One of the biggest difference between “7-mode vs cluster-mode” approach I noticed at the beginning was the term namespace. In 7-mode all volumes were automatically mounted during volume creation to /vol/<vol_name> path. In didn’t matter if the volume was added to vfiler, all volumes on single Data ONTAP 7-mode instance have unique path /vol/<vol_name>. With clustered Data ONTAP the situation is different. Flexible volumes that contain NAS data (basically data served via CIFS or NFS) are junctioned into the owning SVM in a hierarchy.
When the flexvol is created, administrator specifies the junction path for that flexible volume. If you have an experience with 7-mode, it’s safe to compare that junction path for 7-mode was set to /vol/<vol_name>. The junction path is a directory location under the root of the SVM where the flexible volume can be accessed.
Above you can see an namespace, that have couple of junction paths. / is a root path for SVM (also called SVM root volume). vol1 and vol2 are mounted directly under root path, which means that those can be accessed via SVM1:/vol1 and SVM:/vol2.
vol3 Junction path is /vol1/vol3, which means it can be access via SVM1:/vol1/vol3, also customers who have an access to /vol1, can access vol3 by simply accessing vol3 folder (windows) or directory (unix).
dir1 is a simple directory that doesn’t contain any data, but is used to mount vol4 and vol5 to junction path /dir1/vol4, /dir1/vol5 (if you would like to have same juntion paths as in 7-mode environment you would simply call this directory vol instaed of dir1). Finally there is a qtree created on vol5, since it’s junction path is /dir1/vol5, the path to the qtree is /dir1/vol5/qtree1.
This feature have several advantages. For example NFS clients can access multiple flexible volumes using a single mount point. Same with windows clients, they can access multiple flexvols using a single CIFS share. For example, if your project team need additional capacity for their current action, you can just create a new volume for that, and mount it under the volume that this group have already an access to. In fact – a junction path is independent from a volume name. In other words volume1 can be mounted as /volume1 as well as /current_month.
Example: let’s assume that your customers are storing daily reports to SVM1:/current_month location. At the beginning of march you can create a volume called “march” and junction it to /current_month location. At the end of march you can change this junction to /archive/march, and later on create an “april” volume with junction /current_month.
Such operation doesn’t require any action form your customers and doesn’t involve any data movement or data copy on the storage array. It’s a simple modification within your SVM’s namespace.
A namespace consists a group of volumes that are connected using junction paths. It’s a hierarchy of all flexible volumes with junction paths within the same SVM (vserver).
I will create a separate entry about this term. Now I would like to briefly introduce it, to explain you another usage of junction paths. An export policy is used to control client access to a specified flexvol. Each flexvol has an export policy associated with it. Multiple volumes can have the same export policy or all of them can have their unique ones. Also qtrees can have theirs unique export policies. Example: you can create a volume “finance” with junction path /finance that can be access only by selected hosts/protocols. In future, when the finance department need a new volume, you can create new_volume with junction path /finance/new_volume. This volume can be accessed only for hosts/protocols that are in-align with “finance” export policy at least on read-level (and additional new_volume policy).