Hi, thank you for datahike, I am running it for years now ;-) I could use some advice. I inadvertently (in hotfix panic mode, but it turned out to be the reverse proxy) updated datahike from 0.4.1484 to 0.4.1494 in production. So the data on disk was written with 0.4.1484 with the new code being 0.4.1494. After sending a zfs snapshot of the data on disk to my dev machine I found out: 1. Some or all new data is missing from disk (memory still works). 2. http://2.in production :db/retract does not work any more. Luckily I have an nrepl into the production server so I can run an export, and an extra transaction-log so I can replay them if needed. The solution seems clear to me: export from the running system and import back, then replay transactions that are missing. Question I would like some advice on: • Which version do I choose to import into? Do I go back to an earlier version, if so which one or do I go to 0.5.1506? • I am not using any of the new data types, so I use the old import and export into readable files I can edit. Is this still OK or do I need to switch to binary export? • Am I missing something I should be aware of? Thank you in advance! PS: this is a long running db with one handful of users typing/clicking data into it.