Correction:


> 3) Will this auto migration session complete when it gets to the lower
> threshold (60%) or will it keep going until the pool is empty?
>

It will complete when it gets to the low threshold, or when it has migrated
all the data from the largest client.



On Thu, Oct 1, 2009 at 9:52 AM, Wanda Prather <wprather AT jasi DOT com> wrote:

> See answers below.
>>
>>
>> My questions:
>>
>> 1) Since this is an auto migration process kicked off by the thresholds,
>> does it have to complete before any new migration processes can start?
>>
>
> Yes (for the same pool)
>
>
>> Is there any way to change/override that if that is the case?
>
>
> You can try: update stgpool xxxxxxx migprocess=n+m
> where  n is the number of processes currently defined for the pool, and m
> is the additional processes you want to run.  That won't do any good,
> though, if what is in the pool is one big filespace all belonging to 1
> client.
>
>
>> Is only
>> one process running because the only data remaining in that pool is
>> likely, as far as I can see, from a single node?
>>
>
> Yes.
>
>>
>>
>> 2) Could I safely cancel this migration session, then restart four (or
>> six, or eight) new migration sessions?
>
>
> Yes.  Again, you won't get any more processes to run if all the data
> belongs to 1 client.
>
>>
>> 3) Will this auto migration session complete when it gets to the lower
>> threshold (60%) or will it keep going until the pool is empty?
>>
>
> It will complete when it gets to the low threshold, or when it has migrated
> all the data from the largest client, whichever comes first.
>
>>
>>
>>
>>
>>
>>
>>
>> ----------------
>>
>> Kevin Kinder
>>
>> State of WV
>>
>>
>>
>