We have SegmentNotFound exceptions in our logs when trying to compact the repository, we are trying to resolve this by reverting the repo to the last known good state identified by running:
java -Xmx6000m -jar oak-run-*.jar check -d1 -p /path/to/crx-quickstart/repository/segmentstore
I can find out the point in time the repo will be reverted to by running this:
find . -type f -name "data*.tar" -exec sh -c "tar -tvf {} |grep 1ee25c3a-f54a-49f9-ae19-2b177db3b83b:259536" \; -print
Output:
-r-------- 0/0 3168 2018-02-08 17:36 1ee25c3a-f54a-49f9-ae19-2b177db3b83b.cf688f27
./data05499a.tar
But my question is - how do I know from this output what data specifically will be lost, ie will it be all changes made to the Author from 2018-02-08 17:36? Could we potentially lose pages added since that point? Or is just versioned data that is lost?
We have a very large repo (>400G) so want to be clearer on the impact before running. If I look for the id it has identified as the last good state in the journal.log there are a lot of ids below it which I would be removing, but I am not sure what they mean.
