Prismo flagship instance suffers data loss, offers postmortem
Prismo, the lovable link aggregator platform, suffered a fatal data loss the other week when a migration and deployment went awry on the project’s flagship instance, prismo.news. The project’s lead developer issued an apology, along with a comprehensive post-mortem of the incident.
The post-mortem features the elements of a perfect storm: migrating to a new hosting service, forgetting to configure backups for it, a problematic merge during deployment, and a handful of other minor problems.
I deeply regret of what happened and i hope we can work together and rebuild this cool place without looking back. I’m so sorry this happened.
One challenge the developer has to face now is ensuring that the ids for posts and comments don’t cause conflicts with the rest of the fediverse, as they originally only used incremental integers for object IDs.
We’re gonna move from integer IDs to GUIDs which are a unique sequence of 32 chars (read more here). This way, even after starting from scratch, no duplication could happen.
We wish Prismo a speedy recovery, and appreciate the lead dev’s transparency in addressing what happened. Mistakes happen from time to time, and no developer or admin is immune from its effects.