Anyone want to ask me about NetApp 8.0x C-Mode?

kletzly · ‎2011-09-08

happy to dialogue about it

danielmorgenstern · ‎2011-09-09

Matt, I'm definitely interested. We're not ready for C-Mode yet but when we do, I'm interested in whether or not there will be a path from 8 running in 7-mode to move to C-Mode as well as what's needed to put existing storage systems into a C-Mode multiple node cluster.

kletzly · ‎2011-09-09

Today, there is not a "push-button" migration method that takes you automatically from 8.0 7-mode to 8.0 C-mode. However, you are not the first to ask for this capability, which NetApp is well aware of.

I am very excited about C-Mode. It is not brand new. It used to be NetApp's GX product. Can you believe that you can have as many as 24 controllers clustered together using C-Mode.

don_hulley · ‎2012-01-23

What is the process of moving from an existing HA-pair from say ONTAP 8.0 to 8.1 C-mode? Can you point me to any documentation on this?

kletzly · ‎2012-01-25

Hi Don

First question, are you going from 8.0 c-mode to 8.1 c-mode? Or 8.0 7-mode to 8.1 c-mode?

kletzly · ‎2012-01-25

We are working on another Tech OnTap Users Group meeting on Feb 23rd........ here are the topics!

Data ONTAP 8.1 Cluster-mode :

The What’s, When’s, How’s and Why’s

Come to this informative Tech OnTap Live event and learn all about NetApp’s Cluster-mode. Did you know that NetApp Data ONTAP 8 is available in (2) modes, 7-mode and Cluster-mode? Most members of our Users Group use 7-mode, but some members are using "C-Mode" as well. This TOTL session will allow us to discuss the product at length, understand how it differs from 7-mode, when it is an appropriate option, how one deploys and why it is so versatile as a go forward strategy for mature scale-up, scale-out and immortal storage. And, as always, TOTL is a an opportunity to network with others in the area. Attend and enjoy the food, presentation, the networking and the games

don_hulley · ‎2012-01-25

8.0.1P5 to 8.1 c-mode I just want to understand the process involved.

Thanks

Don

kletzly · ‎2012-01-26

Don

I could not see the images you attached.

I am assuming you are going from 8.0.1P5 7-mode to 8.1RC_# C-mode.

This would involve a data migration, especially since 8.1 is still in release candidate mode.

During the Users Group meeting, we can add this question to the agenda.

Thanks

arxcorpse · ‎2012-02-04

I've got a single node c-mode sim running in my lab and I can't figure out how to create a simple nfs export. I've read the manuals but the cli and the gui don't seem to give you the option of specifying the volume path you want to export. The vServer has both cifs and nfs and I can create cifs shares no problem. Can someone give me an example cli to export /vol/vol1?

Also, it looks like c-mode doesn't support persistent ACLs or local groups. Is that accurate?

Thanks.

kletzly · ‎2012-02-06

For now to get you over the hurdle of nfs exports is for you to use OnCommand Systems Manager 2.0R1, downloadable from http://now.netapp.com.

arxcorpse · ‎2012-02-06

Yep, I'm running that now. It appears that you can only apply export policies/rules to locations that exist in the namespace, not the physical file systems. For example I don't see any way of creating an export policy for a path like this:

/vol/vol1/subdir1/subdir2

Also, is there any way to create cifs local groups?

kletzly · ‎2012-02-07

Hello

Sounds like you are over the NFS hurdle right? Yes, on NFS, everyone is tied to the root of the namespace. And the namespace is not constricted per se to the controller as it is in 7-mode. You can use dirs under "root" to tie in specific FlexVols from various controllers in the overall Cluster.

kletzly · ‎2012-02-06

When you export in NFS, you really do not have an /etc/export file -- etc. So, you need to start by using a policy.

There is documentation at the link/pages where you download the C-Mode Sim.

Here is a cut and paste of introductory statements:

Export policy and rule concepts

Export policies enable you to restrict access to volumes to clients that match specific IP addresses

and specific authentication types. Clients cannot access data on a Vserver until you create an export

policy and export rules.

Each volume is associated with exactly one export policy. Each export policy is identified by a

unique name and a unique numeric ID. A Data ONTAP cluster can contain up to 1,024 export

policies. Each Vserver has at least one export policy called default, which contains no rules. This

export policy cannot be deleted, although it can be renamed or modified. Each volume on a Vserver

by default is associated with the default export policy.

etc...............................................................

DPOLEARY1 · ‎2013-05-22

Hi,

Great information - thanks! One question specific to export policies that I'm having trouble finding an answer to: I'm trying to add 6 clients (by name, not IP) to an export-policy. Can I add these all to a single rule in the "-clientmatch" field, or do I need 6 separate rules, one per client? I've know you can specify whole subnets using IPs such as 10.210.50/24 to a single rule, but I need to specify these 6 hosts by name individually. I tried using the "export-policy rule create" command and specify multiple hosts for -clientmatch (seprated by comma, space, etc) but nothing seems to work.

Thanks

kletzly · ‎2012-02-07

... on your question : "Also, it looks like c-mode doesn't support persistent ACLs or local groups. Is that accurate?"

Local Users and Groups are supported.

See the topics in this guide that start with vserver service unix-group ….

http://now.netapp.com/NOW/knowledge/docs/ontap/rel802c/pdfs/ontap/adminref.pdf

vserver services unix-group create -vserver vs0 -name sales -id 94

I’m not sure what is meant by “persistent ACL”. ACL is usually “Access Control List”, and CIFS clients (or NFS V4 clients) usually are the ones that set these, via the client. Once set, they should always be there until changed, so I guess they are persistent. Is that what is meant, or does persistent ACL mean something else?

arxcorpse · ‎2012-03-08

I should have been clearer in asking about the ACL's. I was referring to CIFS. It looks like persistent ACLs are supported on the file system but local CIFS groups are not. I've heard that local CIFS group support is coming in 8.2. I specifically need the "backup operators" group and it's corresponding rights that allow it to bypass file system ACLs.

Thanks.

kletzly · ‎2012-03-13

Hello arx-corpse

Hmmm. One note is:

Local “Unix” groups are supported. I believe “Unix” in this case is a mis-nomer. They are really just groups, and can be used to enable access for CIFS users who are mapped to Unix users in the Name Mapping process. Assigning permissions to Unix groups which are then properly mapped to CIFS users, should result in CIFS users getting the permissions of those Local “Unix” users.

But, I suspect when you ask about Backup Operators, you are referring to RBACs, right? Not sure.