I started full backup, data copy finished pretty quickly (in 3 hours) but look like xtrabackup_logfile had difficulty to catch up. The purpose of full backup is to establish a new slave. As long as the bin log number and position are known, slave should be okay to start. There will be some lag any way. Is it possible to set a stop point so I can finish backup?
Does you mean that backup is taken is finished (in 3 hours) and problem with prepare stage ? Can you share the backup logfile to understand where the problem lies.
Look likes your database is huge or lot of write activity is happening on database during course of backup. By the way, xtrabackup stores the binary log coordinates in xtrabackup_binlog_info file if you initiate backup from master server to create new slave. Check this link to create slave [url]Percona XtraBackup
If you are creating new slave from existing slave then append --slave-info parameter in your backup command. --slave-info parameter store the binary log filename and position for new slave to start. For more info on it check [url]The innobackupex Option Reference
Further, xtrabackup_logfile is use to store changes in InnoDB transaction logs. Whenever data is written to InnoDB tables during backup all changes are written to xtrabackup_logfile in the backup target directory. This is necessary because the backup might take a long time, and the recovery process needs all of the log file entries from the beginning to the end of the backup. For more information check the documentation [url]Percona XtraBackup
Backup time is dependent on many factors including database size, server load, hardware etc. I would suggest to run backup on off peak low traffic time. If you stop the backup in middle backup is not in consistent state and not usable. Backup should be completed successfully including prepare stage in order to use it for restore.
Hope that helps.