Hi King -

The EMC Isilon Guidelines for Large Workloads for OneFS 7.1.1 (docu54994) and the same document for OneFS 7.2, which has been renamed as the EMC Isilon OneFS 7.2 Technical Specifications Guide (docu56230), provide guidelines that may vary by many factors, including variables that are unique to your environment. The guidelines--best practices, really--are provided to help you achieve optimal cluster performance.

The recommended number of Microsoft Active Directory domains (5) and OneFS access zones (20) are guidelines only. The maximum number of supported access zones has yet to be established.

You're certainly free to exceed these guidelines, though doing so puts you outside the boundary of "best practice." If you choose to implement more than the recommended number of AD domains and/or access zones, share your experience with us here. And of course if you encounter any issues, call Support and we'll be happy to help.

Hello Tom,

I have been told that if these numbers are exceeded that the configuration will not be supported by EMC. Again, my understanding is that these numbers are not set in stone, but I could be told to go fly a kite if I need support on my cluster and these numbers are breached.

Please clarify or provide any information that I could reference to supports your statement below.

Thanks,

Wes

I heard that each zone use a chunk of memory true LSAS process. Is there a way to find out how much is use and when it wil became a problem for the cluster?

Hi Wes -

I confirmed that the guidelines are indeed guidelines--or, better put--best practices. The Introduction to the document states that "The figures are guidelines--they can vary by many factors, including variables that are unique to your environment." Throughout the document, you can also find specific statements that further clarify this, as in the following descriptions for NDMP connections per node and NFS exports per cluster:

Exceeding a guideline may affect system performance or system behavior, but at this time there is no hard-and-fast rule because of the many ways in which a feature may be exercised and respond to system load. If you choose to exceed a guideline by an unusual amount, I would suggest contacting Support or your Isilon account manager beforehand. I'd also recommend testing your configuration in a test environment, if possible, before implementing those same configurations on your production cluster.

Thanks,

tom

Hi Alain -

It currently isn't possible to determine how much lsass process is consumed by each access zone. I'll forward your question to the team, however, and see if we can offer some clarification.

Thanks,

tom