OSSV Throttle Policy

[paul_wolf]

I'm trying to get a better understanding of the throttle policy and how it is actually applied. Specifically, we have a policy with 13 clients in it, the throttle policy is set to 1150 KB.  The snapvault log on only one of the clients is showing that the -k setting being applied is 89 KB (which just happens to be 1150 divided by 13).  I have so far been upable to a) find specific documentation that talks about that or b) get an answer from NetApp support in the last three days on if it operates like that.

I would like to get confirmation on that so I can then change how we are setting the bandwidth as my transfers are taking way too long.

Thanks

Paul

[crocker]

  is this one that your team handles or should the question be asked in the NetApp Support Community

[columbus_admin]

Hi Paul,

      The protection manager administration guide states:

"Alloting limited bandwidth to a throttle period causes any backup or mirror

operations that start within period to execute using the alloted bandwidth; On

the other hand, backup or mirror operations that are in progress when a new

period of limited bandwidth starts continue to completion using their original

bandwidth allotment."

So that would mean (number of jobs/bandwidth) at the start of the job is the "allotment" as stated above.

- Scott

[dnewby]

I've expressed my opinion about how pointless this feature is to product development and I encourage everyone else to as well.  What the point of a throttle policy that only works at the start of a job?  The whole point of a scheduled throttling policy is to limit bandwidth during certain periods and then raise the limit on off peak hours.  I ended up running a rsh script as a scheduled task to change the options replication.throttle.incoming.max_kbs at scheduled times. 

[columbus_admin]

I agree, it should be the bandwidth available with dynamic adjustments.  Even if only when a job starts of finishes, it would help out greatly.  We did the script based controls at a previous employer as well to get around it.

- Scott

[paul_wolf]

Thanks for the feedback that confirmed what I suspected. 

@dnewby - I also don't understand the point of setting a throttle policy for x clients that allocates as throttle/# of clients and never changes even as individual sessions within that policy completed. That is no where near what I would even consider dynamic.

Looks like I'm going to remove the throttle policy and go with client side throttle settings in wan.cfg, not a huge fan of this approach as I have 64 remote clients but it seems to be the best available option that I can come up with.

Again, thanks for the feedback!

Paul