Hello, 

 Have you downloaded the MIB files for your firmware release?  

 What are you trying to accomplish?    The in-use space only reflects written blocks.  If the host OS does not send the UNMAP command or the volume is replicated the in-use space will not necessarily reflect the actual in-use space as seen by the hosts. 

 Lastly there is member used space and allocated space.   So once you compile in the MIBs make sure which value you are using.  

  Regards, 

Don 

 

Hi There,

 

Yes we have downloaded the MIB's and did an SNMP walk on the EQL. We have have a Hyper-V cluster network that has access the volumes from the EQL. If a volume fills up is suspends the Hyper-V nodes that have storage on those volumes. We need to know when the volumes are nearing capacity and use the OID to trigger the alert with our management software

 

Thanks

Hello, 

 Are you oversubscribing your volumes?   If not, then you can use 100% of a volume.  EQL allocates in 15MB pages.  But will re-use pages as needed since it's the host that controls what blocks to use and re-use. 

 So the volume should not be getting suspended.  If you have not oversubcribed your array then you can turn off in-use space warning and thresholds by setting  them to 100%. 

 Regards, 

Don 

 

 

Hi Don,

The issue is Cluster side and not on the EQL, its inbuilt safety protocol suspends the cluster nodes when it views the storage as nearing capacity. Unfortunately Cluster Manager has no native alert functions to trigger a warning that the volume it is using is reaching capacity.

Thanks

Robin

Hello Robin, 

 Since the OS knows how much is actually in-use vs. free I am pretty sure what you are seeing is the SCSI based thin provisioned volume features.   Where the storage can tell the initiator that the ACTUAL amount of storage is running out.  In other words an over provisioned device.  Which allows the host to pause VMs to preserve them.  For VMware this is included as part of the VAAI API.  I forget what MS calls it.  Part they are part of the T10 SCSI spec.  

 So if a volume hits the threshold defined on the EQL volume it will alert the host via a SCSI sense code.  Causing the host to react.   

  Check the settings on your volumes and if you are not overprovisioning your pool, then set the values to 100%. 

 Try it on a very small test volume using a clone of some VMs. 

 Regards, 

Don