ONTAP Discussions

cDOT Snapmirror : Current Transfer Priority


In the good old days of 7-mode if the throughput of a snapmirror job slowed down to a trickle, throughput could be restored by aborting the transfer and restarting the job. This was very useful after the clearing of network events and performance spikes on the source/target controllers. Cluster Mode, does not seem to allow this workaround. I have been looking into the data point "Current Transfer Priority" and have precious little other than: it can be set at the beginning of a job (option are normal and low) and the 'snapmirror show' command has this a parameter [-current-transfer-priority {low|normal}].


Q) What is the impact of a snapmirror job changing “Current Transfer Priority” from normal to low?

Q) What are the triggers that cause the “Current Transfer Priority” of a snampmirror job to transition from normal to low?

Q) Is there any situation where a “Current Transfer Priority” will ever transition from low to normal?

Q) What is manual process / workaround for changing a snapmirror job “Current Transfer Priority” from low to normal? If any?






My name is Chriz Ott, I'm a principle architect for NetApp.
Thank you very much for your question, as a matter of fact i have not found an answer to your question in our official documentation.

Having said that my understanding of this option is that it is for manual priorisation of one snapmirror transfer over another.
"Starting with Data ONTAP 8.2, we have the ability to set a priority on SnapMirror relationships. Currently we support two levels: normal and low. Normal transfers are given preference over low-priority transfers."

I understand, that you are looking for a way to "recover" a slow snapmirror transfer. Is that something you are currently experiencing or are you simply curious how you would do that?
If the first, please open a support case and come back to me with the number via mail, so we can get an SLA attached to your request.

Cheers chriz

P.s. If you think this answer is usefull, please make use of the KUDO or correct answer feature, to make it easier for others to find.
P.S. if you feel this post is useful, please KUDO or “accept as a solution" so other people may find it faster.




It doesn't look like anything has changed with this in 8.3 in terms of changing the priority.


The quick way I've worked around it is to run a snapmirror abort --> snapmirror break --> snapmirror resync sequence to restart the job at the 'normal' priority level.


The other question on the priority is can the system change/set the level whist it is under significant load? We seem to have some jobs that were at low priority that were not (intentionally!) set at that level by us as administrators.





Did you ever find a resolution to your questions? I would like to know as well. Some jobs that are set to Normal priority bump down to Low, and restarting the job doesn't correct it.


No I did not receive any answers, online-or-off, to my specific questions beyond what information was already stated in the original post.    They may feel that these operation modes and triggers are a part of their secret sauce and don’t want to share.   I don’t do the day-to-day monitoring anymore, so have not been chasing this information.


Hi there,


I've not recieved a solution to this issue either.


I did have a discussion with NetApp engineering in which I was advised there are essentially a two job queues, one at normal priority and one at low priority, with a limited number of 'slots' available to run a job. Those jobs at normal prioritry are processed in preference to the jobs at low priortiy however once there is a free slot and all normal priority jobs have completed, low priority jobs should be processed without being throttled in any way.


I've no way of verifying that is correct and this obviously doesn't answer the question of how/why jobs get reduced to low priority or how they can be returned to normal priority..





My name is Andreas and  I'm a TSE for NetApp.


The reason why every Running Transfer is showing up as low sch Priority is because this is how the Schedule is scribted in the Code, it will not start any Update of a low transfer if there is still normal ones in the queue.






[-transfer-priority {low|normal}] - Transfer Scheduling Priority

Specifies the priority at which a transfer runs. The supported values are normal or low. The normal transfers are scheduled before the low priority transfers. The default is normal.
the transfer-priority is not  a Throttenling tool for the Transfer, it is just a tool to prioritice Transfers if there is a lag of free transfer slots.
if you feel this post is useful, please KUDO or “accept as a solution" so other people may find it faster.