Hi, We have a large production cluster with Writeback Cache Tier. Recently, we observed that some of the OSDs were near full and got full from the Cache Tier and the cluster is in error state. The target_max_bytes was not set correctly and hence I think the flush and eviction never happened.
I wish to disable caching and remove the cache tier, but before that wish to flush all the contents of the cache. I have set the cache mode to "forward" as suggested here: https://access.redhat.com/documentation/en-us/red_hat_ceph_storage/1.2.3/html/storage_strategies/removing-a-cache-tier. But, rados df still shows same usage of cache tier. It's not automatically flushing and evicting. I tried "cache-try-flush", but this command gets blocked for some reason for many objects (although this is supposed to be non-blocking). How do I make sure all the objects are flushed to backing tier and objects evicted so that I can safely deprecate/remove the cache tier which is currently having many full and near full OSDs (in cache tier). -- Regards, Priya _______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-le...@ceph.io