1
votes

I have a large forest that I want to split in two. So I created a new forest, then attached it to the same database as the first one, and I saw that rebalancing had begun (using the default "Bucket" approach), but eventually I noticed that it had stopped moving documents to the new forest. When I checked in the Admin UI, I saw this message...

There is currently an XDMP-FORESTERR:

Error in rebalance of forest ESI-EDGAR: XDMP-JNLTXNFULL: The journal has exceeded its limit for in-process transaction records: txn_size=46.813MB, current_size=0MB, limit=42.625MB exception. Information on this page may be missing.

Any ideas on how to get around this? And why it happened in the first place? I have plenty of disk space, so that it not the problem.

1

1 Answers

1
votes

If you look at Monitoring History, how does your resource utilization look? Are you pushing the limits on CPU, Memory, or IO?

Try increasing the journal size, and/or maybe lower the throttle on the rebalancer.

The smaller the number for rebalancer throttle, the slower it will go. That may provide enough time to in-flight transactions to commit and complete, and avoid exceeding the journal limit.

https://docs.marklogic.com/messages/XDMP-en/XDMP-JNLTXNFULL

Message Text

The journal has exceeded its limit for in-process transaction records: txn_size=variable1MB, current_size=variable2MB, limit=variable3MB

Cause

A fixed limit of 1/16th of the journal size is a maximum for information about prepared uncommitted transactions. This size limit has been exceeded by the transaction being committed. No more transactions will be allowed to move to being prepared, until the transactions currently using this space have committed or rolled back successfully.

Response

Explicitly commit or rollback XA transactions waiting in a prepared state. Increase your journal size to allow more space for prepared uncommitted transactions.