sylar

Well, I think that did it! I completely missed this setting, which has been autocompleted with this high value when creating the share. Thanks!

The shares are being written from SMB. They are mounted from the proxmox host and bind mounted inside LXC containers. The same applies if I mount one of the share through SMB from my laptop (mac): a new folder is created directly on the data pool, not on the cache. And if I do a simple touch test inside e.g. /mnt/user/backup the file is still created on the data pool.

Thanks! As far as I can tell, all shares configured with cache as primary storage and data pool as secondary do not use cache at all (e.g. archives/, backup/, media/, etc.). unraid-diagnostics-20250720-1016.zip

Hi there, I have just finished my new reorganization of my disks and shares on my unraid (virtualized under proxmox) server (7.1.4 version). I'm running a zfs cache pool (2*500GB SSD in zfs mirror) + a data zfs pool (3*12TB in raidz1). Some of my shares are configured with the cache pool as primary storage, and the data pool for the secondary storage. What I am expecting for these shares is new data to be written to the cache, and then moved to the HDD data pool later according to mover schedule settings . But this is not what is happening: all new data are directly written to the data pool and the cache pool is not used at all ... Also, I'm using the "File Activity Monitoring" plugin, which displays recent file activity on my pools, and what I see is that the data are indeed directly written to /mnt/zfs_data/myshare, so indeed directly on my data pool and thus bypassing the cache. What am I missing here? I had no issue when I was using an unraid array instead of a zfs data pool (together with a btrfs cache pool at this time), so I'm wondering if this can be related to some settings of my zfs data and/or cache pool?