On 8/8/22 20:30, Mark Nelson wrote:
Hi Folks,
We are trying to get a sense for how many people are using
bluestore_compression_mode or the per-pool compression_mode options
(these were introduced early in bluestore's life, but afaik may not
widely be used). We might be able to reduce complexity in bluestore's
blob code if we could do compression in some other fashion, so we are
trying to get a sense of whether or not it's something worth looking
into more.
We also use per-pool compression (snappy) and have CephFS metadata pool
uncompressed. Data / metadata pools share the same OSDs in our cluster.
A datapoint on compression settings: In Octopus (and newer I guess) data
is compressed when it is enabled on the pool and set to "aggressive"
(might be same for force, but haven't tried). We have no bluestore
compression settings defined for OSDs (not in config files or Ceph
config database), so the default applies which is
"bluestore_compression_mode": "none". We use a mix of SSD / NVMe. NVMe
seems to use the value for SSD: bluestore_compression_min_blob_size_ssd.
With "bluestore_min_alloc_size_ssd" set to 4096 we obtain a compress
ratio of 2.
Would it make sense to have a "bluestore_compression_min_blob_size_nvme"
as the performance of such drives is often much higher than SATA / SAS
SSDs (at least have more bandwith available)?
Gr. Stefan
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
