I have noticed on our servers that the WmReprository2 directory in particular has ‘very’ large files. Multiply this by 4 JVM’s per box and the slice (Unix) our IS sits on is always over 75%. This makes turning up logging for troubleshooting, core dumps, etc… very troublesome. I see with the repoui.sh that you can use a database for storage. What I am not finding anywhere is guidance on the particulars of said, and whether that is good/bad etc… Pro’s or con’s … best practices… etc… If anyone has any information or experience with said, it would be greatly appreciated.
Pros and Cons - Refer webMethods Integration Clustering Guide for DataBase and File Repository
wM Clustering Guide 6.5 does provide pro’s and con’s for DB vs file repo.
Each have their own set of trade offs.
In your case I would suggest deploying a JDBC compliant repository as it is more suitable for journaling larger chunks of data. Backing up of that data would be lot more convenient. One thing you have to make a note of is that this is going to be slower compared to a file based repository.
My two cents !
Hope this helps !