Hi everyone. I added a lot more storage to our cluster, and we now have a lot of slower hard drives that could contain archival data. So I thought setting up a cache tier for the fast drives should be a good idea. We want to retain data for about a week in the cache pool as the data could be interesting for at least a week, and then it will probably rarely be accessed. You can see my config below. Would you change any of these values? Do you have any better suggestions? This is my interpretation: I keep the data for 30 minutes before writing it back to the slower drives. We will retain the data for seven days before we evict them from the cache. All accessed objects will be promoted to cache Now we only have the hit_set values left, and I'm not sure the right strategy here. I've read a paper from SUSE where they explain it. https://documentation.suse.com/ses/6/html/ses-all/cha-ceph-tiered.html#ses-tiered-hitset And my understanding from that paper I should change the hit_set_count to 42, so I will have a hit set for the entire duration I want to keep data in my cache pool. Is this correct? We have a lot of overhead when it comes to disk, and the hit sets should be pretty small in this context, so I don't think there should be a reason for not keeping 42 of them. --------------------------------------------- ceph osd pool set cephfs_data_cache cache_min_flush_age 1800 ceph osd pool set cephfs_data_cache cache_min_evict_age 604800 ceph osd pool set cephfs_data_cache hit_set_count 12 ceph osd pool set cephfs_data_cache hit_set_period 14400 ceph osd pool set cephfs_data_cache hit_set_fpp 0.01 ceph osd pool set cephfs_data_cache min_write_recency_for_promote 0 ceph osd pool set cephfs_data_cache min_read_recency_for_promote 0 --------------------------------------------- I appreciate any help you can provide. Best regards Daniel _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |