole scheduling session to finish. The scheduling decision for the jobs in the specified queue is final and these jobs cannot be preempted within the same scheduling cycle. LSB_JOBID_DISP_LENGTH in lsf.conf Not defined By default, LSF commands bjobs and bhist display job IDs with a maximum length of 7 characters. Job IDs greater than 9999999 are truncated on the left. When LSB_JOBID_DISP_LENGTH=10, the width of the JOBID column in bjobs and bhist increases to 10 characters. LSB_FORK_JOB_REQUEST in lsf.conf N Improves mbatchd response time after mbatchd is restarted (including parallel restart) and has finished replaying events. LSB_MAX_JOB_DISPATCH_PER_SESSION in lsf.conf 300 Defines the maximum number of jobs that mbatchd can dispatch during one job scheduling session. LSF_PROCESS_TRACKING in lsf.conf N Tracks processes based on job control functions such as termination, suspension, resume and other signaling, on Linux systems which support cgroups' freezer subsystem. LSB_QUERY_ENH in lsf.conf N Extends multithreaded query support to batch query requests (in addition to bjobs query requests). In addition, the mbatchd system query monitoring mechanism starts automatically instead of being triggered by a query request. This ensures a consistent query response time within the system. Enables a new default setting for min_refresh_time in MBD_REFRESH_TIME (lsb.params). LSB_QUERY_PORT in lsf.conf Disabled Increases mbatchd performance when using the bjobs command on busy clusters with many jobs and frequent query request. LSF_LINUX_CGROUP_ACCT in lsf.conf N Tracks processes based on CPU and memory accounting for Linux systems that support cgroup's memory and cpuacct subsystems. IBM Platform entitlement files Entitlement files are used for determining which edition of the product is enabled. The following entitlement files are packaged for LSF: Chapter 1. Release Notes for IBM Platform LSF 11 v LSF Standard Edition: platform_lsf_std_entitlement.dat v LSF Express Edition: platform_lsf_exp_entitlement.dat v LSF Advanced Edition: platform_lsf_adv_entitlement.dat The entitlement file for the edition you use is installed as LSF_TOP/conf/ lsf.entitlement. If you have installed LSF Express Edition, you can upgrade later to LSF Standard Edition or LSF Advanced Edition to take advantage of the additional functionality. Simply reinstall the cluster with the LSF Standard entitlement file (platform_lsf_std_entitlement.dat) or the LSF Advanced entitlement file (platform_lsf_adv_entitlement.dat). You can also manually upgrade from LSF Express Edition to Standard Edition or Advanced Edition. Get the LSF Standard or Advanced Edition entitlement file, copy it to LSF_TOP/conf/lsf.entitlement and restart you cluster. The new entitlement enables the additional functionality of LSF Standard Edition, but you may need to manually change some of the default LSF Express configuration parameters to use the LSF Standard or Advanced features. To take advantage of LSF SLA features in LSF Standard Edition, copy LSF_TOP/LSF_VERSION/install/conf_tmpl/lsf_standard/lsb.serviceclasses into LSF_TOP/conf/lsbatch/LSF_CLUSTERNAME/configdir/. Once LSF is installed and running, run the lsid command to see which edition of LSF is enabled. What’s new in Platform LSF Version 9.1.2 New and changed behavior Job information security LSF has features for controlling job security. This allows you to set the access control level to jobs by regular users and administrators. This is useful for large environments where many groups may share the same cluster and it may be a security threat to allow some users to view job details and summary information. With access control levels configured, you may prevent users (including user group, queue, and cluster administrators) from viewing other user’s job information through LSF commands (such as bjobs, bjdepinfo, bread, bstatus, bhist, and bacct). There are two kinds of job information which will be viewed by users: v Summary Information: Obtained from bjobs with options other than -l, such as -aps, -fwd, -p, -ss, -sum, -W, -WF, -WP, -WL, etc. v Detail Information: Obtained from bjobs -l, bjobs -UF, bjobs -N, bjdepinfo, bread, and bstatus. There are three parameters available in lsb.params that allow you to control access to job information: SECURE_JOB_INFO_LEVEL, ENABLE_JOB_INFO_BY_ADMIN_ROLE, and SECURE_INFODIR_USER_ACCESS. 12 Release Notes for Platform LSF The parameter SECURE_JOB_INFO_LEVEL in lsb.params allows you to define an access control level for all users (including user group, queue, and cluster administrators). A value between 0 and 4 is defined, with 0 being no security and 4 being the highest security. When a user enters one of the commands to see job information (bjobs, bjdepinfo, bread, or bstatus), the SECURE_JOB_INFO_LEVEL controls whether they see: v Just their own jobs’ information. (level 4) v Their own jobs and summary information from jobs in the same user group. (level 3) v Their own jobs, summary and detail information from jobs in the same user group. (level 2) v Their own jobs, summary and detail information from jobs in the same user group, and summary information from jobs outside their user group. (level 1) v Summary and detail job information for all jobs. (level 0) By default, an administrator’s access to job details is determined by the setting of SECURE_JOB_INFO_LEVEL, the same as a regular user. The parameter ENABLE_JOB_INFO_BY_ADMIN_ROLE in lsb.params allows you to enable the user group, queue, and cluster administrators the right to access job detail information for jobs in the user group, queue, and clusters they manage, even when the administrator has no right based on the configuration of SECURE_JOB_INFO_LEVEL. Note: This does not apply to the primary administrator or root. The parameter SECURE_INFODIR_USER_ACCESS in lsb.params allows you to control whether regular and administrator users (except the primary admin) can see other user’s jobs when using the bhist or bacct command. If enabled (defined as Y), regular users and administrators can view only their own job information when using the bhist or bacct command. LSB_SHAREDIR/cluster/ logdir is readable only by the Primary Administrator. When disabled (defined as N), access to read LSB_SHAREDIR/cluster/logdir returns to default after an mbatchd restart or reconfig. Attention: Requirements An upgrade to LSF 9.1.2 is required for this feature. After enabling this feature, you must setuid of the LSF primary administrator for bhist and bacct binary under LSF_BINDIR. bhist and bacct will call mbatchd to check whether the parameter is set or not when you have setuid for bhist and bacct. Note: When job information security is enabled, pre-LSF 9.1 bjobs and bjdepinfo commands will be rejected no matter who issues them because mbatchd cannot get the command user name. A "No job found" message will be returned. Note: Some batch commands (bkill, bstop, bresume, bchkpnt, bmig, brequeue, bswitch) which use job query API will be affected by ACL (job security) feature. Chapter 1. Release Notes for IBM Platform LSF 13 Global Same for compound resource requirements A "global same" has been introduced as part of a job’s resource requirement expression, to take effect over multiple component subexpressions of a compound resource requirement string. This new functionality can be used in any compound resource requirement, regardless of whether specified at the job, application, or queue level. The "same" requirement can be used within a resource requirement expression of a parallel job, to ensure that all hosts allocated to the job share some common attributes. For example, one might submit a parallel job as: bsub -n 128 -R "same[type]" ./a.out In this case, LSF will allocate 128 slots to the job, all on hosts of the same type. The "same" requirement also works for user defined resources. For example, you might configure a string resource called "rack", and set the value of this resource on each host to the ID of the rack that contains the host. Then, a job may be submitted as: bsub -n 128 -R "same[rack]" ./a.out In this case, LSF will allocate 128 slots to this job, all on hosts within the same rack. Compound resource requirements allow a single job to have different resource requirements on different parts of its allocation. This is useful in clusters, for example, where some hosts are dedicated to IO functions, while other hosts are dedicated to compute. Some jobs may require amounts of each type, such as in the following example: bsub -R "16*{type=io} + 128*{type=compute}" ./a.out This job requests 16 slots on IO hosts, and 128 slots on compute hosts. Previously, LSF had the limitation that the "same" string could only be specified within the simple subexpressions of a compound resource requirement. The "same" would take effect on the set of hosts allocated for each subexpression, but not for the allocation as a whole. In the above job submission, it would not be possible to use the "same" requirement to ensure that all slots come from a single rack. In this release, we remove this limitation. Now, the "same" requirement can be applied to multiple subexpressions of a compound resource requirement expression: bsub -R "{16*{type=io} + 128*{type=compute}} same[rack] " ./a.out This job requests 16 slots on IO hosts, and 128 slots on compute slots. All slots allocated to the job must come from hosts the same rack. Installer configuration templates and initial settings When installing LSF Standard Edition on UNIX or Linux, you may select a configuration template that specifies initial configuration parameter values. This 14 Release Notes for Platform LSF allows you to specify an initial setup that is appropriate for the specific type of cluster you are installing, depending on its purpose. Note: These configuration templates are not available with LSF Advanced Edition. To select a configuration template, edit install.config and uncomment the CONFIGURATION_TEMPLATE parameter: CONFIGURATION_TEMPLATE="DEFAULT" | "PARALLEL" | "HIGH_THROUGHPUT" The following are the valid values for this parameter: DEFAULT This template should be used for clusters with mixed workload. This configuration can serve different types of workload with good performance, but is not specifically tuned for a particular type of cluster. PARALLEL This template provides extra support for large parallel jobs. This configuration is designed for long running parallel jobs, and should not be used for clusters that mainly run short jobs due to the longer reporting time for each job. HIGH_THROUGHPUT This template is designed to be used for clusters that mainly run short jobs, where over 80% of jobs finish within one minute. This high turnover rate requires LSF to be more responsive and fast acting. However, this configuration will consume more resources as the daemons become busier. If you do not select a configuration template, the DEFAULT configuration template is selected by default. The installer uses the DEFAULT configuration template when installing LSF Standard Edition on Windows. Note: Do not specify CONFIGURATION_TEMPLATE for LSF Express Edition and Advanced Edition. These editions have their own default configuration templates for all installations. The installer specifies the following initial configuration file parameter values based on the selected configuration template: v DEFAULT – lsf.conf: DAEMON_SHUTDOWN_DELAY=180 LSF_LINUX_CGROUP_ACCT=Y LSF_PROCESS_TRACKING=Y – lsb.params: JOB_DEP_LAST_SUB=1 JOB_SCHEDULING_INTERVAL=1 MAX_JOB_NUM=10000 NEWJOB_REFRESH=Y SBD_SLEEP_TIME=7 v PARALLEL – lsf.conf: Chapter 1. Release Notes for IBM Platform LSF 15 LSB_SHORT_HOSTLIST=1 LSF_LINUX_CGROUP_ACCT=Y LSF_PROCESS_TRACKING=Y LSF_ENABLE_EXTSCHEDULER=Y LSF_HPC_EXTENSIONS="CUMULATIVE_RUSAGE LSB_HCLOSE_BY_RES SHORT_EVENTFILE" Refer to the Enable LSF HPC Features section for a full description. – lsb.params: JOB_DEP_LAST_SUB=1 JOB_SCHEDULING_INTERVAL=1 NEWJOB_REFRESH=Y v HIGH_THROUGHPUT – lsf.conf: LSB_MAX_PACK_JOBS=300 LSB_SHORT_HOSTLIST=1 – lsb.params: CONDENSE_PENDING_REASONS=Y JOB_SCHEDULING_INTERVAL=50ms MAX_INFO_DIRS=500 MAX_JOB_ARRAY_SIZE=10000 MAX_JOB_NUM=100000 MIN_SWITCH_PERIOD=1800 NEWJOB_REFRESH=Y PEND_REASON_UPDATE_INTERVAL=60 SBD_SLEEP_TIME=3 The installer specifies the following initial configuration parameters for all configuration templates: v lsf.conf: EGO_ENABLE_AUTO_DAEMON_SHUTDOWN=Y LSB_DISABLE_LIMLOCK_EXCL=Y LSB_MOD_ALL_JOBS=Y LSF_DISABLE_LSRUN=Y LSB_SUBK_SHOW_EXEC_HOST=Y LSF_PIM_LINUX_ENHANCE=Y LSF_PIM_SLEEPTIME_UPDATE=Y LSF_STRICT_RESREQ=Y LSF_UNIT_FOR_LIMITS=MB v lsb.params: ABS_RUNLIMIT=Y DEFAULT_QUEUE=normal interactive JOB_ACCEPT_INTERVAL=0 MAX_CONCURRENT_JOB_QUERY=100 MBD_SLEEP_TIME=10 PARALLEL_SCHED_BY_SLOT=Y In addition, the installer enables the following features for all configuration templates: v Fairshare scheduling (LSF Standard Edition and Advanced Edition): All queues except admin and license have fairshare scheduling enabled as follows in lsb.queues: Begin Queue ... FAIRSHARE=USER_SHARES[[default, 1]] ... End Queue v Host groups (LSF Standard Edition on UNIX or Linux): Master candidate hosts are assigned to the master_hosts host group. 16 Release Notes for Platform LSF v User groups (LSF Standard Edition on UNIX or Linux): LSF administrators are assigned to the lsfadmins user group. v Affinity scheduling in both lsb.modules and lsb.hosts. LSF event streaming You can now enable LSF event streaming during installation by specifying ENABLE_STREAM="Y" in install.config before running the LSF installer. Enable LSF event streaming if you intend to install IBM Platform Analytics or IBM Platform Application Center. Block scheduling For applications that are not especially sensitive to network latency, or where you prefer to get throughput, you can allocate slots for a parallel job with a specific block size. The applications specified by the job may be running as threaded processes on groups of n cores, but using MPI applications or other socket connections between blocks. LSF will allocate slots to the job based on block size. LSF tries to pack as many blocks on one host as possible, then goes to next one. Each host is only checked once. It does not matter which host contains the slot blocks. The job can start as soon as any previous job is complete. This packing policy is supported by the keyword block (“span[block=value]”) in the span section of the resource requirement string. “span[block=value]” can also be configured in the RES_REQ parameter in lsb.queues and lsb.applications. When a block size is specified for a job, LSF allocates only a multiple of the block size for the job. The minimum and maximum values in -n min,max are also adjusted to be a value of multiple of the block. Define GPU or MIC resources You can enable LSF so applications can use Nvidia Graphic Processing Units (GPUs) or Intel MIC (Phi co-processors) in a Linux environment. LSF supports parallel jobs that request GPUs or MICs, allowing you to specify a certain number of GPUs or MICs on each node at run time, based on availability. Specifically, LSF 9.1.2 supports the following: v Nvidia GPUs for serial and parallel jobs. Parallel jobs should be launched by blaunch. v Intel MIC (Phi co-processor) for LSF jobs in offload mode, both serial and parallel. v CUDA 4.0 to CUDA 5.5. v Linux x64: MIC supports Linux x64. Linux-based GPUs support x64 for RHEL/Fedora/SLES. LSF also supports the collection of metrics for GPUs and MICs using elims and predefined LSF resources. The previous GPU package is replaced by the new design for LSF 9.1.2. If you want to use the previous GPU package, do the following: For an upgrade install: Chapter 1. Release Notes for IBM Platform LSF 17 1. Replace elim.gpu in $LSF_SERVERDIR with the old GPU elim. 2. Keep the old configuration files and do not use the LSF9.1.2 GPU related resource definition. 3. Restart the cluster For a new install: 1. Replace elim.gpu in $LSF_SERVERDIR with the old GPU elim. 2. Define GPU related resources as the old solution. 3. Restart the cluster. Host based pre- and post-execution processing LSF previously featured job-based pre- and post-execution processing which was intended for sequential jobs, and where pre- and post-execution processing ran only on the first execution host. For release 9.1.2, LSF features host-based pre- and post-execution processing, which is intended for parallel jobs (you can also use this feature for sequential jobs) and runs on all execution hosts. The purpose of this is to set up the execution hosts before all job-based pre-execution and other pre-processing which depend on host-based preparation, and clean up execution hosts after job-based post execution and other post-processing. This feature can be used in a number of ways. For example: v HPC sites can have multiple ways to check for system health before actually launching jobs, such as checking for host or node status, key file systems are mounted, infiniband is working, required directories, files, environment, and correct user permissions are set, etc.) v Administrators can configure site specific policy to run host-based pre- and post-execution processing to set up ssh access to computer nodes. By default, ssh is disabled. However, with host-based pre- and post-execution processing, ssh access to the nodes allocated for the job can be enabled for the duration of job life cycle. This is required for debugging a parallel job on a non-first execution host and will not impact the overall cluster security policy. v Administrators can configure host-based pre- and post-execution processing to create and later remove temporary working directories on each host. The following configuration parameters can be used for both job-based and host-based pre- and post-execution processing: v JOB_PREPROC_TIMEOUT v JOB_POSTPROC_TIMEOUT v v v v v LSB_PRE_POST_EXEC_USER LSB_POSTEXEC_SEND_MAIL JOB_INCLUDE_POSTPROC LOCAL_MAX_PREEXEC_RETRY MAX_PREEXEC_RETRY v REMOTE_MAX_PREEXEC_RETRY v LSB_DISABLE_RERUN_POST_EXEC Kerberos Support for NFSv4 and AFS When using LSF on NFSv4 or Andrew File System (AFS), each process in a sequential job or a distributed parallel job needs to periodically renew its 18 Release Notes for Platform LSF credentials. For this re-authentication to take place in a secure, user friendly environment, a TGT file is distributed to each execution host and the root sbatchd in each execution host renews the TGT. To support AFS, LSF provides an external renew hook mechanism which is called after TGT is renewed. Users can write their own renew logic through this renew hook. More specifically, users can use the demo script named erenew.krb5 in $LSF_SERVERDIR and rename it to erenew. Users can also create an executable named erenew in $LSF_SERVERDIR. This erenew script will be called immediately at job startup time to make sure the user’s job has a valid AFS token. LSF will also automatically call this binary after TGT is renewed. For example, AFS users can use this hook to run aklog for renewing AFS tokens. Note: blaunch krb5 does not support pre LSF 9.1.2 remote execution server, and therefore the renew script will not work in pre 9.1.2 RES. Similarly, blaunch krb5 does not support pre LSF 9.1.2 sbatchd. Therefore, child sbatchds cannot be kerberized and the renew script does not work in pre 9.1.2 root sbatchd Note: No krb on Solaris platforms for LSF 9.1.2. Energy Aware Scheduling LSF offers energy-aware scheduling features for large-scale LSF installations, where the energy requirements for operating large systems are becoming a significant factor in the overall cost of these systems. On large systems with either a long lead period to full production or widely fluctuating workloads many nodes can sit idle for significant time periods. The energy-aware scheduling features of LSF enable administrators to control the processor frequency to allow some applications to run at lower frequency with minor performance degradation. This can lead to overall power savings. Conversely, minimizing the frequency on unused cores can also enable maximum turbo boost to active cores, to increase application performance, and reduce run times. Frequency control allows an organization to balance performance with power savings. As well, manual or policy controlled power management is available to cluster administrators to suspend (S3) and resume on the specified hosts or host groups through LSF command line. In contradiction, if job runtime is more important than energy savings, the CPU frequency can be increased per job. LSF energy-aware scheduling features include the following: v Host-based policies to manage the power state of hosts. v Ability to set the CPU frequency at the job, application, or user level. v Collection and reporting of power usage for an application (assuming exclusive use of nodes). v Benchmarking application power usage and generation of relevant power coefficients. v Prediction of performance, power usage, and runtime of applications at different CPU frequencies. v Automatic CPU frequency selection for jobs based on predictions. Energy aware scheduling features are available only for LSF Standard Edition. Energy aware scheduling features have some extra installation and configuration requirements and dependency on third party tools. For example: Chapter 1. Release Notes for IBM Platform LSF 19 v STREAM and NPB-NAS Parallel Benchmarks v MySQL DB or xCat MySQL database v mysql-connector-odbc or unixODBC must be on the master/master candidate hosts v cpufrequtils package is installed on all compute nodes v The following Linux kernel modules must be installed on all nodes: msr, ibmaem, ipmi_si, acpi_cpufreq v All compute nodes have P-States and C-States enabled v IBM iDataplex is supported, on homogeneous nodes (same hardware, OS, CPU count, memory) v Hyperthreading must be disabled on all nodes v No compute node may be in turbo-boost mode. For information on system requirements for energy aware scheduling features, see the Energy Aware Scheduling chapter in Administering IBM Platform LSF. License Scheduler Basic Edition available with LSF License Scheduler Basic Edition is now available for use with LSF 9.1.2 at no additional charge. License Scheduler Basic Edition monitors the availability of licenses managed by FlexNet, and throttles the workload of a single cluster to avoid dispatching more jobs with license requirements than can run with the available licenses. It can also be used to view the license use of individual jobs. You can replace an elim that tracks licenses managed by FlexNet with License Scheduler Basic Edition. Note that License Scheduler Basic Edition is intended for use only in cases where licenses are not shared by multiple clusters. It does not dynamically balance license use among multiple clusters, or among multiple projects. In cases where such functionality is required, use License Scheduler Standard Edition. To install and run License Scheduler Basic Edition, download and install the License Scheduler packages as described in Using IBM Platform License Scheduler (specifically, in the Installing and starting License Scheduler section), but follow any specific steps for installing and configuring License Scheduler Basic Edition instead of Standard Edition. Changes to default LSF behavior With no new features enabled in a newly upgraded LSF 9.1.x cluster, the following pre-9.1 functionality has changed: v When installing LSF Standard Edition on UNIX or Linux, you may select a configuration template by specifying the CONFIGURATION_TEMPLATE parameter in the install.config file. The LSF installer sets initial configuration parameters based on the configuration template that you select. For more details, refer to “Installer configuration templates and initial settings” on page 14. v By default, bjobs –l reports individual host rusage for a parallel job. Set LSF_HPC_EXTENSIONS=NO_HOST_RUSAGE to enable pre-LSF 9.1.x behavior. v bswitch does not modify the effective resource requirement of a running job based on the resource requirements of the destination queue res req definition. You can define BSWITCH_MODIFY_RUSAGE to enable pre-LSF 9.1.x behavior. 20 Release Notes for Platform LSF v LSF now uses non-privileged ports by default for daemon communication. You can set LSF_NON_PRIVILEGED_PORTS=N in lsf.conf to enable privileged port communication. Also, LSF_MC_NON_PRIVILEGED_PORTS and LSF_NON_PRIVILEGED_PORTS are now fully decoupled, which is different from previous versions. If you are upgrading your master host and leaving some server hosts still running older versions, do the following: v If LSF_NON_PRIVILEGED_PORTS is already set to Y or N, continue with upgrade. v If LSF_NON_PRIVILEGED_PORTS is not set, but LSB_MAX_JOB_DISPATCH_PER_SESSION is set to a value greater than 300, do the following: 1. Shut down the cluster 2. Set LSF_NON_PRIVILEGED_PORTS=Y 3. Upgrade the master host 4. Restart the cluster v If neither LSF_NON_PRIVILEGED_PORTS nor LSB_MAX_JOB_DISPATCH_PER_SESSION is set, do the following: 1. Set LSF_NON_PRIVILEGED_PORTS=N. 2. Upgrade the master host. 3. Start LSF on the master host. See Upgrading IBM Platform LSF on UNIX and Linux for detailed upgrade steps. New commands New commands added for LSF 9.1.2. v bentags: Used with energy policy, or the energy aware scheduling feature. The bentags command queries or removes information about the energy policy tag from mbatchd which is saved in the database. This command displays all the energy tag names that have been generated by the user and can remove energy policy tags. New configuration files New configuration file added for LSF 9.1.2. v lsb.threshold: To enable the automatic select CPU frequency feature of energy aware scheduling, you must define the lsb.threshold configuration file, using the energy tags (accessed using bentags). The lsb.threshold file is available at the location specified by the parameter PERFORMANCE_THRESHOLD_FILE in lsb.params. The default location is $LSF_ENVDIR/lsbatch/cluster_name/configdir/lsb.threshold. New and changed commands, options, and output The following command options and output are new or changed for LSF 9.1.2: badmin v New subcommand option hpower is used to manually switch hosts between a power saving state suspend or a working state resume. v Subcommand options hist and hhist can be used to retrieve a host’s history of power state changes. Both badmin and policy (job)-triggered power related events are logged as type HOST_POWER_STATUS. Chapter 1. Release Notes for IBM Platform LSF 21 bapps v HOST_PRE_EXEC: The host based pre-execution command for the application profile. The HOST_PRE_EXEC command runs on all execution hosts before the job associated with the application profile is dispatched to the execution hosts. If job based pre-execution PRE_EXEC was defined at the queue-level/application-level/ job-level, the HOST_PRE_EXEC command runs before PRE_EXEC of any level. The host-based pre-execution command cannot be executed on Windows platforms. v HOST_POST_EXEC: The post-execution command for the application profile. The HOST_POST_EXEC command runs on all execution hosts after the job finishes. If job based post-execution POST_EXEC was defined at the queue-level/applicationlevel/job-level, the HOST_POST_EXEC command runs after POST_EXEC of any level. The host-based post-execution command cannot be executed on Windows platforms. bhist v The -t option displays job events chronologically, including new events for energy aware scheduling, JOB_PROV_HOST and HOST_POWER_STATUS. bhosts v The -l option displays host power states when PowerPolicy is enabled (in lsb.resources). Final power states are on or suspend. Intermediate power states are restarting, resuming, and suspending. If the host batch status becomes unknown (power operation due to failure), the power state is shown as a dash (“-”). bjobs v PROV has been added as a possible value for JOB STATUS in the long format output (-l). This status means the job has been dispatched to a power-saved host that is waking up. Before the job can be sent to the sbatchd, it is in a PROV state. v If the job was submitted with an energy policy, to automatically select a CPU frequency, -l will show the Combined CPU frequency (the CPU frequency selected for the job based on the energy policy tag, energy policy and threshold file). If the job was submitted with a user defined CPU frequency (using bsub –freq), -l will show the Specified CPU frequency for the job. bmod v The -freq option specifies a CPU frequency for a job. The submission value will overwrite the application profile value and the application profile value will overwrite the queue value. The value is float and should be specified with SI units (GHz, MHz, KHz), for example bmod -freq 2.5GHz. If no units are specified, GHz is the default. bqueues v HOST_PRE_EXEC: The host based pre-execution command for the queue. The HOST_PRE_EXEC command runs on all execution hosts before the job associated with the queue is dispatched to the execution hosts. If job based pre-execution PRE_EXEC was defined at the queue-level/application-level/job-level, the HOST_PRE_EXEC command runs before PRE_EXEC of any level. The host-based pre-execution command cannot be executed on Windows platforms. v HOST_POST_EXEC: The post-execution command for the queue. The HOST_POST_EXEC command runs on all execution hosts after the job finishes. If job based post-execution POST_EXEC was defined at the queue-level/application- 22 Release Notes for Platform LSF level/job-level, the HOST_POST_EXEC command runs after POST_EXEC of any level. The host-based post-execution command cannot be executed on Windows platforms. bresources v The -p option displays the currently defined energy aware scheduling policies and exits. Shows the PowerPolicy settings as they are in lsb.resources. An additional line is included with the PowerPolicy settings to indicate whether it is currently Applied (Y) or not (N). bsub v The -freq option specifies a CPU frequency for a job. The submission value will overwrite the application profile value and the application profile value will overwrite the queue value. The value is float and should be specified with SI units (GHz, MHz, KHz), for example bsub -freq 2.5GHz. If no units are specified, GHz is the default. New and changed configuration parameters and environment variables install.config v CONFIGURATION_TEMPLATE: Selects the configuration template for this installation, which determines the initial LSF configuration parameters specified when the installation is complete. CONFIGURATION_TEMPLATE="DEFAULT" | "PARALLEL" | "HIGH_THROUGHPUT" v ENABLE_STREAM: Enables LSF event streaming for Analytics or PAC. ENABLE_STREAM="Y" | "N" lsb.applications v HOST_PRE_EXEC: Enables host-based pre-execution processing at the application level. The HOST_PRE_EXEC command runs on all execution hosts before the job starts. If job based pre-execution PRE_EXEC was defined at the queue-level/application-level/job-level, the HOST_PRE_EXEC command runs before PRE_EXEC of any level. HOST_PRE_EXEC is not supported on Windows platforms. HOST_PRE_EXEC=command v HOST_POST_EXEC: Enables host-based post-execution processing at the queue level. The HOST_POST_EXEC command runs on all execution hosts after the job finishes. If job based pre-execution POST_EXEC was defined at the queue-level/applicationlevel/job-level, the HOST_POST_EXEC command runs after POST_EXEC of any level. HOST_POST_EXEC is not supported on Windows platforms. HOST_POST_EXEC=command v CPU_FREQUENCY: Specifies the CPU frequency for an application profile. All jobs submit to the application profile require the specified CPU frequency. Value is a positive float number with units (GHz, MHz, or KHz). If no units are set, the default is GHz. This value can also be set using the command bsub –freq. The submission value will overwrite the application profile value, and the application profile value will overwrite the queue value. CPU_FREQUENCY=[float_number][unit] lsb.params v SECURE_JOB_INFO_LEVEL: Defines which jobs all users can see information for. A value between 0 and 4 is defined, with 0 being no security and 4 being the highest security. Chapter 1. Release Notes for IBM Platform LSF 23 SECURE_JOB_INFO_LEVEL=0|1|2|3|4 v ENABLE_JOB_INFO_BY_ADMIN_ROLE: Enables user group, queue, and cluster administrators the right to access job detail information for jobs in the user group, queue, and clusters they manage, even when the administrator has no right based on the configuration of SECURE_JOB_INFO_LEVEL.You may define one or more of the values, usergroup, queue, or cluster. ENABLE_JOB_INFO_BY_ADMIN_ROLE=[usergroup] [queue] [cluster] v SECURE_INFODIR_USER_ACCESS: Controls whether regular users can see other user’s jobs when using the bhist or bacct command. If enabled (defined as Y), the primary administrator will still be able to view all job information in lsb.event and lsb.acct. SECURE_INFODIR_USER_ACCESS=Y | N v MAX_JOB_PREEMPT_RESET: Does not reset the preempted count for MAX_JOB_PREEMPT when the started job is requeued, migrated or rerun in SSUSP state. MAX_JOB_PREEMPT_RESET=Y|N v POWER_ON_WAIT: Configures a wait time (in seconds) after a host is resumed and enters ok status, before dispatching a job. This is to allow other services on the host to restart and enter a ready state. The default value is 0 and is applied globally. POWER_ON_WAIT=time_seconds v POWER_RESUME_CMD: Defines the resume operation script that will be called when handling a power resume request. POWER_RESUME_CMD=command v POWER_RESET_CMD: Defines the reset operation script that will be called when handling a power reset request. POWER_RESET_CMD=command v POWER_STATUS_LOG_MAX: Configures a trigger value for events switching. The default value is 10000. This value takes effect only if PowerPolicy (in lsb.resources) is enabled. If a finished job number is not larger than the value of MAX_JOB_NUM, the event switch can also be triggered by POWER_STATUS_LOG_MAX, which works with MIN_SWITCH_PERIOD. Not available with LSF Express edition. POWER_STATUS_LOG_MAX=number v POWER_SUSPEND_CMD: Defines the suspend operation script that will be called when handling a power suspend request. POWER_SUSPEND_CMD=command v POWER_SUSPEND_TIMEOUT: Defines the timeout value (in seconds) for power suspend, resume, and reset actions. When a power operation is not successful (for example, sbatchd does not reconnect when resuming a host) within the specified number of seconds, the action will be considered failed. POWER_SUSPEND_TIMEOUT=integer v PERFORMANCE_THRESHOLD_FILE: Specifies the location of the performance threshold file for the cluster. This file contains the cluster-level threshold values for the minimize energy and minimize time policies used for automatic CPU frequency selection. PERFORMANCE_THRESHOLD_FILE=full_file_path lsb.queues v HOST_PRE_EXEC: Enables host-based pre-execution processing at the queue level. The HOST_PRE_EXEC command runs on all execution hosts before the job starts. If job based pre-execution PRE_EXEC was defined at the queue-level/application- 24 Release Notes for Platform LSF level/job-level, theHOST_PRE_EXEC command runs before PRE_EXEC of any level. HOST_PRE_EXEC is not supported on Windoows platforms. HOST_PRE_EXEC=command v HOST_POST_EXEC: Enables host-based post-execution processing at the queue level. The HOST_POST_EXEC command runs on all execution hosts after the job finishes. If job based pre-execution POST_EXEC was defined at the queue-level/applicationlevel/job-level, theHOST_POST_EXEC command runs after POST_EXEC of any level. HOST_POST_EXEC is not supported on Windoows platforms. HOST_POST_EXEC=command v CPU_FREQUENCY: Specifies the CPU frequency for a queue. All jobs submit to the queue require the specified CPU frequency. Value is a positive float number with units (GHz, MHz, or KHz). If no units are set, the default is GHz. This value can also be set using the command bsub –freq. The submission value will overwrite the application profile value, and the application profile value will overwrite the queue value. CPU_FREQUENCY=[float_number][unit] lsb.resources v PowerPolicy section: Enables and defines a power management policy. lsf.conf v LSB_AFS_BIN_DIR: If LSB_AFS_JOB_SUPPORT=Y, then LSF will need aklog in AFS to create a new PAG and apply for an AFS token. You can then use LSB_AFS_BIN_DIR to tell LSF the file path and directory where aklog resides. If LSB_AFS_BIN_DIR is not defined, LSF will search in the following order: /bin, /usr/bin, /usr/local/bin, /usr/afs/bin. The search stops as soon as an executable aklog is found. LSB_AFS_BIN_DIR=path to aklog directory v LSB_AFS_JOB_SUPPORT : When this parameter is set to Y|y, LSF assumes the user’s job is running in an AFS environment, and calls aklog -setpag to create a new PAG for the user’s job if it is a sequential job, or to create a separate PAG for each task res if the job is a blaunch job. LSF then runs the erenew script after the TGT is renewed. This script is primarily used to run aklog. Finally, LSF assumes that JOB_SPOOL_DIR resides in the AFS volume. It kerberizes the child sbatchd to get the AFS token so the child sbatchd can access JOB_SPOOL_DIR LSB_AFS_JOB_SUPPORT=Y|y|N|n v LSB_EAUTH_DATA_REUSE: When set to Y, blaunch caches authentication data returned by eauth -c when connecting to the first remote execution server in memory. blaunch uses this cached data to authenticate subsequent first remote execution servers. If set to N, blaunch does not cache authentication data. Every time blaunch connects to a different authentication, it calls eauth -c to fetch new authentication data. LSB_EAUTH_DATA_REUSE=Y|y|N|n v LSB_BSUB_ERR_RETRY: In some cases, jobs can benefit from being automatically retried in the case of failing for a particular error. When specified, LSB_BSUB_ERR_RETRY automatically retries jobs that exit with a particular reason, up to the number of times specified by RETRY_CNT. LSB_BSUB_ERR_RETRY=RETRY_CNT[number] ERR_TYPE[error1 error2 ...] v LSF_RES_ALIVE_TIMEOUT: Controls how long the task res on non-first execution hosts waits (in seconds) before cleaning up the job. If set to 0, this parameter is disabled. SF_RES_ALIVE_TIMEOUT=time_seconds Chapter 1. Release Notes for IBM Platform LSF 25 v LSF_DJOB_TASK_REG_WAIT_TIME: Allows users/admin to define a fixed timeout value to override the internal timeout set by LSF in order to avoid task registration timeout for a large parallel job. Can be configured in lsf.conf or set as an environment variable of bsub. The environment variable will overwrite the lsf.conf configuration. If neither is present, jobRES will use the default value. When it is specified by the environment variable or configured in lsf.conf, the value will be directly used by LSF without any adjusting. LSF_DJOB_TASK_REG_WAIT_TIME=time_minutes v LSB_FANOUT_TIMEOUT_PER_LAYER: Controls how long sbatchd waits until the next sbatchd replies. Can also be set as an environment variable. LSB_FANOUT_TIMEOUT_PER_LAYER=time_seconds v LSB_DEBUG_EBROKERD: Sets the debugging log class for the new daemon ebrokerd. Only messages belonging to the specified log class are recorded. Used in combination with LSF_LOG_MASK which sets the log level. LSB_DEBUG_EBROKERD="log_class [log_class...]" v LSF_DEFAULT_FREQUENCY: Sets the default CPU frequency for compute nodes when nodes start and when node has finished a job that uses a different CPU frequency. Value is a positive float number with units (GHz or MHz). If no units are set, the default is GHz. If nothing is set for this parameter, the host’s nominal CPU frequency will be used. LSF_DEFAULT_FREQUENCY=[float_number][unit] v LSF_MANAGE_FREQUENCY: Uses a keyword value (N, CORE, or HOST) to set whether the CPU frequency is set for the core (CPU) or by host (node). If the value CORE is set, jobs will require affinity resource requirements. The default value for this parameter is N (not set). LSF_MANAGE_FREQUENCY=N | CORE | HOST v LSF_COLLECT_ENERGY_USAGE: Determines if the collection of job and node energy usage is enabled on the LSF cluster. This is used for CPU frequency management and energy usage reporting. The default value is N. LSF_COLLECT_ENERGY_USAGE=Y | N New and changed accounting and job event fields The following job event fields are added or changed for LSF 9.1.2. lsb.events v HOST_POWER_STATUS: LSF logs this event when a host power status is changed, whether by power policy, job, or by the command badmin hpower. The HOST_POWER_STATUS event is logged to reflect the power status changes. v JOB_PROV_HOST: When a job has been dispatched to a power saved host (or hosts), it will trigger a power state change for the host and the job will be in the PROV state. This event logs those PROV cases. v cpu_frequency was added to JOB_START, JOB_NEW, and JOB_MODIFY2 to show the CPU frequency at which a job runs. lsb.acct v cpu_frequency was added to JOB_FINISH to show the CPU frequency at which a job ran. Documentation changes This section summarizes major changes and corrections to the LSF documentation since the release of LSF 9.1.1. 26 Release Notes for Platform LSF v Administering IBM Platform LSF now contains content from “Using Platform LSF HPC Features” guide which is no longer published. Most of the features that are technically supported and relevant to LSF 9.1.x users can be found described in the chapter, "Job Execution and Interactive Jobs". v Updates to the IBM Platform LSF Command Reference and IBM Platform LSF Configuration Reference have been made since the released build of the man pages. Please consult these guides for more information on the following: – LSF_AFS_BIN_DIR – LSB_AFS_JOB_SUPPORT – LSB_EAUTH_DATA_REUSE – LSF_RES_ALIVE_TIMEOUT – LSB_FANOUT_TIMEOUT_PER_LAYER – DEFAULT_JOB_OUTDIR – HOST_PRE_PROC and HOST_POST_POC – JOB_POSTPROC_TIMEOUT and JOB_POSTPROC_TIMEOUT – bsub -freq Known issues v Host based pre- and post- execution processing timeout algorithm does not consider the program execution time on each execution host in the allocation, which could be too short to abort the processing. LSF admin should configure JOB_PREPROC_TIMEOUT to a value to indicate the maximum program runtime expected, the timeout algorithm will consider it as a factor. If the job’s host based pre-execution fails, its host based post-execution will be started immediately. It may end up both programs running concurrently if the job is re-dispatched to the same hosts. Configuring JOB_INCLUDE_POSTPROC can avoid the situation. v brequeue does not transfer new TGTs to mbatchd. If a job is re-queued by the brequeue command, the TGT job used is the one cached by mbatchd. v Library requirement for OpenAFS integration permitting jobs with valid Kerberos credentials to access AFS shared directories: libkopenafs.so must be provided in one of the following locations: /lib, /lib64, /usr/lib, /usr/lib64, /usr/local/lib, /usr/local/lib64 v CPU and memory affinity scheduling has the following limitations. – When reservation is enabled, affinity reservation allocations appear as part of the allocated resources in bhosts -aff Jobs that are submitted with a membind=localprefer binding policy may overcommit the memory of the NUMA node they are allocated to bhosts -aff output may occasionally show the total allocated memory on the NUMA nodes of a host as exceeding the maximum memory of the host, this is because the reservations that show in bhosts -aff overcommit the NUMA node. However, LSF will never allow the allocation of running jobs on a host to exceed the maximum memory of a host. – When reservation is enabled, and an affinity job requests enough resources to consume an entire node in the host topology. (for example, enough cores to consume an entire socket), LSF will not reserve the socket for the job if there are any jobs running on its cores. In a situation when there are always smaller jobs running consuming cores, then larger jobs that require entire sockets will not be able to reserve resources. The workaround is to require that all jobs have estimated runtimes, and to use time-based reservation. Chapter 1. Release Notes for IBM Platform LSF 27 v LSF does not check the contents or exit code of the erenew script. If erenew contains the wrong command, AFS tokens may not be renewed and LSF will not report any error in the log file. Therefore, users must ensure that the commands in erenew can renew AFS tokens successfully. v bmod cannot change the memory requirement for a running job if a MEM general resource limit is defined for the user in lsb.resources. v Application checkpointing is not supported on 64-bit Windows 7. v LSF 8.3 blimits does not work with 9.1.x binaries. v For GSLA, a job may pend or receive fewer slots than expected when you ask for a range of slots. Limitations v Parallel restart cannot be used if the mbatchd is configured to use duplicate event logging (LSB_LOCALDIR is configured in lsf.conf). v Processor number is not detected correctly on POWER7 Linux machines v NUMA topology may be incorrect after bringing cores offline. v Cannot remove the energy tag from a job. The workaround is to kill the current job, and submit a new one. Bugs fixed The December 2013 release (LSF 9.1.2) contains all bugs fixed before 8 October 2013. Bugs fixed between 25 June 2013 and 8 October 2013 are listed in the document Fixed Bugs for Platform LSF 9.1.2. This document is available on Platform LSF’s IBM Service Management Connect at http://www.ibm.com/ developerworks/servicemanagement/tc/. 28 Release Notes for Platform LSF Chapter 2. Platform LSF product packages The Platform LSF product consists of the following packages and files: v Product distribution packages, available for the following operating systems: Operating system Product package IBM AIX 6 and 7 on IBM Power 6 and 7 lsf9.1.2_aix-64.tar.Z HP UX B.11.31 on PA-RISC lsf9.1.2_hppa11i-64.tar.Z HP UX B.11.31 on IA64 lsf9.1.2_hpuxia64.tar.Z Solaris 10 and 11 on Sparc lsf9.1.2_sparc-sol10-64.tar.Z Solaris 10 and 11 on x86-64 lsf9.1.2_x86-64-sol10.tar.Z Linux on x86-64 Kernel 2.6 and 3.x lsf9.1.2_linux2.6-glibc2.3-x86_64.tar.Z Linux on IBM Power 6 and 7 Kernel 2.6 and 3.x lsf9.1.2_linux2.6-glibc2.3-ppc64.tar.Z Windows 2003/2008/2012/XP/7/8 32-bit lsf9.1.2_win32.msi Windows 2003/2008/2012/XP/7/8 64-bit lsf9.1.2_win-x64.msi Apple Mac OS 10.x lsf9.1.2_macosx.tar.Z zLinux Kernel 2.6, glibc2.4 SLES 10 lsf9.1.2_lnx26-lib24-s390x-64.tar.Z Cray Linux XE6, XT6, XC-30 lsf9.1.2_lnx26-lib23-x64-cray.tar.Z ARMv7 Kernel 3.6 glibc 2.15 lsf9.1.2_linux3.6-glibc2.15-armv7.tar.Z v Installer packages: – lsf9.1.2_lsfinstall.tar.Z This is the standard installer package. Use this package in a heterogeneous cluster with a mix of systems other than x86-64 (except zLinux). Requires approximately 1 GB free space. – lsf9.1.2_lsfinstall_linux_x86_64.tar.Z Use this smaller installer package in a homogeneous x86-64 cluster. If you add other non x86-64 hosts you must use the standard installer package. Requires approximately 100 MB free space. – lsf9.1.2_no_jre_lsfinstall.tar.Z For all platforms not requiring the JRE. JRE version 1.4 or higher must already be installed on the system. Requires approximately 1 MB free space. – lsf9.1.2_lsfinstall_s390x-64.tar.z Installer package for zLinux platform. Includes zLinux specific JRE. Requires approximately 300 MB free space. The same installer packages are used for LSF Express Edition, LSF Standard Edition, and LSF Advanced Edition. v Entitlement configuration files: – LSF Standard Edition: platform_lsf_std_entitlement.dat – LSF Express Edition: platform_lsf_exp_entitlement.dat. – LSF Advanced Edition: platform_lsf_adv_entitlement.dat. v Documentation packages: – lsf9.1.2_documentation.tar.Z © Copyright IBM Corp. 1992, 2013 29 – lsf9.1.2_documentation.zip Downloading the Platform LSF product packages Download the LSF installer package, product distribution packages, and documentation packages from IBM Passport Advantage: www.ibm.com/software/howtobuy/passportadvantage. 30 Release Notes for Platform LSF Notices This information was developed for products and services offered in the U.S.A. IBM® may not offer the products, services, or features discussed in this document in other countries. Consult your local IBM representative for information on the products and services currently available in your area. Any reference to an IBM product, program, or service is not intended to state or imply that only that IBM product, program, or service may be used. Any functionally equivalent product, program, or service that does not infringe any IBM intellectual property right may be used instead. However, it is the user's responsibility to evaluate and verify the operation of any non-IBM product, program, or service. IBM may have patents or pending patent applications covering subject matter described in this document. The furnishing of this document does not grant you any license to these patents. You can send license inquiries, in writing, to: IBM Director of Licensing IBM Corporation North Castle Drive Armonk, NY 10504-1785 U.S.A. For license inquiries regarding double-byte character set (DBCS) information, contact the IBM Intellectual Property Department in your country or send inquiries, in writing, to: Intellectual Property Licensing Legal and Intellectual Property Law IBM Japan Ltd. 19-21, Nihonbashi-Hakozakicho, Chuo-ku Tokyo 103-8510, Japan The following paragraph does not apply to the United Kingdom or any other country where such provisions are inconsistent with local law: INTERNATIONAL BUSINESS MACHINES CORPORATION PROVIDES THIS PUBLICATION “AS IS” WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF NON-INFRINGEMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Some states do not allow disclaimer of express or implied warranties in certain transactions, therefore, this statement may not apply to you. This information could include technical inaccuracies or typographical errors. Changes are periodically made to the information herein; these changes will be incorporated in new editions of the publication. IBM may make improvements and/or changes in the product(s) and/or the program(s) described in this publication at any time without notice. Any references in this information to non-IBM Web sites are provided for convenience only and do not in any manner serve as an endorsement of those Web © Copyright IBM Corp. 1992, 2013 31 sites. The materials at those Web sites are not part of the materials for this IBM product and use of those Web sites is at your own risk. IBM may use or distribute any of the information you supply in any way it believes appropriate without incurring any obligation to you. Licensees of this program who wish to have information about it for the purpose of enabling: (i) the exchange of information between independently created programs and other programs (including this one) and (ii) the mutual use of the information which has been exchanged, should contact: IBM Corporation Intellectual Property Law Mail Station P300 2455 South Road, Poughkeepsie, NY 12601-5400 USA Such information may be available, subject to appropriate terms and conditions, including in some cases, payment of a fee. The licensed program described in this document and all licensed material available for it are provided by IBM under terms of the IBM Customer Agreement, IBM International Program License Agreement or any equivalent agreement between us. Any performance data contained herein was determined in a controlled environment. Therefore, the results obtained in other operating environments may vary significantly. Some measurements may have been made on development-level systems and there is no guarantee that these measurements will be the same on generally available systems. Furthermore, some measurement may have been estimated through extrapolation. Actual results may vary. Users of this document should verify the applicable data for their specific environment. Information concerning non-IBM products was obtained from the suppliers of those products, their published announcements or other publicly available sources. IBM has not tested those products and cannot confirm the accuracy of performance, compatibility or any other claims related to non-IBM products. Questions on the capabilities of non-IBM products should be addressed to the suppliers of those products. All statements regarding IBM's future direction or intent are subject to change or withdrawal without notice, and represent goals and objectives only. This information contains examples of data and reports used in daily business operations. To illustrate them as completely as possible, the examples include the names of individuals, companies, brands, and products. All of these names are fictitious and any similarity to the names and addresses used by an actual business enterprise is entirely coincidental. COPYRIGHT LICENSE: This information contains sample application programs in source language, which illustrates programming techniques on various operating platforms. You may copy, modify, and distribute these sample programs in any form without payment to IBM, for the purposes of developing, using, marketing or distributing application 32 Release Notes for Platform LSF programs conforming to the application programming interface for the operating platform for which the sample programs are written. These examples have not been thoroughly tested under all conditions. IBM, therefore, cannot guarantee or imply reliability, serviceability, or function of these programs. The sample programs are provided "AS IS", without warranty of any kind. IBM shall not be liable for any damages arising out of your use of the sample programs. Each copy or any portion of these sample programs or any derivative work, must include a copyright notice as follows: © (your company name) (year). Portions of this code are derived from IBM Corp. Sample Programs. © Copyright IBM Corp. _enter the year or years_. If you are viewing this information softcopy, the photographs and color illustrations may not appear. Trademarks IBM, the IBM logo, and ibm.com® are trademarks of International Business Machines Corp., registered in many jurisdictions worldwide. Other product and service names might be trademarks of IBM or other companies. A current list of IBM trademarks is available on the Web at "Copyright and trademark information" at http://www.ibm.com/legal/copytrade.shtml. Intel, Intel logo, Intel Inside, Intel Inside logo, Intel Centrino, Intel Centrino logo, Celeron, Intel Xeon, Intel SpeedStep, Itanium, and Pentium are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries. Java™ and all Java-based trademarks and logos are trademarks or registered trademarks of Oracle and/or its affiliates. Linux is a trademark of Linus Torvalds in the United States, other countries, or both. LSF®, Platform, and Platform Computing are trademarks or registered trademarks of International Business Machines Corp., registered in many jurisdictions worldwide. Microsoft, Windows, Windows NT, and the Windows logo are trademarks of Microsoft Corporation in the United States, other countries, or both. Other company, product, or service names may be trademarks or service marks of others. Privacy policy considerations IBM Software products, including software as a service solutions, (“Software Offerings”) may use cookies or other technologies to collect product usage information, to help improve the end user experience, to tailor interactions with the end user or for other purposes. In many cases no personally identifiable information is collected by the Software Offerings. Some of our Software Offerings can help enable you to collect personally identifiable information. If this Software Notices 33 Offering uses cookies to collect personally identifiable information, specific information about this offering’s use of cookies is set forth below. This Software Offering does not use cookies or other technologies to collect personally identifiable information. If the configurations deployed for this Software Offering provide you as customer the ability to collect personally identifiable information from end users via cookies and other technologies, you should seek your own legal advice about any laws applicable to such data collection, including any requirements for notice and consent. For more information about the use of various technologies, including cookies, for these purposes, See IBM’s Privacy Policy at http://www.ibm.com/privacy and IBM’s Online Privacy Statement at http://www.ibm.com/privacy/details the section entitled “Cookies, Web Beacons and Other Technologies” and the “IBM Software Products and Software-as-a-Service Privacy Statement” at http://www.ibm.com/software/info/product-privacy. 34 Release Notes for Platform LSF Printed in USA GI13-3413-03