With the changes in ISPN-15172 it added better handling of compactor and additional checks for file changes. However, it does not account for compaction starting for a log file and a new entry being written before it completes its compaction check. This means at the end of the it may log a single message such as

WARN [blocking-thread--p3-t9] o.infinispan.persistence.sifs.Compactor  : ISPN029022: Compaction skipping a corrupted entry for key null, at 0:712444|712444 that doesn't have enough bytes for header [keyLength=26, valueLength=0, metadataLength=0, internalMetadataLength=0,seqId=7364, expiration=-1]

Note the size of the file is the same as the offset. The size is only captured at the beginning of compaction and thus may change later. We should stop doing expiration compaction only up to the file size to prevent this.

Note that this does not cause any data consistency or processing issues, is just an annoying WARN message in the log files.