ONTAP Hardware

Flexclone process can take a long time


I'm curious to know what factors can impact the speed of the flexclone and volume creation process?

We have a Powershell script which takes a flexclone of an SQL database every 30 minutes. By running "snap list <volname>" I can see the snapshot is taken instantly - as expected. However the volume creation from that snapshot can take anywhere from 2 minutes to 20 minutes. By the time the LUN contained within the flexcloned volume is mounted to the server, the script can sometimes have been running for 35-40 minutes which means the next job is skipped.

During working hours, CPU usage on the filer is around 60% on average, so I doubt this is having such an impact on time to create a volume? There are no snapmirrors running at the same time and de-dup operations are scheduled out of hours....

Any thoughts?




this is really hard to say. if it's split, easy to understand. it's copying data out from snapshot. but i think u r not talking about that.

so it is best to open a case and collect some perfstat to analyse it.



I noticed there were a lot of 'dead' FC paths on the ESXi host, rescanning the HBAs removed these and seemed to speed up the process of mounting a flex clone to the Guest OS. Not sure why!


I ve the same problem with a FAS 3240 with Dataontap 8.1 ( and 8.1P3 )  .Randomly , flexclone process take a long time to complete . We have experimented this problem with the following command  ( vol clone create , vol offline <clone_volume> , vol destroy <clone_volume> ) .

Ha ve you solve  this problem , If yes  , I'm very very interested to know that you have do