nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Oracle Communities

Oracle Analytics and AI

Child Item

WebCat replication in OBIEE 12c

AM_1

Hi gurus!

We have two OBIEE 12c servers in different sites - one in US and another one in UK.

The users usually connect to the US site, but UK site must be up to date in case of downtime in US.

We are looking for a way to seamlessly and automatically synchronize the web catalogs.

For example, once a catalog in US is modified, e.g. a new report was saved, the report should be automatically replicated to another site.

According to Oracle support, there might be few techniques which can synchronize the catalogs automatically but those are not tested by oracle and there is no documentation.

Did anybody implemented such functionality?

For example, since our servers are Linux machines, can we simply use Linux rsync on webcat folder / subfolders?

Thanks in advance,

Alex

Find more posts tagged with

Accepted answers

All comments

Joel Acha

You can probably achieve this using a mechanism from Linux like rsync as you mentioned to identify if any catalog changes have been made.

See - Oracle Support Document 2067527.1 (OBIEE 12c: Quick Reference For Architecture, Lifecycle Management and File Locations) can be found at: https://support.oracle.com/epmos/faces/DocumentDisplay?id=2067527.1 for information of the OBIEE 12c catalog location .

You can then use the BI Archive (BAR) files as a means of exporting importing the catalog -

Oracle Support Document 2067513.1 (OBIEE 12c : What Are The BI Archive (.BAR) Files And Are There Any Predefined Default BAR files Available?) can be found at: https://support.oracle.com/epmos/faces/DocumentDisplay?id=2067513.1

AM_1

Hi Joel,

Thanks for answering.

I am currently using bar files.

However the process is pretty heavy since we have a large catalog.

Also we are frequently getting strange errors during bar file import.

Since webcat is a regular Linux folder, I was wondering if it's possible just to replicate changed files from one server to another, probably with an exception to the system folder...

Thanks

Alex

Joel Acha

I haven't tried that and wouldn't recommend that approach.Thinking out aloud, you could create a python script (or any other programming language of your choice) which scans your catalog on the file system to detect any changes objects (possible looking at the timestamp?).

You can then use the runcat.sh utility to create an archive file for this object (an archive file per changed object) and then unarchive these into the instance that you're replicating to making sure that you unarchive the archive file into the right location of the replicated catalog.

AM_1

Joel,

Thanks again!

Another question: Doesn't the OBIEE cluster provide this kind of functionality?

Alex

Gianni Ceresa

Hi Alex,

I would say the cluster doesn't have any fancy thing for that as it is you supposed to deal with it when you setup the Singleton Data Directory (SDD).

https://docs.oracle.com/middleware/12212/biee/BIESG/GUID-60B2E4E4-11AC-444B-A9A6-50DF4CA0F6A0.htm#BIESG9273

PS: hoping I didn't got the thread wrong when quickly reading through

AM_1

Thanks, Gianni.

Unfortunately this doesn't look like a viable option in our case since the servers are not located in the same LAN (one - US, another - UK), and therefore there's no such shared disk that will be 100% of the time available to both instances to possible network issues.

Alex

Gianni Ceresa

Yeah, I posted more as reference but it was clearly not an option.

There is (maybe) a thing I read about in 11g (I guess, hoping it wasn't in 10g), an utility which role was exactly what you look for, sync catalogs between 2 separate instances. It was poorly or not documented and there wasn't much about it. I'm trying to find a reference or a link to it as it could give you something to maybe look for. If I ever manage to get my hands on that again I will post it back.

Rajesh Dhanapal

Hi Gianni,

SDD doen't work I have tried this option. In our scenario both sites are residing in the same country but different sites. But when you copy the ../bidata folder as specified in Oracle Doc to both the site and after restarting both instances, recently started OBIEE instance throws the following error:

OBIS server is in shutdown status and not able to login to analytics. [nQSError: 13042] Repository password is wrong

Gianni Ceresa

SDD is for a proper cluster (and has been posted as OP asked if cluster didn't have that functionality). If you follow the full doc on how to setup the cluster it must work (if it doesn't you better open a new thread describing your issue), at least many have it up and running without an issue.

It just isn't an option for the OP, but as he asked if cluster could have a functionality for that the answer is no because the cluster expect to work on SDD, which don't cover the need of the OP.