GOLDEN GATE 078 - GoldenGate

Can any one please assist us to resolve below issue
when start the extract getting the below error
2010-09-15 08:07:13 GGS WARNING Z1-078 Oracle GoldenGate Capture for Oracle,
org.prm: No valid default archive log destination directory found for thread 1.
Thanks in Advance 

What I would do is:
1. Make sure the system user who is starting GG processes has set correct environment variables, such as ORACLE_BASE, ORACLE_HOME, ORACLE_SID
2. If you have all environment variables set, there is a chance that your database is writing archive logs to other directory than the default one. In that case you should use "ALTARCHIVELOGDEST" option of "TRANLOGOPTIONS" Extract parameter to set the path where archive logs are being saved into.

In addition to what Domink said, if this is RAC then make sure the archive logs from each node are available to every node.
Good luck,

Thanks a lot folks, I did , its working fine


How to configure with several DB on source

Hello all,
I configured correctly 2 servers linux with GG 10.4 and oracle 10.2 for simple tables replication.
So i had one source database on server1 and one target on server2 ( 2 datasources on my GG director)
I need now to had a new source database on server1 and replicate tables on the same server2 (other destination schema).
How can i do this ?
Do i need a new GG installation on server1 (as manager process seems to bee dedicated on one DB) ?
Or simply add a new extract from my other db ?
Thanks for any clue. 
If the second source database is the same version of Oracle as the first and running on the same box then all you need to do is add another extract and explicitly set the ORACLE_SID (if you're not using an TNS alias to connect, but remember it still needs access to the redo logs) and possible the ORACLE_HOME if it happens to be different. To override any environment variable in your extract or replicat use the SETENV parameter before the login information (USERID parameter). For example:
If the second source database is a different (major) version then you'll need to download the appropriate build and hence work out of another OGG home.
Remember, a replicat can only read from one trail so you're going to have at least two replicats on your target - one for each source trail.
Hope this helps,
Hi Joe,
I used same database version, so setenv (ORACLE_SID=orcl2) and a new replicat works fine, thanks

OGG Archivelog files

Hi all oracle gurus, as of now we are storing our archivelog files in flash recovery area, i hope a read permission on these files will help OGG to read archive log files right?
One more thing if we are storing archivelogs according to date wise i mean, a folder is created on daily basis for storing our archive logs, so does OGG will be able to figure out in case redo logs not available to read from which directory and which file, is OGG that intellegent enough?
One more thing when we start our extract process, we see in report file unable to read archive log, since they are located in default space, so can anybody tell me how to make OGG process read archive logs from other than default directory?
Thanks in advance. 
You can specify the location for GoldenGate to read/extract archived redo logs. For example, to read from /oradata/archive/log2:
Extract on the source will connect to Oracle and query the data dictionary for archived log locations. 
Thanks Steven, so just read permissions on those files is sufficient right?
And since we are using some folders in archive log directory to save files according to date, will OGG will be able to pick from folders where the data is?
Thanks for you help 
"The Extract process requires an operating system user that has read access to the
transaction log files, both online and archived. On UNIX systems, that user must be a
member of the group that owns the Oracle instance."
If by changing the date for the destination path name (on a daily basis), how are you doing that? Are you altering archive_log_dest_n every day to point to a new location? Or are you relying on Oracle to do that with how it manages the FRA? The path to the file is known to Oracle, and hence, to GoldenGate.
If you are using only the FRA for archived redo logs, you have other management issues to worry about down the road. What is the harm in using an additional location for one of the 31 archive_log_dest_n locations available? 
I tool has this problem earlier in one of my test machines. i.e GG is unable to find the archive log when using FRA (storing in date folders). Up on the suggestion from Steve, I have created alternate archive location for GG and it is working good now.
Hi Vijay,
Thanks for the feedback.

GoldenGate question

I replicate a database from to using GG
I need to make my extract process to read from redo and archivelogs so I put in :
Is this is sufficiant to be sure taht GG will read fropm redo and archivlogs or I need to add another options?
You don't need to tell GG about where archives are stored. It will automatically find that out.
But if there is some alternate archive location (eg where you move the archives after certain duration) then you can use the parameter for alternate archive dest and make GG aware of it. 
First I'm sorry I post the message twice :(
So if I keep my archivelog in the diskgroup for 24 hours it seams that GG will be able to get 24 hours of transaction!?
looks ok,but make sure you're able to connect with ASM instace by sys#ASM_DB as this required some changes at listener.ora file. 
Yes. When it comes to know that the information it needs is no more there in the online redo logs, it will switch to archives. If it can't find the required archive log there also then it will abend with archive log missing error. 
Thank you very much for you answers..

Not able to start Extract

Dear All ,
I am not able to start one of my extract and ggate error log shows the following message.
2013-01-22 22:48:14 ERROR OGG-01668 Oracle GoldenGate Capture for Oracle, e1_*.prm: PROCESS ABENDING
I have inadvertently deleted some archive files and my extract is expecting files from the deleted ones.Unfortunately i didn't took backup also. Please help me to start extract.
Is there any way out to start from the current arhive logs and make the destination in synch.Kindly post your comments
How out of sync is your data? One option is to rebuild the target. Without the missing ARLs, you're hosed anyway, unless you don't care or need the data consistency.
Good opportunity to have or develop an emergency rebuild script and obey files and use those to perform disaster recovery as it relates to GoldenGate. 
Thanks for the reply
Rebuild target means - > export /import and rebuild indexes ?
My data is out of synch for 32 hrs . Could you please let me know ? What will be the implication if i execute the command
alter extract begin now ? 
Well, you could have massive amounts of ORA-1403s - no data found. An index problem in of itself isn't necessarily going to kill you, but missing data certainly will (assuming you do not have any exception handling).
How long did it take to build the target? Being out of sync that long, combined with the missing logs needed to get caught up in the first place is kind of a show stopper. 
Well Steve in order to rebuild the target i may need to spend a couple of hours . Is there any way out to start the replication with the current archive logs ? 
You can alter extract and have it start "now" - but when you hit an operation where there is missing data on the target, you need to incorporate exception handling. Otherwise, every time you hit the error, it's game over. Find and fix the error, start again. How many times do you want to repeat that drill? 
So Steve as a best practice we shall rebuild the target and take necessary steps not to repeat the incident like archive deleting . Any way thanks for the guidance 
Beginning in OGG version 11.1 for Oracle databases 10.2 and greater, there is a means to register an Extract with the database to prevent RMAN from deleting archive logs that the Extract needs. Search 'logretention' in the OGG Reference guide.

Golden Gate Parameter File Maintenance

Hi Team, We are working on Golden Gate implementation and Its new to me. wanted to understand regarding the parameter files. As we understand the Parameter files should be under the following the path. /ggadmin/gg_home/dirprm This path has the mgr.prm and all other prm files relate to extration and replication. In our case the mgr.prm maintenance will be done by the DBA and all the extration & replication replication prm files will be maintained by developer team. So It is possible to have mgr.prm and replicate and extraction prm at different folder. If yes, where to make the changes to specifiy the location of these files. Regards & thanks in advance
Hi, Yes, You can create a parameter file in a directory other than dirprm, but you also must specify the full path name with the PARAMS option of the ADD EXTRACT or ADD REPLICAT command when you create your process groups. Once paired with an Extract or Replicat group, a parameter file must remain in its original location for Oracle GoldenGate to operate properly once processing has started.But I think it is not possible to change the Manager Parameter file to different location.Please refer the below document for more details,,Veera
Hi Veera, Your input is very helpful. Just once clarification. Above documentation is for version 12. Do you know if the same thing will work for 11 also. RegardsMohit
Hi , Yes it works for GoldenGate version 11 also... Regards, Veera  Sent from Samsung Mobile