Design Considerations for a Repository Manager in a clustered environment

Hi,
I am currently building a repository manager for a backend document management system.  Our portal is installed in a load balanced clusterd environment and we are creating the following functionality in the repostiory manager -
  -  Browse content
  -  Read/Write document Metadata 
  -  Upload/download documentation
  -  Reserve/Unreserve documentation
  -  Search
How will a clustered environment affect the implementation of the repository manager?  Would I need to check for a clustered installation in any kind of way in my code?  Currently I do not see why this would be necessary, but i'm not entirely sure and I need this manager to work in a clustered environment.
I could see this as being an issue if we are caching information about the documents in the repository.  However, we are not.  We will only be caching user information which can be stored separately in a single cache on each server. 
Thanks for your help,
Scott

Hi Scott!
There might be two reasons for knowing whether other cluster nodes are running or not and for communicating with them:
- synchronizing access to the backend
- synchronizing caches
If your repository manager must care for serialized access to the backend you will have to synchronize cluster-wide. You don't have to, if your backend can handle parallel access on its own (like a web-server).
If you cache data in your repository and you want to update the caches of the repositories on the other cluster nodes (maybe because you don't get events for changed data from your backend), you must think of sending messages between cluster nodes. You don't have to, if you think that the expiration of your cache entries is shorter than the required timeliness of the data.
Kind Regards, Dirk

Similar Messages

  • Design considerations for mobile version of website

    My company has just implemented a new version of our coporate website using Oracle Portal, and ADF. However we do not have comprehensive mobile support but it is required. From my research I've learned we should be using ADF mobile. However, the resources I have found have been geared more at developers and I work as a Business Analyst. Can you point me to online resources (white papers, tutorials etc.) that deal with requirements and design considerations for porting over to a mobile version?
    Thanks in advance.

    My apologies for lateness in checking this thread.
    Here is an article I wrote that might help. Most of the paper is geared toward developer, but the first few sections talks about best practices for mobile browser app in general, not just for iPhone. http://www.oracle.com/technology/pub/articles/huang-iphone.html
    You can also check out this link: http://www.orato cle.com/technology/tech/wireless/adf_mobile.html.
    As for design consideration, a few rules of thumb:
    - First, obviously use ADF Mobile and Trinidad components. We put a lot of effort in adding rendering support for different mobile devices.
    - Next, figure out what devices you want to support. Browsers found in smartphones varies greatly, and in consumer/feature phones, the support is even less consistent. In general you should be able create one app/set of screens for all mobile devices, but you should plan on having ability to test it out on different devices.
    - Determine what information is really needed by mobile user. Not all information available in desktop application may be applicable for mobile users
    - Design your mobile screens based on a few key principals:
    - Display data as user needs it, instead of trying to display everything. For example, instead of displaying master-detail data on the same screen, break it out into 2 screens. Master data may be a list, and user would click on a particular piece of data to look at the details of that master record.
    - Provide navigation buttons on each screen, and ensure they are easy to get to on a page. For example, using the iPhone paradigm, there is a navigation bar at the top of the page where you can go between views.
    - Place command buttons at location where it's easy for user to get to. For example, if you need to allow user to modify a long list of fields, you may want to place save button on top and button of the screen, so user can easily get to it without having to scroll around too much.
    - Use style sheets to achieve mobile-platform optimized UI. For example, if you intend to support touch screen devices (iPHone, BB Storm, etc), then style your application so buttons, command links, etc, are big enough so it's easy to get to. For non-touch screen devices, then it's OK to compress the UI, but ensure user can easily flow between controls to get to the functionality they need. For example, using a trackball to scroll to a button.
    Thanks,
    Joe Huang

  • Help needed: Search function for custom Repository Manager

    Hi there,
    i'm writing my own Repository Manager for EP 6.0 SP2 PL4.
    I've serious problems getting the search function running. I'm using the 'new', Netweaver-based API.
    I think, there are some general questions to answer:
    - My Repository Manager exposes lots of custom 
       properties to the system. I create the properties
       for example using following code:
       IProperty newProp = new StringSingleValueProperty(
         PropertyName.getPropertyName(
                  "{}" + myAttName),
         myAttVal,                         PropertyFlags.constructFlags(
                  true, false, true, false)
        Is there any convention for the propertie's name?
        I think it has to be written in lowercase only?
        Can I use ANY namespace I want instead of the
        empty default namespace? (For example:
    - After indexing my resources (of course, containing my
       custom properties), the TREX-Monitoring screen for my
       index shows all my custom properties in the 'Document
       Properties' area, but all uppercase chars are escaped
       somewhat. Is this OK?
    - Trying to use the index above to search for particular
       resources in my repository by my custom properties
       fails all the time ('no docs found...'); searching for
       any system property works fine. What's the reason?
    - How can I force the (Property-) search uses my
       custom namespace-manager's methods
       isQueryPropertyNameUnderstood() and
       getQueryBuilder()
       to perform a 'pure' property search by the repository
       manager? It seems that the TREX always handles the
       queries by itself, because my methods are never called.
    Thanx for you help,
    Daniel

    I created a new manager and copied the classes exactly from the "simple" example and I see similar behavior. Actually, when I do this, folders are recursively copied, but not files. Again, browsing and viewing of files work fine. Again, I am not seeing any unexpected exceptions.
    The only thing that changed were the names and namespaces. Nothing else at all was changed.
    This makes me think I am missing some configuration somewhere or there is something special about the example project that the wizards are not creating. The only difference I see in the project file is an additional nature (com.tssap.selena.cmf.PatternNature), but I am not sure that is relevant.
    Wow, I am really stumped.
    -Alex

  • Possible reasons for a repository manager not displayed under KM Content

    Hi,
    I developed a repository manager and deployed the par file Archive uploader.
    Most time after deployment, I can see immediately an instance of the repository manager under KM Content. However, sometimes, when I made some changes to the code, it doesn't appear under KM Content. The deployment log in SYS/global/config/cm/reports shows no error.
    I tried to find some log in System Administration - Monitor, but found nothing.
    So, could anyone tell me:
    1. What might be the reason for the problem that the repository manager cannot display under KM Content after deploying the par?
    2. Where can I find the log explaining why my repository manager cannot display?
    Thanks,
    Ray

    Hi Ray,
    Deploying a repository manager means a hot KM deployment, which isn't supported at all, see https://www.sdn.sap.com/irj/sdn/go/portal/prtroot/docs/library/uuid/bb1f18be-0801-0010-ffa0-e3495f7401bc and SAP note 894884. After such a deployment, the KM is restarted, but you have some kind of "shadow KM" in memory, which leads for most things to an unpredictable behaviour.
    In contrast to the documents reffered to above, also the standard KM delivery from SAP still behaves in an uncontrolled way for many issues.
    That's why after a KM hot deployment, you should restart your server if you realize anything "unusual". A workaround is to debug your code and using the hot code replacement feature (which only works in some situations, but that's better than nothing...).
    Hope it helps
    Detlev

  • Special consideration for coding clients while using weblogic clustering

              Hi.
              I wanted to know whether there are any special consdirations to
              be taken while coding clients(servlets in our case) for EJBs when
              we are using clustering of App servers.
              In our setup, we are following the recommended 2-tier architecture
              for Weblogic5.1
              thanks and regards,
              Amol
              

    Amol wrote:
              > Hi,
              > adding a couple of points to the previous mail:
              > 1. In case of servlet failure after doXXX() method has been
              > called and before the method returns, if the app server
              > fails, will doXXX() be called again on the servlet in the
              > secondary server?
              > 2. How much time does it take for the servlet session on
              > one machine to be replicated on the second machine when
              > using in-memory replication.
              It depends on the size of the object and complexity.
              -- Prasad
              >
              > Thanks,
              > Amol
              >
              > "Amol" <[email protected]> wrote:
              > >
              > >Hi kumar (and everybody),
              > >
              > >Your opinion resolves the doubt i had in mind, since even
              > >WebLogic
              > >docs do not say anything about clustering affecting the
              > >coding
              > >of bean clients.
              > >
              > >However may I request you to please evaluate the same
              > >query with
              > >respect to the following conditions:
              > >1. The Session bean calls a stored proc on the db. After
              > >the call
              > >to stored proc and before the stored proc returns with
              > >the results,
              > >the app server fails.
              > >Here there could be two cases of interest, a) Stored proc
              > >does
              > >a write on db -OR- b) Stored proc does only a read on
              > >db
              > >2. The Session bean calls an Entity bean which "writes"
              > >to the
              > >db. The db write is successful, but before the entire
              > >call to the
              > >session bean is completed(that is before user gets his
              > >results)
              > >the Server fails. Now permanent changes have been made
              > >to the db,
              > >but client doesnt get the results.
              > >
              > >Please note the fact that we are using recommended 2-tier
              > >architecture
              > >and in this case server failure would imply failure of
              > >Servlet
              > >as well as the beans.
              > >
              > >Regards,
              > >Amol
              > >--------------------------
              > >
              > >Kumar Allamraju <[email protected]> wrote:
              > >>Absolutely not. Make sure the objects you are putting
              > >>in the session are
              > >>serializable.
              > >>
              > >>
              > >>
              > >>Amol wrote:
              > >>
              > >>> Hi.
              > >>> I wanted to know whether there are any special consdirations
              > >>to
              > >>> be taken while coding clients(servlets in our case)
              > >>for EJBs when
              > >>> we are using clustering of App servers.
              > >>>
              > >>> In our setup, we are following the recommended 2-tier
              > >>architecture
              > >>> for Weblogic5.1
              > >>>
              > >>> thanks and regards,
              > >>> Amol
              > >>
              > >
              Cheers
              - Prasad
              

  • Design considerations for custom C223

    Hi all,
    I'm building a custom report using C223 as a base. I think I've determined that the struct mkal-expand is used to store the record set returned upon executing the transaction. I want to add a couple of fields to the struct by including another struct, which I define, in the mkal-expand struct. Is that a good aprroach for adding my fields to the report? Also, when I view the screen layout in design mode all of the fields which are displayed in the executed report are NOT displayed? Why...and would I add my fields in that case?
    regards,
    Mat

    Hi WalterLaan
    Thanks again for your assistance. I still think I might be better off drawing my own charts because I know that I have a bunch of unique stuff that I have to add to the charts (results of custom market analysis techniques)... e.g. if I have to draw custom annotations above (and/or between) the 5th and 10th bars on the chart, it will be easier for me to do this if I have drawn the bars myself (i.e. I'll know exactly where the bars are on the screen... to the pixel)... whereas there may be guesswork involved if I have to find the location on the screen of the 5th bar drawn by a third-partly library (unless they explicitly provide that in their API, to determine the x and y co-ordinates of a particular bar).... but even if they provide this, I may have to deal with the complexity of having to add a transparent pane above the chart element (or some other complexity) just so that I can go ahead and draw my own custom annotations to the chart.
    Again, for now I'm thinking I should just draw the charts myself... that doesn't sound too hard. The most important thing I needed to get my head around was the interactions between the data processing piece and the GUI components. I think you have laid out a pretty reasonable system for me in that regard... so thank you.
    Just a quick question.... Does repaint() have to be called from the Event Dispatch Thread? I'm thinking not because as far as I understand it just schedules a paint event. I think what you said made sense... just call repaint after each new piece of data is processed.
    Thanks for all your help
    Eddie

  • Design considerations for wrapping higher level process

    So Ive used workflow now on a few projects and have become comfortable building/coding with it.  Recently a discussion was started to design a new process to automate a part of the business Im in and there are several different departments, each with
    their own different functions and processes that they perform.  During this discussion, the thought came to me - why stop with just the "small" picture where I was using WF to contain and execute business rules, but also wrap the higher level
    abstractions also.
    For example, a department performs steps A, B and C.  Each step might have some branching in it (if/then etc).  Its only the "steps Im using workflow currently.
    Above the department, there is another layer, lets call it "group".  There would be several different groups, each with its own departments, some with the same name as other groups, but the internal structure might be different for a given
    department between groups.  So this means that if I modeled a department in one group, I cant assume I can reuse that department in another groups workflow.
    My main unknown however, is how would a user interact with a workflow that was this high level.  Abstracting higher levels, means users dont need to interact with or execute the whole workflow each time, only a part of it.  So am I looking at pausing/resuming
    the bigger workflow?  How can I "skip" to a certain stage or area of a workflow, bypassing the stuff that has already been done? 
    Correct me if Im wrong, but this seems to be where I need to look at bookmarks?  
    In my case, something similar to an order in a shopping cart is what would be the input type for the bigger workflow.  There are several different groups of users that have to set or add properties to associated objects (sql tables) related to the order
    before it can be marked "complete".
    How would you tie together a "wizard" type concept with workflow?  This seems to be the type of model we are considering.  A user takes an order a few steps, then another user from a different group/department adds some attributes and
    the workflow/wizard continues a few more levels down.  Finally, after all the rules have been matched and required data provided, the workflow can "finish".
    Hope my description seems clear.  

    My main unknown however, is how would a user interact with a workflow that was this high level.  Abstracting higher levels, means users dont need to interact with or execute the whole workflow each time, only a part of it.  So am I looking at pausing/resuming
    the bigger workflow?  How can I "skip" to a certain stage or area of a workflow, bypassing the stuff that has already been done? 
    If you want to skip to certain state or bypass what stuff that has been done, you need give logic judgment first. Bookmarks are the mechanism that enables an activity to passively wait for input without holding onto a workflow thread. When an activity signals
    that it is waiting for stimulus, it can create a bookmark. So bookmark is available for you during create this application.
    Here are some workflow applications, see:
    Workflow Application
    hope it helps.
    Regards,
    Angie
    We are trying to better understand customer views on social support experience, so your participation in this interview project would be greatly appreciated if you have time. Thanks for helping make community forums a great place.
    Click
    HERE to participate the survey.

  • Getting error while starting Managed server in clustered environment

    Hi All,
    I created a cluster in weblogic 9.2 and two manged servers under that cluster through WLST scripts.
    I started Node manager through connecting WLST and now when i am starting the managed servers its giving the following error :
    wls:/Dist_3/serverConfig> start('managed1','Server')
    Starting server managed1 ....NMProcess: <Jan 30, 2009 2:57:27 PM> <Warning> <I/O
    error while reading domain directory: java.io.FileNotFoundException: Domain dir
    ectory 'C:\bea\weblogic92\common\bin' invalid (domain salt file not found)>
    NMProcess: java.io.FileNotFoundException: Domain directory 'C:\bea\weblogic92\co
    mmon\bin' invalid (domain salt file not found)
    NMProcess: at weblogic.nodemanager.server.DomainManager.initialize(DomainMa
    nager.java:80)
    NMProcess: at weblogic.nodemanager.server.DomainManager.<init>(DomainManage
    r.java:52)
    NMProcess: at weblogic.nodemanager.server.NMServer.getDomainManager(NMServe
    r.java:242)
    NMProcess: at weblogic.nodemanager.server.Handler.handleDomain(Handler.java
    :212)
    NMProcess: at weblogic.nodemanager.server.Handler.handleCommand(Handler.jav
    a:105)
    NMProcess: at weblogic.nodemanager.server.Handler.run(Handler.java:66)
    NMProcess: at java.lang.Thread.run(Thread.java:595)
    NMProcess:
    I/O error while reading domain directory: java.io.FileNotFoundException: Domain
    directory 'C:\bea\weblogic92\common\bin' invalid (domain salt file not found)
    Traceback (innermost last):
    File "<console>", line 1, in ?
    File "<iostream>", line 1250, in start
    WLSTException: 'Error occured while performing start : Error starting the server
    Error occured while performing start : Server with name managed1 failed to be st
    arted Use dumpStack() to view the full stacktrace'
    wls:/Dist_3/serverConfig> start('managed2','Server')
    Starting server managed2 ....NMProcess: <Jan 30, 2009 3:05:51 PM> <Warning> <I/O
    error while reading domain directory: java.io.FileNotFoundException: Domain dir
    ectory 'C:\bea\weblogic92\common\bin' invalid (domain salt file not found)>
    NMProcess: java.io.FileNotFoundException: Domain directory 'C:\bea\weblogic92\co
    mmon\bin' invalid (domain salt file not found)
    NMProcess: at weblogic.nodemanager.server.DomainManager.initialize(DomainMa
    nager.java:80)
    NMProcess: at weblogic.nodemanager.server.DomainManager.<init>(DomainManage
    r.java:52)
    NMProcess: at weblogic.nodemanager.server.NMServer.getDomainManager(NMServe
    r.java:242)
    NMProcess: at weblogic.nodemanager.server.Handler.handleDomain(Handler.java
    :212)
    NMProcess: at weblogic.nodemanager.server.Handler.handleCommand(Handler.jav
    a:105)
    NMProcess: at weblogic.nodemanager.server.Handler.run(Handler.java:66)
    NMProcess: at java.lang.Thread.run(Thread.java:595)
    NMProcess:
    I/O error while reading domain directory: java.io.FileNotFoundException: Domain
    directory 'C:\bea\weblogic92\common\bin' invalid (domain salt file not found)
    Traceback (innermost last):
    File "<console>", line 1, in ?
    File "<iostream>", line 1250, in start
    WLSTException: 'Error occured while performing start : Error starting the server
    Error occured while performing start : Server with name managed2 failed to be st
    arted Use dumpStack() to view the full stacktrace'
    wls:/Dist_3/serverConfig>

    hi kartheek,
    i think u just missed some of steps while starting managed server using nodemanager in WLST.
    If u dont mind i'm sending some of steps just check it out. Once u try the same steps.
    1.start the ADMIN SERVER
    2.start the WLST
    java weblogic.WLST
    3.Connect to AdminServer
    connect('username',password','t3://adminhost:listenport')
    example: connect('weblogic','weblogic','t3://localhost:7001')
    4.start the nodemanager in WLST
    startNodeManager()
    5.connect to nodemanager
    nmConnect('username','password','nodemanagerhost',
    'nodemanagerlistenport','domainname','path to domain')
    example: nmConnect('weblogic','weblogic','localhost','5556','base_domain'
    'C:/bea/user_projects/domains/base_domain')
    6.start the managed server in WLST
    start('managedservername','Type','t3://managedhost:listenport')
    example:start('man1','Server','t3://localhost:7003')
    7.If u want to start cluster in WLST
    start('clustername','Type')
    example: start('cluster1','Cluster')
    If u want to start managed in Cluster u follow STEP-6 only
    regards
    abhi

  • Customizing creation iview for Repository Manager

    Hi All,
    How to customize the creation iview (NEW menu) for Custom Repository Manager?
    For example  if we try to use “New submenu” under folder option for CM repository instance “documents” manager then we have lots of option like creation of (folder,upload,textfile,htmlfile,link,external link, Form), where as if  you try to use “New “ submenu for  “ume”  instance, it has very less customized option.
    In fact how do I add the creation of some new type other that specified one in New submenu?, or remove one for my repository from the menu?
    Please help
    Regards
    Arati

    hi
    just check in this link
    see the zip file in it
    https://www.sdn.sap.com/irj/servlet/prt/portal/prtroot/docs/library/uuid/fcc4b3c7-0701-0010-55ab-86e3ec0ce275
    Download Zip file for RepositoryManager (com.sap.bc.rf.manager.ZipfileRepositoryManager.zip)

  • PCD Repository manager

    Hi,
    I am looking for PCD Repository manage feature provided with SP 14. It says that for configuration of the PCD repository manager, choose Content Management ® Repository Managers ® PCD Repository. After applying SP 14 I am unable to find this 'PCD Repository' link inside.
    Can any one tell me how can I see this link inside repository management.
    Thanks
    Vishal

    Hi Vishal,
    the full navigation path is System Administration - System Configuration - Knowledge Management - Content Management - Repository Managers - PCD Repository.
    I expect that you know this way - but I expect that you forgot to click "Show Advanced Options" beneath the heading "Repository Managers", right?!
    Last but not least, this repository is a feature of KMC SPS14, not of the EP itself; but I also expect that you have upgraded both (just that the question is within the wrong forum).
    Hope it helps
    Detlev

  • How to develop repository manager with initial folder structure

    Hi 2 all!
    I wonder if it is possible to create new repository programmatically with defined folder structure. I'm interested in solutions when repository manager is works with db or file system. Can you provide any example or tell me where can I define my structure?
    thnx

    Hi Nata!
    Basicly this task consists of two steps:
    1. Deploy a new configuration object for the repository manager
    2. Create the intial folder structure
    Step one is simple. Here is the configuration object of the repository manager for /documents. It's been exported from the KM configuration by using the standard export mechanisms. This can be included into a portal application project in Netweaver Studio.
    documents.co.xml
    <?xml version="1.0"  encoding="UTF-8" ?>
    <!-- file generated by configuration framework Jun 27, 2008 9:16:36 AM -->
    <Configurable configclass="CmRepositoryManager" owner="km.appl.rep.mgr.prjconfig">
    <property name="ro-expiry-time" value="0"></property>
    <property name="fsconttrack" value="true"></property>
    <property name="poolid" value="dbcon_wcmdb_cm"></property>
    <property name="securitymgr.aclcacheid" value="ca_rsrc_acl"></property>
    <property name="localroot"></property>
    <property name="serverfarm" value="false"></property>
    <property name="propertysearchmgr.class" value="com.sapportals.wcm.repository.manager.cm.CmPropertySearchManager"></property>
    <property name="keepVersionHistory" value="false"></property>
    <property name="securitymgr.aclcfg" value="acl_jdbc_rep"></property>
    <property name="versioning.localroot"></property>
    <property name="persistence.mode" value="db"></property>
    <property name="hidden" value="false"></property>
    <property name="fsnosync" value="false"></property>
    <property name="cacheid.versionhistory"></property>
    <property name="prefix" value="/documents"></property>
    <property name="fsContentCacheMaxEntries" value="-1"></property>
    <property name="fsContentCache"></property>
    <property name="name" value="documents"></property>
    <property name="cacheid.content"></property>
    <property name="compressLimit" value="-1"></property>
    <property name="cacheid.foldercontent"></property>
    <property name="cacheid.properties"></property>
    <property name="active" value="true"></property>
    <property name="cacheid.rid"></property>
    <property name="ignorerootdisplayname" value="false"></property>
    <property name="class" value="com.sapportals.wcm.repository.manager.cm.CmRepositoryManager"></property>
    <property name="cclist"></property>
    <property name="intLinksDefaultDynamic" value="true"></property>
    <property name="contentmgr.class" value="com.sapportals.wcm.repository.manager.cm.CmNamespaceAndContentManager"></property>
    <property name="cacheid.documents"></property>
    <property name="versioningmgr.class" value="com.sapportals.wcm.repository.manager.cm.CmVersioningManager"></property>
    <property name="securitymgr.w2kcfg.systemid"></property>
    <property name="lockmgr.class" value="com.sapportals.wcm.repository.manager.cm.CmLockManager"></property>
    <property name="cacheid.locks"></property>
    <property name="securitymgr.class"></property>
    <property name="persistence.db.instance-id" value="documents"></property>
    <property name="cacheid" value="ca_cm"></property>
    <property name="namespacemgr.class" value="com.sapportals.wcm.repository.manager.cm.CmNamespaceAndContentManager"></property>
    <property name="autoversioning" value="false"></property>
    <property name="description"></property>
    <property name="securitymgr.ref" value="AclSecurityManager"></property>
    <property name="fsContentCacheLimit" value="100"></property>
    <property name="propertymgr.class" value="com.sapportals.wcm.repository.manager.cm.CmPropertyManager"></property>
    <property name="services" value="accessstatistic, com.dkv.DKVApprovalAndTranslation, com.dkv.jobtool.repository.JobToolCreationService, comment, discussion, feedback, layout, personalnote, properties, rating, statemngt, subscription, svc_acl, tbp"></property>
    <property name="sendevents" value="true"></property>
    </Configurable>
    This is a DB repository. The config object for a filesystem repository might differ.
    Step 2, the creation of the initial file structure may be done with in two different ways:
    a) Include a Repository Service with your repository, which creates the folder structure on the first access of your repository.
    b) Include a Portal app which has to be called once to create the folder structure.
    For more info about using the API to create km folders take at look at this:
    https://www.sdn.sap.com/irj/sdn/thread?threadID=933638&messageID=5663711#5663711
    Cheers,
      Jürgen

  • Database design (ERD )for Inventory Management System

    Dear All,
    I am going to develop a simple Inventory Management System software using C# .NET for my learning. After searching different forums, many people have suggested to first create a database design for the software. I want a database design, in short, an ERD
    diagram for simple Inventory Management System which shows proper entities(tables), attributes and relationship between entities.
    It would be highly helpful for me as I am newbie to C# and databases.
    Thanks,
    momersaleem

    Dear Rebecca,
    Thanks for you suggestions.
    As I am going to develop IMS for learning purposes so I think I wouldn't need to go in detail regarding Customer name and addresses. However, I am still thinking of adding country attribute in customers' table which I think will be helpful to sort out customers.
    What's the difference between a purchase and an order?  They're usually the same thing, which doesn't mean you're
    wrong, but what are you picturing here? Purchase entity will be used to keep record of purchases you made and an order entity will be used to keep record of orders that cutomers placed.
    Pricing:
    Any order system needs to manage two very distinct bits of data that are easy to confuse. The price in the Product entity is the current
    price. The price in the Order entity is the selling
    price. Not at all the same thing--current price is almost certainly going to change over time. Selling price won't.
    Does it mean that I'll change the price attribute for product to current_price and add selling_price to order table which will help to keep record of price at the time of order?
    Why did you include a quantity field in the Products table? Is it meant to represent stock on hand?
    Yes you are right. It represents stock in hand.
    Could you please recheck the entities relationships as I am not confirmed whether these are correct or not?
    Thanks,
    momersaleem

  • DB type repository manager for information broadcasting?

    I read some article which says only FSDB type of repository manager can be used for BI informaiton broadcasting. I want to broadcast to MOSS server, which uses IIS, but stores everything in database. Is there anyway to use DB or WSDL type repository manager for broadcasting?
    Thanks
    Jane Zhou

    Where did you get this information from? Is there any document that states that? Below is the information I got, it says FSDB has to be used:
    Can I use information broadcasting to distribute precalculated queries, Web applications, and workbooks to a third-party file server, Web server or document management systems?
    Yes. With information broadcasting, you can precalculate queries, Web applications, and workbooks and publish them into the Knowledge Management of the SAP NetWeaver Portal.
    In KM, you can easily create a Repository Manager (CM repository with persistence mode FSDB) that is attached to a file system directory (for example, the directory of an Internet Information Server (IIS)). You have to create a link in the KM folder of documents to the folder of the CM Repository attached to the file system or you can define your CM Repository as an entry point in KM. For more information, see SAP Note 827994 (SMP login required).
    Information broadcasting can automatically put a new report on the third-party file server (for example, using the data change event in the process chain). KM offers repository managers for many different file servers, Web servers, and document management systems (such as IIS and Documentum):
    1.                            Create CM Repository attached to file system.
    2.                            Use iView KM Content to create subfolder in file system (optional).
    3.                            Set permission to Administrator (optional).
    4.                            Create link in /documents to folder of CM Repository attached to file system or define CM Repository as entry point. (See SAP Note 827994.)
    5.                            Schedule Broadcasting Settings that export to a linked folder of CM Repository.
    Because documents created via Information Broadcasting have additional attributes attached to them which mark them as broadcasted documents, it is not possible to store these kind of documents in a "pure" file system repository because such a repository usually only stores properties like "last changed", "creator", etc. Fortunately, KM provides a mechanism to nevertheless use a file system repository to store the documents. The additional properties will be stored in the database. Details are given here and here.
    The "persistence mode" of the repository must be "FSDB" to allow this kind of behavior. Please note that because of the distributed storage of file and additional properties, the property assignment will be lost when moving around the document in the file system using some non-KM tool like windows explorer.

  • Get user manager in workflow SharePoint Designer 2013 for SharePoint online list item

    Hello,
    I would like to create a workflow in SharePoint Designer 2013 for a list in a site on SharePoint online and one of the action would be to send an e-mail to the manager of the user that created an item in a list.
    How can I get the manager of the user and then use it to send an e-mail?
    Thanks!

    In SP Designer there's the following action that I use.
    Find Manager for this user (then output to variable:manager) Action
    This is for SPD 2010 not sure if it's still available in SPD 2013
    Thanks,
    KP
    KP

  • Special consideration for BRF rules related to configuration managment

    Hello All.
    I am looking for a document or oss note that discribes if there are any special consideration for BRF rules related to configuration management and transport control?  I just learned how to spell BRF so, I would appriciate if you could direct me to the information requested above.
    Regards,
    k

    Hello All.
    I am looking for a document or oss note that discribes if there are any special consideration for BRF rules related to configuration management and transport control?  I just learned how to spell BRF so, I would appriciate if you could direct me to the information requested above.
    Regards,
    k

Maybe you are looking for