4 ms·
It was definitely MinIO related, I probably should have made that clearer. We noticed that with zero fault tolerance, MinIO objects would randomly become corrup
by adamcharnock 9mo ago
It was definitely MinIO related, I probably should have made that clearer. We noticed that with zero fault tolerance, MinIO objects would randomly become corrupted, which MinIO would present as "you're making too many requests, please slow down". We were certainly not making too many requests.
- maxloh 9mo agoWhat is your plan after MinIO enters maintenance mode?
- adamcharnock 9mo agoWe're looking at alternatives, I've made some previous comments on that front. Sadly MinIO was the only option with sufficient performance for this particular situation. Thankfully we're not using any MinIO-specific features, so at least the migration path away is clear.
- fakebizprez 9mo agoCeph. The answer is always Ceph.