You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, I’ve been testing Engraphis as the persistent memory provider for Hermes and have run into what may be either an Engraphis bug or an issue with my integration/configuration.
I’m not sure which, so I wanted to check with you before spending time trying to repair the database myself.
2026-08-16 21:27:24 Engraphis provider is unavailable (EncryptionError)
2026-08-16 21:46:14 Engraphis provider is unavailable (EncryptionError)
There may also be a related earlier observation involving the dashboard.
While the dashboard was running, I observed the SQLite WAL/SHM file descriptors becoming shown as "(deleted)" while still open. The Engraphis dashboard service is currently not running. The Hermes Desktop Engraphis Agent Plugin toggle was already off and has not been changed.
I’m not assuming that caused the later database failure, but I’m mentioning it because it occurred beforehand.
I have not attempted to repair, migrate or restore the database since the EncryptionError appeared because I didn’t want to destroy useful diagnostic evidence.
I also have known-good backups of the encrypted DB, SQLCipher key and configuration from before the failure.
Before I go further, could you advise:
Does "receipt chain has no unique structural head" indicate a known Engraphis failure mode?
Could the later EncryptionError plausibly be related, or would you suspect a separate SQLCipher/key/configuration problem?
Are there changes in current main/newer releases that may address this?
Would you recommend restoring the known-good DB, upgrading/migrating the current DB, or preserving it for diagnosis first?
Is there any specific diagnostic output you would like me to collect before changing anything?
Happy to provide logs, hashes, schema information or other diagnostic output if useful.
I’m deliberately holding off on deeper debugging because I’d rather check whether this is already understood upstream before trying to reverse-engineer or repair Engraphis internals myself.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Hi, I’ve been testing Engraphis as the persistent memory provider for Hermes and have run into what may be either an Engraphis bug or an issue with my integration/configuration.
I’m not sure which, so I wanted to check with you before spending time trying to repair the database myself.
Environment:
Engraphis 1.6 from commit a23850b
Hermes v0.20.1
Python 3.11.15
Ubuntu 24.04.4 LTS
SQLCipher-enabled database
Hermes memory.provider = engraphis
Initial acceptance testing passed successfully, including:
write
recall
supersession
encrypted database operation
The system then developed the following sequence of failures.
First:
sqlite3.IntegrityError: receipt chain has no unique structural head
After this appeared, both memory writes and memory prefetch/recall began failing.
Relevant log entries:
2026-08-16 19:52:36 Engraphis prefetch failed (IntegrityError)
2026-08-16 19:54:42 Engraphis turn persistence skipped (IntegrityError)
2026-08-16 20:56:52 Engraphis turn persistence skipped (IntegrityError)
Later the provider became unavailable with:
EncryptionError
Log entries:
2026-08-16 21:27:24 Engraphis provider is unavailable (EncryptionError)
2026-08-16 21:46:14 Engraphis provider is unavailable (EncryptionError)
There may also be a related earlier observation involving the dashboard.
While the dashboard was running, I observed the SQLite WAL/SHM file descriptors becoming shown as "(deleted)" while still open. The Engraphis dashboard service is currently not running. The Hermes Desktop Engraphis Agent Plugin toggle was already off and has not been changed.
I’m not assuming that caused the later database failure, but I’m mentioning it because it occurred beforehand.
I have not attempted to repair, migrate or restore the database since the EncryptionError appeared because I didn’t want to destroy useful diagnostic evidence.
I also have known-good backups of the encrypted DB, SQLCipher key and configuration from before the failure.
Before I go further, could you advise:
Does "receipt chain has no unique structural head" indicate a known Engraphis failure mode?
Could the later EncryptionError plausibly be related, or would you suspect a separate SQLCipher/key/configuration problem?
Are there changes in current main/newer releases that may address this?
Would you recommend restoring the known-good DB, upgrading/migrating the current DB, or preserving it for diagnosis first?
Is there any specific diagnostic output you would like me to collect before changing anything?
Happy to provide logs, hashes, schema information or other diagnostic output if useful.
I’m deliberately holding off on deeper debugging because I’d rather check whether this is already understood upstream before trying to reverse-engineer or repair Engraphis internals myself.
All reactions