Hello. I'm trying to solidify my understanding of space reclamation for LUNs connected to VMware and could use some help. Our VMware environment is vCenter 6.0U1 with mostly ESXi 5.0U3 hosts. They are connected via iSCSI to LUNS on a FAS8020 running CDoT 8.3.1P1. One datastore to one LUN. The datastore is the same size as the LUN in most cases. Smaller in others. We do not have space-allocation enabled on our LUNs, so we are seeing the situation where the datastore free space in VMware is different than the LUN free space.
Here is my question: Is it possible that the LUN will fill up before the datastore, thereby causing the LUN to offline? If so, how is that possible? Outside of enabling space-allocation (which requires LUN downtime) is there a way to prevent it? Thanks.
Is it possible that the LUN will fill up before the datastore, thereby causing the LUN to offline? Answer is NO. The LUN contains exactly what VMware puts in the datastore. If VMFS thinks the file system is full, then the LUN is full. That may affect VMs depending on their config (thick vs. thin), but won’t take the LUN offline. What can take the LUN offline is space outside the LUN either in the volume (consumed by NetApp snapshots) or the aggr if the volume and LUN are thin provisioned.
The following TRs give detail information on these topics. TR-3749, 4068.
Hope this helps.
If this post resolved your issue, help others by selecting ACCEPT AS SOLUTION or adding a KUDO.