On running oak run check checkpoint not found

[NituDh]

HTTP ERROR: 503

Problem accessing /. Reason:

    AuthenticationSupport service missing. Cannot authenticate request.

I am getting 503 Authentication handler missing.

The issue was in running server, jpa log file wasn't rotating and it took around 650 GB and aem repository get corrupted. The configuration is not upgraded its

AEM 6.4, Java 8 and oak-run-1.8.2 on Premise. I have seen some previous question which were giving good revision by running 

java -Xmx6000m -jar oak-run-*.jar check crx-quickstart/repository/segmentstore/

but in my case I get the output which is added in screenshot. It is like a deadlock for me I have tried a lot of things but nothing seems working Can some one guide me to get last good revision so that I can run recover or compact activity

[HrishikeshKagne · Accepted solution]

Hi @NituDh ,

Your AEM instance is failing because the repository has become corrupted (likely from the huge unrotated JPA log), so the Oak segment store cannot find a valid checkpoint to start from. That’s why oak-run check reports nothing usable and AEM throws the “AuthenticationSupport service missing” 503 error. In this state, you can’t “fix” the current repository — instead you need to restore from the last known good backup (filesystem or datastore backup). If no backup exists, the only option is to attempt an oak-run offline compaction (oak-run compact) or a checkpoints rm-unreferenced cleanup to try to free it, but recovery is not guaranteed. In practice, the safe path is restoring a valid backup; otherwise, data loss is very likely.

Thanks!

[NituDh]

I have created new publish instance. nothing worked in my case.

[SantoshSai]

Hi @NituDh,

Looks like your AEM could not bootstrap its repository because the Oak Segment Store is corrupted. The oak-run check output confirms this: multiple SegmentNotFoundException and no valid checkpoints found, which means the repository state cannot be reconstructed normally.

Here are the steps you can take:

1. Stop AEM immediately
   Do not attempt to restart repeatedly, as this may worsen corruption.

2. Take a full backup
   Even if the repository is corrupted, back up your entire crx-quickstart/repository directory and datastore before doing anything else. This way, you can retry recovery attempts without losing further data.

3. Check available checkpoints
   Run:

   java -Xmx6g -jar oak-run-*.jar checkpoints crx-quickstart/repository/segmentstore/

   If valid checkpoints are listed, you may be able to roll back to one of them with:

   java -Xmx6g -jar oak-run-*.jar checkpoints crx-quickstart/repository/segmentstore/ rm-unreferenced

   This removes broken/unreferenced checkpoints.

4. Try an offline compaction
   Run:

5. java -Xmx6g -jar oak-run-*.jar compact crx-quickstart/repository/segmentstore/

   Sometimes compaction can rebuild the store enough to allow AEM to start again.

6. If compaction fails

   • Attempt recovery with oak-run restore if you have a datastore backup.

   • If you have FileDataStore / S3 / Azure Blob Store, check that the binaries are intact and accessible.

7. If no recovery works
   Unfortunately, the repository is beyond repair. The only reliable option is restoring from your last filesystem or datastore backup. If you do not have one, you may need to rebuild the repository (re-import content packages, reconfigure users, etc.).

[kautuk_sahni]

@NituDh Just checking in. Were you able to resolve your issue? We’d love to hear how things worked out. If the suggestions above helped, marking a response as correct can guide others with similar questions. And if you found another solution, feel free to share it — your insights could benefit the community. Thanks again for being part of the conversation!

[NituDh]

I have created new instance of publish, nothing worked in my case