zfs dataset constantly unmounting itself after some time, others still available

Question

I have a linux box, that i putty into, which is my backup system. It has a zfs linear span array with 3 2TB drives. Every night i use freefilesync (great program) to sync files to this mount, which is mapped as a network drive.

$ zfs list
NAME                                USED  AVAIL  REFER  MOUNTPOINT
san                                3.31T  2.04T  2.87M  /san
san/vault                          3.31T  2.04T   136K  /san/vault
san/vault/falcon                    171G  2.04T   100K  /san/vault/falcon
san/vault/falcon/snapshots          171G  2.04T   171G  /san/vault/falcon/snapshots
san/vault/falcon/version            160K  2.04T    96K  /san/vault/falcon/version
san/vault/gyrfalcon                 564K  2.04T   132K  /san/vault/gyrfalcon
san/vault/gyrfalcon/snapshots       184K  2.04T   120K  /san/vault/gyrfalcon/snapshots
san/vault/gyrfalcon/version         184K  2.04T   120K  /san/vault/gyrfalcon/version
san/vault/osprey                    170G  2.04T   170G  /san/vault/osprey
san/vault/osprey/snapshots         24.2M  2.04T  24.2M  /san/vault/osprey/snapshots
san/vault/osprey/version            120K  2.04T   120K  /san/vault/osprey/version
san/vault/redtail                  2.98T  2.04T  17.2M  /san/vault/redtail
san/vault/redtail/c                 777M  2.04T  72.9M  /san/vault/redtail/c
san/vault/redtail/c/AMD            4.44M  2.04T  4.24M  /san/vault/redtail/c/AMD
san/vault/redtail/c/Users           699M  2.04T   694M  /san/vault/redtail/c/Users
san/vault/redtail/d                1.59T  2.04T   124K  /san/vault/redtail/d
san/vault/redtail/d/UserFiles      1.59T  2.04T  1.59T  /san/vault/redtail/d/UserFiles
san/vault/redtail/d/archive         283M  2.04T   283M  /san/vault/redtail/d/archive
san/vault/redtail/e                1.34T  2.04T   124K  /san/vault/redtail/e
san/vault/redtail/e/PublicArchive  1.34T  2.04T  1.34T  /san/vault/redtail/e/PublicArchive
san/vault/redtail/e/archive         283M  2.04T   283M  /san/vault/redtail/e/archive
san/vault/redtail/snapshots         184K  2.04T   120K  /san/vault/redtail/snapshots
san/vault/redtail/version          44.3G  2.04T  43.9G  /san /vault/redtail/version

When looking at the linux, via putty, the mounts are there one minute, then a few minutes later they are gone(zfs list always shows them, you have to traverse into them to see if they have been unmounted, they will be empty, or not in the parent dir at all). These are datasets that are losing their mounts.

san/vault/redtail is empty almost every time when i come back the next morning, or a couple hours later, just after freefilesync starts its synch, but before i can start moving files.

I have tried to export and import , same problem happens still. My data is still in tact...

this command will fix all of them momentarily again(for an unknown duration).. zfs mount -a

This all started (coulpe weeks ago) after i made quite a few child datasets inside the parnt san, where before it was only the san dataset parent(i NEVER had any problems), no child datasets existed when all the data was layed in there. I have since moved out the data, made the datases, and moved back int he data.

My machine that keeps network traversing it to synchronize, the mount gets lost in the duration and my backups have yet to finish in over a month now almost. Before it can tally the files (say 15 minutes to an hour) the fricken mount is gone again, and it hangs. Back ups are in limbo.

I may have fubar'd something when i crated the datasets. I had to delete a few after because i wasnt happy with them. The data wasnt either visible or what, but once i had all said and done, things looked perfect after several reboots and double checks!!

After this, things seemed good, but I looked into nto one of its child data sets san/vault/redtail/c and that fricker is empty too, i think it came unmounted.

before i destroy shit. i need to know whats going on. This data is a duplicate backup of a current healthy system, but this backup is the only, therefore is at the mercy of the health of the source drives.. So i cant afford it to be offline at all..

One more note on SNAPSHOT's. I made a snapshot, for the first time, just prior to the seeming brakage. could that have caused it? Can i possibly use this snapshot to fix something? Is this coincidence, or symptomatic?

see this post, Bulk remove a large directory on a ZFS without traversing it recursively if you want information as to why i created a bunch of datasets, with children...

Edit: The question has just been rewritten, due to lack of attention on this post, please read it again..