Replication Synchronization Mirroring

I have a project where I've been asked to provide some (!) fault tolerance. The application is a JDBC/Web application. I've partitioned all the JDBC/SQL stuff into my 'DAO' classes. The app runs on a server with it's own database. My boss wants the app to synchronize with another server (the Database). This effectively means the 2 boxes are mirror copies of each other (sessions don't matter...). I'm thinking I need to use (fast) RMI between the two servers and some clever framework logic to achieve this... has anyone heard of any projects or frameworks on these lines??? I'd rather not start from scratch. Any help appreciated.
Thanks

Thank you, but what I'm looking for is a 'simple' framework that can be applied to a layer above the data access object (DAO), that manages the synchronization of 2 (or more) databases. The problem I've got is that the database is currently MS SQL Server 2000, but the other (secondary) machine could be Linux and MySQL and a third machine could be NT with Interbase.... My DAO needs to store and retrieve the data, but another 'manager' layer needs to coordinate the updates. I have the Core JINI book; Jini looks good, but there's too many additional services that need to be run for it to work and that just means more components that can break....
Thanks again, but I'm still looking, anyone any thoughts on this?

Similar Messages

Replication / synchronization with MaxDB using Windows

Hi,
is there somewhere a tutorial how to use replication / synchronization with MaxDB using Windows?
I found information about the Synchronization Manager but it seems that it's no more available for Download.
The Database Studio description points out that it will be useable for MaxDB 7.7 and upper versions,
but I can't also find V7.7 anywhere for download...
So do I need and where can I get it?
Thanks,
Andre

Hi Wolgang,
thanks for your support - this is the output:
C:\Program Files\sapdb\indep_prog\bin>Exception in thread "main" java.lang.Unsat
isfiedLinkError: no swt-win32-2128 in java.library.path
        at java.lang.ClassLoader.loadLibrary(Unknown Source)
        at java.lang.Runtime.loadLibrary0(Unknown Source)
        at java.lang.System.loadLibrary(Unknown Source)
        at org.eclipse.swt.internal.Library.loadLibrary(Library.java:104)
        at org.eclipse.swt.internal.win32.OS.<clinit>(OS.java:42)
        at java.lang.Class.forName0(Native Method)
        at java.lang.Class.forName(Unknown Source)
        at org.eclipse.swt.internal.Library.getPlatform(Library.java:49)
        at org.eclipse.swt.internal.Library.loadLibrary(Library.java:83)
        at org.eclipse.swt.internal.Callback.<clinit>(Callback.java:31)
        at org.eclipse.swt.SWT.getPlatform(SWT.java:2227)
        at org.eclipse.swt.SWT.<clinit>(SWT.java:2368)
        at org.eclipse.jface.action.ActionContributionItem.<clinit>(ActionContri
butionItem.java:60)
        at org.eclipse.jface.action.ContributionManager.insertAfter(Contribution
Manager.java:268)
        at com.sap.sdb.syncMan.gui.SyncManGUI.drawMSG_LRU(SyncManGUI.java:1422)
        at com.sap.sdb.syncMan.gui.SyncManGUI.createMenuManager(SyncManGUI.java:
428)
        at org.eclipse.jface.window.ApplicationWindow.addMenuBar(ApplicationWind
ow.java:223)
        at com.sap.sdb.syncMan.gui.SyncManGUI.<init>(SyncManGUI.java:230)
        at com.sap.sdb.syncMan.gui.SyncManGUI.main(SyncManGUI.java:821)
Thanks,
Andre

Replication (synchronization mode) between EE and XE

Hello,
I'm developing an application that needs a DB, but the Mobile PC (running the application) is not always connected to the network.
I have a master repository running under Oracle EE 11g
Since the application needs 5 tables and 8 views to work, I'm planning to install, on the Mobile PC, Oracle XE because of its limited footprint.
How can I do to replicate data from master DB to mobile DB, since on the XE part the application must update rows.
So it's a bi-directional replication (synchronization) but the master repository runs under Enterprise Edition.
Could someone help ?

I succeeded to do the replication using Basic Replication method.
Tables in a fast mode, Views in a complete one, with a replication group.

Setup Transactional Replication on Mirrored database

Hi All,
Environment details:
SQL server 2005 SP4 on all servers below.
Server A -- PRINCIPAL
Server B -- MIRROR
Server C -- REPLICATED DB (need to set this up from A to C.)
Database: test
Actions done:
1. DB mirroring (without witness) has been setup from Server A to Server B.
I need to setup replication for a subset of tables (say 3) and only some of the columns of these tables.
Questions:
1. Is this possible?
2. Are there steps which I can follow?
NOTE: In our scenario, it is ok for the replication to break when there is a DB mirror failover from Server A to Server B. Requirement is only to setup replication from Server A to Server C.

1. We can do the replication provided tables has PK.
2. Once added the publisher and subscriber script out the replication which will help you to drop and recreate as and when needed.
3. Keep distribution On C server
4. If tables are large enough create individual publishers for each of them.
Regards, Pradyothana DP. Please Mark This As Answer if it solved your issue. Please Mark This As Helpful if it helps to solve your issue. ========================================================== http://www.dbainhouse.blogspot.in/

Database Mirroring and Replication in SQL Server 2008 R2

I have configured the mirroring and replication between 4 servers (A,B,C,D).i.e, Mirroring between A to B and C to D, Replication between A and C. The configuration was a success and I am able test the replication(B to C) when I have failed over the mirroring
dbs(A to B). The replication works fine after the mirroring fail over but I am not able to check its status in the Replication monitor. When I am having any insert in a table which is replicated in B, it is reflected back to C..it means the replication is
working fine.
Any thought on how I can view the status of replication from mirrored server. I tried adding the publisher in the monitor, but no luck. If I check the snapshot agent status, it says could not retrieve the info, same with the log reader agent status.
Any suggestions on this please.
Thanks, Siri

For example in your case...
Server A is principal and Server B is Mirror with either Manual or Automatic Failover.
Server A is replicated to Server B ( publisher & B is subscriber )
In Server A Database named Test_Mirror_Replication is configured for both mirroring and Replication.
Now you have failed over your Database 'Test_Mirror_Replication' from Server A to Server B.
After the Failover Server A will act as Mirror for DB 'Test_Mirror_Replication' & Server B will act as Principal for DB 'Test_Mirror_Replication'
Hope my understanding is correct ?
If yes then have you tried monitoring the replication after registering in other instance with current principal database sql instance name ? I mean current Publisher database name sql instance ? not your old sql instance name which was before the mirroring
role change or failover...
Hope you are trying with mirror database server name ?
Raju Rasagounder MSSQL DBA

How to do Mirroring of Oracle Database ?

Dear Experts,
I'm working on the oracle 10g database. i need the database replication or mirroring, ( backup of the same database stored into two locations but i want to work on single database server.
Is it possible to do this ?
Could you please, send me the solution.
I need the solution both in software oriented and hardware oriented solution
[email protected]
[email protected]

Hi,
You will need to read the first 6 chapters of oracle data guard to get a handle on it. From there it will be easy to institute using grid control. Here I just finished doing a project.
Best of luck.
al

Does SAP use SQL server's snapshot and transactional replication?

Gurus:
Could you help with this?
Does SAP use SQL server's snapshot and transactional replication?
Thanks!

Hi Christy,
no, SAP does not directly leverage these functions.
But none the less, it is up to you to use these on your system. I regulary use the snapshot functionality when applying Support Packages. In case somehing goes wrong a snapshot is the easiest way to roll back the import process (not exactly the best choice when talking about production and users keep on working while importing).
Have a look at this [document |http://download.microsoft.com/download/d/9/4/d948f981-926e-40fa-a026-5bfcf076d9b9/SAP_SQL2005_Best Practices.doc]. It deals with Best Practices and also covers snapshot, replication and mirroring.
Sven

Some question about replication procedure?

1.At the function __rep_process_message in the file D:\db-4.5.20\rep\rep_record,
there are several types of messages to be handled.
REP_LOG/REP_LOG_MORE : It indicates starting to transfer log record from
client to master
REP_PAGE/REP_PAGE_MORE: What are these messages used to ?
               Does it indicate in some special case, master
will transferdatabase file directly to client to
accomplish replication synchronization ?
2. When does the master start to send log records to client?
I mend source code of ex_rep_base example provided by BDB as follow:
          ret=dbenv->txn_begin(dbenv,NULL,&txn,0);
          //step 1
          Sleep(3000);
          printf("---------------txn_begin Waken--------------------\n");
          if (ret !=0 )
               dbenv->err(dbenv,ret,"transaction begin fail");
               goto err;
          if ((ret = dbp->put(dbp,
          txn, &key, &data, 0)) != 0) {
               dbp->err(dbp, ret, "DB->put");
               goto err;
          //step 2
          Sleep(3000);
          printf("---------------Waken--------------------\n");
          ret=txn->commit(txn,0);
          //step 3
          if (ret!=0)
               dbenv->err(dbenv,ret,"transaction commit fail");
               goto err;
I print message being processed at client.
At step 1: no message is processed.
At step 2: two REP_LOG message are processed.
At step 3: one REP_LOG message is processed.
Does it mean: the master don not need to wait for the put transaction commit to send log records to clients?
And when does the master send log records to clients ?
3. A test case is tried as follow:
     Step 1: Start Master and client;
     Step 2: Add records to master.
     Step 3: kill client and delete the database file of the client.
     Step 4: restart client, and query client's database to check if it has
caught with master.
     (there is not write requests to master at step 4)
     Step 5: send write requests to master.
Result: At step 4, it takes a long period of time for the client to catch
up with master.
     But after step 5, the synchronization procedure is very fast. It takes a
very short period of time for the client to catch up with master.
     What is the reason?
Thanks a lot !

1. LOG and LOG_MORE messages convey log records, usually from the master to the client. (If using client-to-client synchronization, log records may be copied from one client to another, in certain circumstances.)
PAGE and PAGE_MORE messages convey the contents of database pages during "internal init" (which is mentioned in db-4.5.20/docs/ref/rep/mastersync.html). Again, this is usually from master to client, but could be from client to client in some circumstances.
2. The master generally sends log records to clients as soon as the operations that generated them occur (db->put, txn->commit, env->txn_checkpoint, etc). (See also the discussion of Bulk Transfer, however: db-4.5.20/docs/ref/rep/bulk.html)
3. The client realizes it has "caught up" with the master (STARTUPDONE event) when it can process the first "live" log record generated by the master. Thus it relies on the master doing some new write requests after the client has started synchronization.
Alan Bram
Oracle

Merge Replication: subscriber database error

Hello
We have a principal application database running on SQL Server 2005, which has 300 GB information.
This database is synchronized with multiple subscribers using merge replication agent, allowing changes at both the publisher and subscribers.
Client applications connect to the subscribers and not the publisher, so the information is loaded in the subscribers.
Last week, one of the subscribers had storage level error (broken discs) that caused corruption in the file system of the server.
Then, the merge replication with this server began to fail due to bad sectors of the file system. But the database remained online, and the application continued to function normally, so new information was loaded in database.
Because disk errors, it was not possible backed up the affected database or copy databases files to save the information. Subscribers regular backups were NOT performed.
After resolving disk errors and repaired the file system, database was restored,
but had consistency errors which were repaired by the DBCC CHECKDB command with
repair_allow_data_loss option.
Comparing the number of records before and after repair, we estimates that some records were lost in the database.
Replication has not been restored yet.
Now we have the following questions:
- What happens with lost data in the subscriber database if the merge synchronization is reactivated? We estimate that the lost data will be replicated to the Publisher, and want to avoid that.
- Is there any way to sync (download) data from the Publisher to the Subscriber to load the missing records again? Avoiding of course the upload from the subscriber to the publisher.
- If the above point is possible, may you then make the upload to the publisher to sync the information loaded into the database subscriber?
- If these actions are not possible, have we another ways to sync the information, to avoid losing data in the subscriber?
Thanks in advance
Javier
Javier Mariani

Hi,
You can configure the Synchronization direction in Article Property. Right-click the Publication, click Properties. In the Articles page, set the properties of the articles. See the image:
Here is a thread for your reference:
http://social.msdn.microsoft.com/Forums/sqlserver/en-US/dc21b614-b736-409b-883e-3af6c75ab546/merge-replication-synchronization-direction-as-downloadonly-to-subscriber-allow-subscriber?forum=sqlreplication
Thanks.
Tracy Cai
TechNet Community Support

ACS's replication

Hi,
I need an advice about ACS's replication. I have 2 independent ACSs, The first one ACS1 is for routers/switches, remote VPN and some servers using MS-CHAP v1/2, EAP-TLS and PEAP-GTC. The second one ACS2 is only for Wifi APs using EAP-TLS and PEAP-GTC.
I'd like to make a replication between them. As I read, the one should be a primary ACS and the other one should be as secondary ACS.
If I understood well I should move Wifi authentication to ACS1 and authenticate all devices (routes/switches/VPN/wifi) to this ACS1 (with all methods MS-CHAP v1/2, EAP-TLS and PEAP-GTC) as primary. ACS2 will be as secondary ACS2 with the same configuration (routes/switches/VPN/wifi) but authenticates them only in case of its failure.

Hi,
This is actually a right practice. I would also suggest you to let primary ACS authenticate all kind of session including Wi-Fi and use secondary ACS as a standby.
ABOUT DB REPLICATION
====================
Database replication creates mirror systems of ACSs by duplicating parts of the primary ACS setup to one or more secondary ACSs. You can configure your AAA clients to use these secondary ACSs if the primary ACS fails or is unreachable. With a secondary ACS whose ACS internal database is a replica of the ACS internal database on the primary ACS, if the primary ACS goes out of service, incoming requests are authenticated without network downtime, provided that your AAA clients are configured to fail over to the secondary ACS.
I am sending you one link for Setting Up Replication for Cisco Secure ACS, I am sure this example with screen shots gives you better understanding.
Please visit the below suggested ULR:
http://www.cisco.com/en/US/products/sw/secursw/ps2086/products_configuration_example09186a00800e518a.shtml
HTH
JK
Plz rate helpful posts-

Disaster Recovery Solution

I have 2 copes of the same database on 2 different server at two different locations. I have to synchronize the data 100% between the two databases... Please help me to select an appropriate solution to accomplish this.
The 2 databases should be always online
What I can think of is
- Merge Replication.
- Transactional Replication with Updatable subscription..
If you can point some other solutions that would be really great... how about Mirroring.. If I use this I may need a 2 way mirroring... is this possible.

Hi Sudhakar,
According to your description, there are lots of built-in ways in SQL Server
to make two servers synchronize. Some points you should take into account, for example, how much bandwidth and latency between two server, and the size of two database, how often and how much do data change and so on. you can choose the method
most suited for your needs. Merge replication allows various sites to work autonomously and later merge updates into a single. As other post, you can choose replication as a disaster recovery solution. In database mirroring, secondary database
is in an offline mode while mirroring is in place. This can ensure integrity in the secondary database in the event of a failover being required.
For more information, see:
Selecting the Appropriate Type of Replication:
http://technet.microsoft.com/en-us/library/ms152565(v=sql.105).aspx
Replication VS Mirroring, and what to use for a Disaster Recovery setup:
http://simplesql.blogspot.in/2011/01/replication-vs-mirroring-and-what-to.html
Thanks,
Sofiya Li
If you have any feedback on our support, please click here.
Sofiya Li
TechNet Community Support

Db file in Replica was created in wrong folder

Hello!
I created db file at \BDB_HOME\BDB_SUBx in Master, but in Replica site db file was created in wrong folder which is BDB_HOME.
Do I miss something than setting path to DatabaseEnvironmentConfig.CreationDir ?
\BDB_HOME
|
-BDB_SUB1
-BDB_SUB2
-BDB_SUB3
FYI
Ver.5.1.19 (HA)
C#
VisualStudio.Net 2010
Thanks,
Osamu

Hi Osamu,
This is a known limitation with BDB HA at the moment. This issue is tracked internally in SR #19041. We hope to fix it, but do not guarantee that it will make it, in the next BDB release, 5.2.
Osamu, if you have access to MOS (MyOracle Support) please consult Note 1250654.1 (Doc ID 1250654.1) -- How to Correctly Replicate a Directory Structure Involving Subdirectories Between Master and Client. This article explains in detail this issue and how to work around it for now.
For reference, to explain the issue in more detail, suppose on the master the environment directory structure looks something like this:
\BDB_ENV_HOME
     \BDB_SUBDIR_1
     - DbFile_1_1.db
     - DbFile_1_2.db
     - DbFile_1_N.db
     \BDB_SUBDIR_2
     - DbFile_2_1.db
     - DbFile_2_2.db
     - DbFile_2_N.db
     \BDB_SUBDIR_M
     - DbFile_M_1.db
     - DbFile_M_2.db
     - DbFile_M_N.dbOn the client site, the environment directory structure will be incorrectly replicated and will look something similar to this (with all database files ending up under one of the subdirectories):
\BDB_ENV_HOME
     \BDB_SUBDIR_1
     \BDB_SUBDIR_2
     \BDB_SUBDIR_M
     - DbFile_1_1.db
     - DbFile_1_2.db
     - DbFile_1_N.db
     - DbFile_2_1.db
     - DbFile_2_2.db
     - DbFile_2_N.db
     - DbFile_M_1.db
     - DbFile_M_2.db
     - DbFile_M_N.dbThe workaround to resolve this issue depends on a manual hot backup. AUTOINIT should be turned off via rep_set_config with DB_REP_CONF_AUTOINIT and 0. Than the client should be bootstrapped with the databases and logs from the master and recovery should be run as described in the Berkeley DB Programmer's Reference Guide, in the Hot failover section.
Basically, copy all the databases, copy the logs, run catastrophic recovery on the new client environment directory, and start the application. By turning off AUTOINIT you might need to take care of handling the REP_JOIN_FAILURE event error.
To summarize, the workaround implies the following actions:
1. On the client environment call DbEnv::rep_set_config like this:
DbEnv::rep_set_config(DB_REP_CONF_AUTOINIT, 0);With the C# API it's something like:
DatabaseEnvironmentConfig envConfig = new DatabaseEnvironmentConfig();
envConfig.RepSystemCfg = new ReplicationConfig();
envConfig.RepSystemCfg.AutoInit = false;
DatabaseEnvironment env = DatabaseEnvironment.Open(envHomeDir, envConfig);or you could use the DatabaseEnvironment's class RepAutoInit property.
2. Handle DB_REP_JOIN_FAILURE on the client environment, which could be returned from DbEnv::rep_process_message. Handling implies setting an event callback function via set_event_notify for DB_EVENT_REP_JOIN_FAILURE like this:
DbEnv::set_event_notify(eventsHandler);where eventsHandler is a callback function that respects this signature:
void (*db_event_fcn)(DB_ENV *dbenv, u_int32_t event, void *event_info)and internally checks to see if event is DB_EVENT_REP_JOIN_FAILURE. If it is, it may trigger copying of the backup files from the master onto the client.
With the C# API you will set an EventNotifyDelegate delegate using the EventNotify property and check the NotificationEvent.
3. Start the client. Because AUTOINIT is turned off, internal init will not be performed, hence it will not remove the directory structure on the client environment.
4. If the event callback function gets notified of the DB_EVENT_REP_JOIN_FAILURE it means that the backup of the master (databases and log files of the master) is required on the client. Take a manual hot backup as per the guidelines in the Hot failover section, and copy the database and log files onto the client; this is something that the event callback function can do, when it handles DB_EVENT_REP_JOIN_FAILURE by using something like rsync to copy the files between the machines.
5. If a backup was copied onto the client, run a catastrophic recovery in the client environment either manually via db_recover -c, or programmatically by specifying the DB_RECOVER_FATAL flag to the client's DbEnv::open call (with the C# API you specify it via RunFatalRecovery)
6. After the catastrophic recovery will have been performed and the client will have started, all the replication synchronization will work as expected, replication messages will be processed correctly, and the directory structure will be maintained.
Regards,
Andrei

Toplink in near realtime systems

Hi,
The idea is to have toplink commit db transactions to a in-memory-db and then async process posts these transactions to the oracle DB. What is the best way of doing this? We are in the financial market.
Are there any utils/apis in toplink which can do a delayed transaction commit? Ideas/suggestions welcome.
thanks

With TopLink you can postpone commit for a long time and still be able to see modification in the result of your query without sending update to the db.
If you want to use a 2 steps DB for changes -> memory db then Oracle db, that's complicating the queries if they need to represent most recent data. To persist change for memory db to Oracle db you can use popular replication/synchronization scheme. Microsoft Access document some scheme in details. You will have to code the implemention of the synchronization, could be using TopLink for the task.
To get your queries represent the most recent data you could have two TopLink sessions, pointing to both db, then by piggybacking on conform in unit of work feature of TopLink you could get TopLink code do most of the work. By reading first the memory db, then transfer the data with the right state in the uow used to query on the oracle db, simulating modifications to UOW, you can then get your query against oracle db be conformed as usual.

ACS load balancing

if i have CSS and i want to load balance between 2 ACS . do i have to make one of them active and seoncde backup or i can load balance between bother server .
if yes is this will not effect the authentication and databse .
if there is any artical it wile be more better

Hi,
Cisco ACS has a replication feature that allows you to have more than one (1) ACS servers/appliances to provide high-availability/ redundancy. In this case, you will have one primary and more than one secondary (backup) servers.
The database replication creates mirror systems of ACSs by duplicating parts of the primary ACS setup to one or more secondary ACSs. Without load-balancer, you need to add both primary and secondary ACSs in all AAA clients as backup if the primary ACS fails or is unreachable. With a secondary ACS whose ACS internal database is a replica of the ACS internal database on the primary ACS, if the primary ACS goes out of service, incoming requests are authenticated without network downtime, provided that your AAA clients are configured to fail over to the secondary ACS.
The following url provides you with details on how the ACS replication is performed:
http://www.cisco.com/univercd/cc/td/doc/product/access/acs_soft/csacs4nt/acs40/user/sad.htm#wp756102
http://www.cisco.com/univercd/cc/td/doc/product/access/acs_soft/csacs4nt/acs40/user/index.htm
I am not sure about load-balancing two ACSs, but you probably can try this. Behind a load-balancer, maintain the primary/secondary server setup to enable replication (selected items only) from primary to secondary ACS. But pls bear in mind, in replication, only the Primary ACS can send update to backup server, not bidirectional. Backup/secondary ACS can only receive updates. Use the replication features as an update tool between the servers. All changes/updates must be made in your primary ACS only.
In normal ACS replication, all AAA clients need to specify primary and secondary ACS server as backup. With load-balancer, only one (1) IP need is required, which is the virtual IP assigned by load-balancer to represent the two ACSs.
Rgds,
AK

Data replication and synchronization in Oracle 10g XE.

We are trying to do data replication and synchronization sort of thing for all our servers. We are using Oracle 10g. XE. I guess there are some features in oracle already for replication but I am not very sure about them.
To explain it more clearly - we will have individual database servers in our sub-divisions and then divisions and centers and then main server. We need to synchronize at various levels. So If any body is aware of any techniques, please let me know.

Hi,
Could you tell me what exactly synchronisation your talking about..?
we will have individual database servers in our sub-divisions and then divisions >>and centers and then main serverIf you have mulitple DB servers then you can connect it by DB links. also if you are talking DB synchronisation then you can have Triggers,Materialized views.
we also have two independent severs which are synchronised(atleast schema levels).
Regards!

Replication Synchronization Mirroring

Similar Messages

Maybe you are looking for