Synchronization questions

I’m developing a replicated Berkeley DB application by using the replication framework.
The application is running but the client and the server are not stable (specifically, the replication layer is very unstable).
There’s some behaviour that I don’t understand.
- When the client is synchronizing with the master, the server always transfer the whole databases, even when I’m importing the databases from the server. Basically, the client clears the whole databases and sync again with the server.
- The priority of the client is 0. But sometimes, when the master is not available, the client becomes the new master. I thought that when the priority is 0, the environment can never be a master site.
Before adding the replication layer, the application was running very well. Since I added the replication layer, the application is unstable and crash regularly due to databases locking.
Huu-Dong Quach

- When the client is synchronizing with the master,
the server always transfer the whole databases, even
when I’m importing the databases from the server.
Basically, the client clears the whole databases
and sync again with the server.I don't understand this part of your question. What do you mean by "importing"?
- The priority of the client is 0. But sometimes,
when the master is not available, the client becomes
the new master. I thought that when the priority is
0, the environment can never be a master site.Yes, this is a bug in the Replication Framework. When there are more than 2 sites in a group, if the master fails, the remaining sites hold an election. In that case we do correctly pass the 0 priority to rep_elect. However, if there are only two sites, when the master fails there is no point holding an election with just one remaining site, so we just appoint ourselves master. We should be checking for the 0 priority in that case.
This bug will be fixed as soon as possible.
Alan Bram
Oracle

Similar Messages

An interesting synchronization question

hey all, I have an interesting synchronization question to ask you.
Suppose that I have 20 tasks and 8 worker threads, each worker is assigned a task by the main thread, and then main thread suspends until one worker has finished its task. Then main thread create another task and handles to a new worker thread until all tasks are finished.
so code in main thread is something like this.
boolean has_more_task, has_more_worker;
do {
              if (has_more_task && has_more_worker)
                           * create new thread and assign it a new task
              else if ( has_more_task && !has_more_worker)
                        wait();
               else if ( !has_more_task && has_more_worker)
                          wait();
} while (has_more_task || has_more_worker)the problem is that I don't quite familiar with synchronized key word. so I don't know how to share a synchronized vaiable between main thread and worker thread. can anybody do me a help?
Remember that main thread should be waken up by either of the 8 worker threads !!

you can follow producer consumer conept
create a queue. main thread will push task in queue.
and worker threads will consumer task from queue.
you just need to synchronize operation on queue.
when there is no task in queue then worker thread will notify to main thread and
will call wait().
then main thread again put the more task in queue and send notification signal to worker thread and then call wait().
Class JobQueue
//should be called by worker thread
public synchronized getNextTask()
//shoule be invoked by main thread
public synchronized addMoreTask()
}

Synchronize question in servlet

Hello everybody!
I have some question regarding synchronization in a servlet, and hope someone can help me.... (see example code below)...
1) The methods "getNames(), addNames(), removeNames()" need to be synchronized due to that the "names" vector can be accessed by many at the same time, right?
2) Do I need to synchronize the method "addANumber()"? why/why not?
3) Do I need to synchronize the methods "setString(), getString()" placed in the "Test" class? why/why not?
4) All "global" variables, i.e. the variables declared as object or class variable, need to be synchronized. Is this valid for servlet methods as well? For example the addANumber() method, or whatever method (that doesnt use global variables) I place in the Servlet class (ASR-class)
Please all your experts out there, help me.....
Thanx in advanced
TheMan
import java.io.*;
import javax.servlet.*;
import javax.servlet.http.*;
import java.util.*;
public class ASR extends HttpServlet {
    private Vector names;
    private ServletConfig config;
    public void init(ServletConfig config) throws ServletException {
          synchronized(this) {
            this.config=config;
            names = new Vector(100,10);
    public void doGet(HttpServletRequest request, HttpServletResponse response) throws IOException, ServletException {
        response.setContentType("text/plain");
        PrintWriter out = response.getWriter();
        //add a name to the vector
        String theName = request.getParameter("inpName");
        addNames(theName);
        //get the vector and print that name
        Vector tmp = getNames();
        int pos = Integer.parseInt(request.getParameter("number"));
        out.println("a person " + tmp.get(pos));
        addANumber(pos);
        //remove a name
        removeNames(pos);
        //create a test object
        Test t = new Test();
        //set a string in the test object
        t.setString("yyy");
        //print that string
        out.println("test string: " + t.getString());
    public void doPost(HttpServletRequest request, HttpServletResponse response) throws IOException, ServletException {
        doGet(request, response);
    //needs to be synchronized?
    public int addANumber(int number){
        int adding = number + 10;
        return adding;
    public synchronized Vector getNames(){
        return names;
    public synchronized void addNames(String theName){
        names.add(theName);
     public synchronized void removeNames(int pos){
         names.remove(pos);
class Test {
    public String str = "xxx";
    //needs to be synchronized????
    public void setString(String str){
        this.str = str;
     //needs to be synchronized????
    public String getString(){
        return str;
}

The teachers, which are responsible for the
servlet-courses that I have taken, never spoke about
the member variables, in the same way you did. They
used member variables and synchronized methods or
synchtronized code blocks to access them. They never
said that "one should not/try not to use member
variables". I can understand this in the context of using Servlets to gain understanding on some of the issues involved with multithreading. In a classroom, servlets may pose an easier laboratory (so to speak) for learning threading concepts than something like nachOS or something like that.
For example, they made a ordinary "page counter",
i.e. a counter which increases by one for each "hit"
on the page. I that example they, and I, thought it
would be a good idea to use a member variable (a int
counter), since all instances of the servlet will use
the same variable. How should one construct such page
otherwise? Local variables and databases is one way.As mentioned in an earlier post, in the real world I can think of very few reasons as to why a particular servlet may need to keep state. In terms of keeping a counter, what you would probably do is to analyze the access logs...Yes -- not a programmatic solution, but a very doable, low maintenance, and quick one. One important lesson to learn is that not all problems are best solved in code. This is probably not the right answer for a test though...
I just got a bit confused, my teachers had no problem
with member variables, but you guys recomend me to
use local variables instead. Which leader should I
follow :) ?Capitalist answer: whoever pays you :) Seriously, why not talk to your teachers about what you have learned here versus what they are teaching?
- N

Synchronize question

When programing servlet, I found out there is some concepts I do not understand very well; so would like to ask to make clear.
I have a servlet in which a service will go to database to obtain a file name; then generating file object and return back to the client.
public class AServlet extends ...{
    static AService aService = new AService();
    public void doGet(...){
        aService.performActionsInDB();
        aService.getFileName();
public class AService{
    private String fileName;
    public void performActionInDB(){
        this.fileName = .....//assign a file name variable obtained from db to local variable fileName
    public String getFileName(){
        return this.fileName;
}Now my question arises. The 'fileName' is a global variable in which different threads can access (so it needs syncrhonize key word). But what I am not clear is where should I add the synchornize key word would have less performance impact and also prevent concurrent access problem (a thread is reading fileName that is being updated by another thread)?
My first thought is to add sycnhronize key word enclosing
        aService.performActionsInDB();
        aService.getFileName();above code block, but it would also lock the execution while performing DB actions which performe several tasks as well.
Other place I can think of is the e.g.
synchronized(this.name){
this.fileName = ...// .//assign a file name variable obtained from db to local variable fileName
}And
public synchronized String getFileName(){...}But this seems would have problem because a thead is reading the fileName which may be updating by other thread.
What is the correct to place the key word?
I appreciate any suggestion. Thank you very much.
Edited by: shogun1234 on Mar 29, 2010 5:43 PM

Just synchronize the smallest block that updates the name.

Trying to understand threads; interesting synchronize question

Ladies and Gentlemen,
what would happen if:
class c {
public synchronized void a() {
//do some stuff
b();
public synchronized void b() {
// do some stuff
this should cause a deadlock situation, should it not? The compiler doesnt complain when I try this.
Can someone confirm if this is correct:
any class method can be synchronized; it doesnt have to be a method in a thread you ahve created. Presumable, this method and its object are being manipulated by threads, so synchronization of data is necessary. I have a program that has multiple threads. Each thread get a reference to manager object (there is one object for all the threads, not one for each). Each thread has its own instance of an analysis object. The analysis object takes teh reference to the manager object in its constructor, so the end result is mulitple threads each have their own analysis object, and all of these objects point to one manager object. I want the methods of the manager object to be synchronized to avoid read/write conflicts and inconsistencies.
Thank you in advance

You are right it is not officially deadlock. but it
is a situation that will produce an error if one is
not careful. b is called in a, and they both require
a lock, so the processing in b cant be done while a is
running. No, I'm telling you that is not the case. You can call b() from a() because if a thread is in a() it already has the lock needed to call b(). There is no possiblity of deadlock with the code you have written.
In order for a deadlock to be possible, you'd need to do something like this:
class Example
   final lockObjectA = new Object();
   final lockObjectB = new Object();
   void a()
      synchronized(lockObjectA)
          b();
   void b()
      synchronized(lockObjectB)
          a();
}

ISA 5.0 / NWDI : web-j2ee-engine.xml / synchronization question

Hello experts
I wonder whether this is normal that whenever I start a synchronization on the crm/isa/web/b2c DC (or in my custom DC which synchronizes the used DCs including crm/isa/web/b2c), nwds asks me to create an activity to take into account changes on the web-j2ee-engine.xml file ?
I select cancel each time but maybe I'm wrong ?
Is there a solution that nwds not ask me more ? Is there a mistake in any configuration file ?
Thank you very much and have a nice day
Julien

hi
sap friends
i have configured B2B in CRM 5.0 sr2 (ABAP+JAVA) on same host VIA XCM
while login into this link
http://idescrmsr2.iserviceglobe.com:50000/b2b/b2b/init.do
http://idescrmsr2.iserviceglobe.com:50000/isauseradm/useradmin/init.do
with user test , this user is having the following roles
SAP_ALL
sap-new
sap_j2ee_admin
sap_crm_eco_isa*
sap_crm*
sap_crm_isa*
all the b2b related roles i have given, but still am unable to login into above URLs
it is giving error message like :"<b>LOGON IS INVALID,CHECK UR ENTRIES</b>"
is there any configuration related users configuratiom
plz help me anbody regarding this problem

DSEE and Directory Synchronizer Question

I'm investigating using Directory Synchronizer to sync up our DSEE 6.x directory with AD.
We do not have our directory set up to where users can change their password directly. They must go to the authoritative source (our web portal) and change their password there, then have it synced out to DSEE. We are using SSHA passwords for this, but also have a SHA-1 hash of the password.
So, how will Directory Synchronizer be able to get passwords into AD? I don't think AD uses or recognizes this password format, does it?

http://docs.sun.com/app/docs/doc/819-0993/gbfza?l=Ja&a=view should help you.

BW Landscape synchronization question

Hi, Guru
We have two BW DEV and QA systems in the system landscape: BWD & BWQ is the project phase1 (finance) D & Q system, respectively. BCD & BCQ is the phase2 (supply chain) D & Q system, respectively. Phase 1 and 2
share a single production system BWP.
Phase 1 is already live for over a year, but there is discrete
enhancement work done over the time in BWD & BWQ environment.
Phase 2 work is done in BCD/BCQ environment. Phase 2 is going live soon. We are attempting to consolidate the two landscapes into one after phase2 go-live. This way only need to maintain one D & Q system from that point on.
The decision is to keep BWD and abandon BCD since phase 1 is more critical. So we are planning to MANUALLY RECREATE all phase 2 work in BWD, this include everything:
infoCube, ODS, update rule, process chain, infoPackage, user exit, queries, etc...After that BCD, the original development system for phase 2, will be removed from the landscape.
Phase 2 work is to be transported to production from BCD-BCQ route.
The concern we have now with this approach is that we are afraid of transport for future enhancement work for phase 2, which will be done on top of the manually created objects in BWD, won't work. For example, even objects are now recreated and look exactly same, they are going to bear different IDs in BWD than BCD. Would transport work then?
There may be other things we have not considered yet? Is this approach feasible?
We know it is a rare case (rebuilding development system beneath Q and P). But has anyone done this before? can anyone foresee a problem with this approach? Remember it is BW, not R/3. Thanks.

Hi Latha ,
Normally we would do a Prod system Copy to Dev' and then start the upgrade. Also ensure that you make a GAP analysis of all Inactive objects and unused objects and delete them on DEV system using a transport request and move it to Q and prod, before you start the upgrade.
We followed this recently in our project for upgrade and it worked fine for us.
Cheers
Raja

Ace fail over / synchronization question

Hey all,
I have a customer who has a ace HA pair, the primary ace is shut down, and they've been making changes to the standby ace which has been working ok.
They want to bring up the primary ace again, but I just want to confirm the process so I don't overwrite the configuration of the current standby ace when the primary ace is brought back online.
I don't have any experience with these boxes yet. But I was thinking about turning preemption off and increases the standby priority to make it the primary?
Thoughts?
Many thanks.
Sent from Cisco Technical Support iPad App

Hi,
If you want to sync the config then you dont have to use the following command.
no ft auto-sync running-config
no ft auto-sync startup-config
Start as follows:
(1) Configure a FT VLAN interface & FT PEER on “new replacement ACE”.
Configure all FT groups BUT DO NOT “configure them “inservice”.
Make sure you have IP connectivity OVER FT VLAN to “currently ACTIVE ACE”.
Make sure there is a TCP connection setup OVER FT VLAN (show conn should provide you that information).
(2) Please make sure “preemption” is NOT enabled for the FT group. If enabled please do remove it and re-add after the module is successfully replaced.
Example:
Example:
               ft group 1
                                    peer 1
                                    no preempt <=====================
                                    peer priority 150
                                    associate-context test
(3) Once you have IP connectivity over FT VLAN to “primary ACE”, now mark the FT GROUP “inservice”.
Example:
               ft group 1
                                    peer 1
                                    no preempt
                                    peer priority 150
                                    associate-context test
                          inservice <===============================
(4)   At this time I expect the “auto-sync” to “sync” configs between “currently ACTIVE ACE” & “new standby ACE”.
show ft group detail
show ft peer detail
These “show commands” should help you with verifying the state of FT configuration.
(5) Repeat the above procedure for all context one by one ( Bring Admin context FT "inservice" at the end )
In case if you have are using SSL offloading in any context refer the following thread:
https://supportforums.cisco.com/thread/2156101?tstart=0&viewcondensed
Hope that helps.
regards,
Ajay Kumar

To Synchronize or not to synchronize that is the question

I find myself using static methods a lot since they are very convenient in that you don't have to instantiate an object and populate members, etc. rather its often more convenient just pass something in and get something out.
However, I'm not crystal clear on when its necessary to synchronized a method, it appears most of the time I can get away without synchronizing.
Could you post a simple java example or describe a scenario when its necessary to synchronize a static method?

917903 wrote:
//If there were 1000 Threads doing this in various environments, calling getUnique contending for a unique value, would the method need to be synchronized?The method does need to be static, because it's being called from a static context. If it weren't being called from a static context, then it wouldn't necessarily have to be static.
However whether it has to be synchronized depends solely on what it does. Whether it has a "static" modifier makes no difference at all.
And if you're going to synchronize it, you do have to be careful about what object you choose to be the monitor for that synchronization, it's true. However that is above and beyond the simple "To synchronize or not to synchronize" question.

UserManagerLookupService findGroupMembers restricted to 200?

Hi All,
I have a hopefully simple question.
Scenario:
1) I have LCES3 connected & synchronizing to the clients MS AD nightly. (Enterprise Domain setup in the LC Adminui)
2) I need to periodically query a specific group and do something with each user.
3) The specific group contains 975 users
4) Im using the Find Group & findGroupMembers operations from the UserManagerLookupService to retrieve a list of users.
Problem:
1) The findGroupMembers operation is only returning 200 users.
What I have tried / Noted:
1) Modifying the offset & resultSize values of the findGroupMembers operation to try get the next “batch” of users – no luck.
2) I opened the GROUP in LC AdminUI and noted that there are only 200 users in the GROUP – I resynchronized
2 a) Then in LC AdminUI, I went to “Edit Enterprise Domain” -> Directory -> Directory Group. I clicked TEST and saw the GROUP. I clicked on the GROUP, and noted in the “member” Attributes Names in the Attribute Value it listed all 975 users.
3) I have resynchronzied the Enteprise domain multiple times, but still no reflection (all users showing in the group in AdminUI or via Workbench findGroupMembers query).
4) The users that need to be in the group, however are showing in the LC AdminUI as users, however in the user’s user groups, it shows that it is not associated with that group.
5) Googling - on one of the adobe blogs, it speaks about a config file - but doesn't say where it is? (http://blogs.adobe.com/apugalia/lifecycle-of-livecycle-domain-synchronization/)
Question:
1) Why does LC AdminUI findGroupMemmbers not return the full set of users in a group in a synchronized Enterprise domain (it limits it to 200)?
Im hoping this is a simple adminui update, setting or config file change.
Thanking you all in advance

Just while im at it...
So after doing the above, you need to go back to Domain Management and Edit your Authentication & Directory for the specific domain.
You will need to put the password in again (as per warnings in step 1 above in previous post).
When putting in your password, it may give you an error saying the "username and password is incorrect".
And you will probably find a nice error in your server error log:
javax.naming.AuthenticationException: [LDAP: error code 49 - 80090308: LdapErr: DSID-0C0903A9, comment: AcceptSecurityContext error, data 775, v1db0]
Go to https://confluence.atlassian.com/display/CONFKB/LDAP+Error+Code+49 to figure out what the bold and underlined value means.
In my example 775 means user account is locked. (why... when I did the config.xml import, alot of things happened on the server error log and assume it locked the AD account)

New Entrant to WLSP4.0. Facing plenty problems

Deployed wls6.1sp1. then wlsportal. then ebcc
now trying to run petflow application.
It starts fine (took me some time to fish thru the password - "weblogic" from this
forum).
When i invoke the URL (localhost:7501/petflow), it gives some configuration exception.
Came to know (from this forum itself) that i have to do synchronisation thru ebcc.
When i try to do that thru ebcc, it first asks me username/password. i give system/weblogic.
Then it opens a box and puts lot of files in "incomplete documents" space.
Finally, cant synchronize
Questions:
which usename/password should i give.
What needs to be done to run this petflow app?
Thanks in Advance,
Prashant Gupta

Prashant,
Welcome! The mostly likely explanation is that you are not sync'ing to the
correct URL. What is the Data Sync URL you are using? This should be the
"context URI" of the DataSync Web Application deployed within the petstore
EAR (you can check these in the console).
Usually it is something like:
http://localhost:7501/p13nAppDataSync/DataSyncServlet
(for the P13N EAR).
Sincerely,
Daniel Selman
"Prashant Gupta" <[email protected]> wrote in message
news:3cec97b3$[email protected]..
>
Deployed wls6.1sp1. then wlsportal. then ebcc
now trying to run petflow application.
It starts fine (took me some time to fish thru the password - "weblogic"from this
forum).
When i invoke the URL (localhost:7501/petflow), it gives someconfiguration exception.
Came to know (from this forum itself) that i have to do synchronisationthru ebcc.
When i try to do that thru ebcc, it first asks me username/password. igive system/weblogic.
Then it opens a box and puts lot of files in "incomplete documents" space.
Finally, cant synchronize
Questions:
which usename/password should i give.
What needs to be done to run this petflow app?
Thanks in Advance,
Prashant Gupta

New user - questions about email setup and synchronization

Hi,
I'm a not-particularly-techno-savvy new owner of a BlackBerry Curve 8900. My questions are:
1 I seem to only have the option to use a work email account with a BB Enterprise server and not the option to create or add an email addy. How do I resolve this?
2 I've installed the BlackBerry Desktop Manager but when I try to synchronize data from my device - organiser stuff and contacts - I get a message saying there are no applications configured, with subsequent instructions as to how to do this but I then have to choose from a huge list of folders and other stuff and have no idea what to do!
I'm really hoping someone will be kind enough to try and help as I'm floundering dreadfully.
Thanks in advance.
Jan

Jan,
When you install Desktop Manager it gives you the option to setup with an enterprise server or internet email, I can't tell from your post which you have.
However, to pick the applications to sync, assuming you have Desktop Manager 4.7,
open Desktop Manager
Click on Synchronize on the left side of the screen
On the next screen that opens, choose Synchronize under the word "Configure" (also put check marks in all the boxes that will allow you to check)
On the next screen to open choose Synchronization on the "right" side
In a couple of seconds another screen will open up where you can choose what applications to sync (calendar in Outlook, tasks in Outlook, Contacts in Outlook etc) just choose each program you need to sync and choose the program on your PC to use such as Outlook.
I also have the 8900 and it syncs with my Outlook 2003 for contacts, calendar, memo and tasks.
Hope that helps.
Steve

The question about portlet customization and synchronization

I have a question about portlet customization and synchronization.
When I call
NameValuePersonalizationObject data = (NameValuePersonalizationObject) PortletRendererUtil.getEditData(portletRenderRequest);
portletRenderRequest.setPortletTitle(str);
portletRenderRequest.putString(aKey, aValue);
PortletRendererUtil.submitEditData(portletRenderRequest, data);
Should I make any synchronization myself (use "synchronized" blocks or something else) or this procedure is made thread-safe on the level of the Portal API?

HI Dimitry,
I dont think you have to synchronize the block. i guess the code is synchronized internally.
regards,
Harsha

User list synchronization and Unique userid questions

I am new to Oracle portal and LDAP and learning more and more about Portal every day.
Hi
I am using Oracle 9iAS portal 3.0.9 version. I have a requirement to integrate 3rd party LDAP with Oracle Portal Single Sign On. I have white paper on Configuring Oracle9iAS Portal for LDAP authentication. I have following questions
Paragraph from white paper (Background information):
When using LDAP authentication or any other external repository, for that matter- the list of users for authentication is held on the external repository. However, there is also a list of users held on the Login server, which is used to associate privileges to the user. Ideally, this list is maintained transparently and automatically. In fact, if a user account is created on LDAP and a user attempts to log in, the login will succeed, and an entry is automatically created on the login server for that user, after a successful login.
Questions:
1.     What privileges user will be granted when synchronization process create new portal user on login server automatically?
2.     Is it possible to customize whatever default privileges new portal user gets? If yes, how? Please provide some forum link or documentation or example.
Unique user Id scenario.
Our LDAP repository is setup for customers from different companies. We have requirement to integrate LDAP users with Oracle 9iAS portal 3.0.9 Single Sign On. I have two userid with same name on the LDAP from different company. For example userid jsmith from company A and jsmith from company B. Both user id do not exist on Portal Login Server. Both userids will be created automatically in Login server when LDAP and Portal synchronize user list.
1.     How oracle portal will handle such scenario when portal requires unique userid?
2.     Can I customize portal login screen? For example when they login they can provide userid, passwd and domain name. Where domain name could be company name.
Let me know if you need more information. Feel free to send direct e-mail also.

Dumlu,
For 1, you can have the user removed from OUL in OOB scenarios, but behind IP phones it's difficult since we won't know when the PC is offline from there. Only way to know that is when CAM receives a MAC-Notification of a new MAC address being learnt. In IB, you can use heartbeat timers to log them out
For 2, when a new MAC address is seen on the port, the MAC-Notification is sent out, and depending on your port profile the switchport will change or not. Check your port profile settings for more details on how you have it setup.
HTH,
Faisal

Synchronization questions

Similar Messages

Maybe you are looking for