I have a question regarding Access Zones in OneFS 8.2.2.
The goal is to have a production and test access zone, each with their own AD/domain. Two 10Gb interfaces are connected to the switch using LACP. On Isilon there are 2 groupnets (since DNS servers are different with different search domains) and 2 subnets, each with their own vlan-id.
Two access zones are created (prod and test) with path: /ifs/isilon1/prod /ifs/isilon1/test
I would like to segregate network traffic according to best practices and to move data access from System zone. Therefore, there should be access zone for each protocol. In my case only SMB is used, so I need only one access zone for SMB. The synciq and ndmp are also used, but they don't require the access zone.
Since the best practice is to use unique paths for access zones, what should I use in this case?
For example, should the root directory for SMB zone look like: /ifs/isilon1/smb
or I need to use shared access zones, like: /ifs/isilon1/prod/smb
It seems logical, that all shares should be created under prod/test zones and related path, but this would mean use of shared access zones to separate SMB traffic, if I understood everything correctly.
I think the wording is unclear. I believe what we're trying to say there is "put front-facing client protocol traffic in an access zone other than System", not "put SMB and NFS in different access zones". As you point out, that doesn't make any sense.
Access Zones are a way of separating different groups that (can) use different authentication directory services. If you are implementing multi-protocol, there has to be a common source of identities for that to work and so all protocols you wish to use would be active in the same zone.
This is what I've found in "h16463-isilon-advanced-networking-fundamentals" guide as a best practice for Access zones:
=====
When a PowerScale cluster is first configured, the System Zone is created by default. The System Zone should only be used for management as a best practice. In certain special cases, some protocols require the system zone, but generally speaking, all protocol traffic should be moved to an Access Zone. If nothing else, NFS and SMB should have protocol specific Access Zones. Moving client traffic to Access Zones ensures the System Zone is only used for management and accessed by administrators. Access Zones provide greater security as administration, and file access is limited to a subset of the cluster, rather than the entire cluster.
======
Maybe I misunderstood something, but according to above you should separate SMB and NFS for example, to different access zones. In my case, customer won't use multiprotocol shares, only SMB.
maybe some confusion comes from the observation that the System zone and custom access zones will overlap and therefore look like "shared" zones. But that is normal, the System zone is alway at the top level and does not count as "shared" zone with respect to user file access via SMB, NFS and other protocols.
No, I'm not confused with the system zone. Below part is confusing to me:
===
If nothing else, NFS and SMB should have protocol specific Access Zones. Moving client traffic to Access Zones ensures the System Zone is only used for management and accessed by administrators.
If you are implementing multi-protocol, there has to be a common source of identities for that to work and so all protocols you wish to use would be active in the same zone.
tenortim
36 Posts
3765
0
Posted September 14th, 2020 08:00
I think the wording is unclear. I believe what we're trying to say there is "put front-facing client protocol traffic in an access zone other than System", not "put SMB and NFS in different access zones". As you point out, that doesn't make any sense.
Access Zones are a way of separating different groups that (can) use different authentication directory services. If you are implementing multi-protocol, there has to be a common source of identities for that to work and so all protocols you wish to use would be active in the same zone.