Is it normal, a feature, or bug in V6 to have the min free space of a share to include the size of the cache drive if set to use it?  I have 150GB min free space set on a share with 125GB free on the cache drive (and 20GB as cache min free space setting).  When I write to the share (/mnt/user/some_share) I get a syslog message saying "...shfs/user: share cache full" and the file goes dirtectly to the array spinning up the parity drive and the file is less then 5GB.  It did not do this in V5 as I could write until the cache was full before being written directly to the array.

bump...no one else has this issue, or no one else has a small cache drive?

Just saw this.  In the past, one of the ways this could occur was if the user misinterpreted the units being used, when they entered a min value.  In other words, a user might *think* they were entering something equivalent to 20GB, but actually evaluated to 200GB.  It's the only idea I can come up with...

Thanks for the reply.  The units were correct as I though the same thing.  To verify I entered a size of min free space for a share that was more than what was available on the cache drive ("125GB" with 120GB actual free at the time on the cache drive) and it wrote the file (~6.5GB) directly to the array and reported the following in syslog "Oct 20 23:22:22 Tower shfs/user: share cache full".  I reduced the min free space for the share to a size smaller than what was free on the cache drive ("100GB") and it wrote it to the cache drive with no problem.  As I started adding more files and reached the 100GB actual free on the cache drive the next file went to the array with another "share cache full" line in syslog.  It didn't do this in 5.0 as I could let it run completely out of space on cache before writing directly to the array.  The cache drive min free space setting is 20GB.  If need diagnostics they are attached to previous post on this http://lime-technology.com/forum/index.php?topic=43153.msg411876#msg411876