I am Scott Peiffer, Sr. Director from the ONTAP System Manager Product Management & Technical Marketing team. Thank you to our many passionate customers and partners for sharing your feedback on System Manager 9.8. This post is to provide you with background on the changes being made to ONTAP, information on how to learn more about these changes, and guidance on how best to provide feedback to NetApp.
ONTAP System Manager simplification is to directly address your changing needs as you work to support larger data sets, with new applications while expanding into hybrid multi-cloud environments. The simplification effort has been data driven. Significant time was spent on determining and setting smart defaults based on NetApp best practices, automating, and simplifying many of the everyday tasks, and then prioritized workflows that were most frequently used. Since ONTAP provides AIQ with metrics on feature usage, we were able to understand what features our customer use most and what is infrequently used.
The feedback that has been received is being considered and prioritized. To be transparent, there are other items which will not be addressed as they are considered complex, infrequent, or unneeded based on our desire to provide our customers with self-managing storage.
We also hear analysts, for example Gartner stating by 2023, 60% of organizations will use infrastructure automation tools as part of their DevOps toolchains, and improve application deployment efficiency by 25%. (source: Market Guide for Infrastructure Automation Tools, May 2020). API driven integrations will be essential to our customers success. Anecdotally, we have heard in our virtual EBCs that many of our globals and large enterprise customers are managing their infrastructure with tools such as Service Now and Remedy, and even homegrown tools that consolidate service management across various vendors in their data center, all based on API services. There are a number of such tools that operate in infrastructure and application context that we are looking to support...
Please continue to share your feedback though your NetApp account teams! This is how we can best aggregate feedback, address your needs and respond. Help us along this journey with your improvement suggestions.
Sincerely, Ravi Chhabria (VP Engineering) Scott Peiffer (Sr. Director ONTAP System Manager Product Management & Technical Marketing)
Unfortunately I'm going to have to disagree with you on this post.
I have liked using NetApp over the years but the new UI is a real issue. This interface was not around for very long before it was forced on users.
When the interface was first released, we had two meetings about features missing or usability issues and nothing we brought up has been addressed.
Were any customers consulted before these changes were designed? If you look at any online forum discussing the interface its full of unhappy users that are now being forced to use a broken, incomplete interface. I understand the need to change/update the interface but it shouldn't be forced on users until all the issues are worked out.
Scott, you mention NetApp’s desire to provide self-managing storage. While this sounds like a good idea in concept, there are many use cases where providing such an abstracted view of a complex situation is a large detriment. This seems more like a desire from NetApp, after looking at their competitors (like Pure), and deciding they can do it better. Not taking into consideration why NetApp is actually better than and leading most of the market. The devops/API driven automation doesn’t actually touch the GUI, so why you conflate these customers with an interface change feels like a diversionary tactic.
I have lost a lot of respect for NetApp right now.
With NetApp, I just had complains with the hybrid flash array performances, but we overcame that by adding more flash-cache. Since then, we've added aff-200, aff-300, aff-400 almost 500TB of flash storage. But the reason us sticking to NetApp and migrating all our vnx, ibm, isilon to NetApp was the ease of using GUI on the NetApp ONTAP older version. We trained our team members who can easily present the storage, perform other day to day tasks. We recently upgraded to a new version which brings the new GUI which is difficult to use, annoying and confusing. I hope they fix this horror and bring in more automation tools which could simply storage administration.
As someone who stood up a replacement cluster tonight I would like to pass along a bit of wisdom.
GO ---- YOURSELF. The 9.8 GUI is garbage. I don't care how much you you think you have tested this or how much feedback you think you have. I have been using Netapp for the better part of 15 year now and spend the bulk of my time at the CLI. I thought I would try our the GUI tonight to create a peer relationship and it was incapable. The new GUI is as much garbage as is your response.
So. I rebuilt my lab system (again - nothing like digging into wafl until you break it - which is sadly way to easy) today. FAS-8040 cluster. I thought I would "attempt" to build the cluster entirely within the 9.8P1 GUI.
The GUI is now so "simple-(minded)", It is incapable of performing all but one of (creating the aggregate) the simplest of tasks.
>Creating the Cluster
One of the 4 cluster ports was disconnected. The GUI reported the network was in error and would not proceed. even after resolving the issue, reseating the twinax cable and rebooting all nodes. So step #1 had to be completed at the CLI. Strike 1.
>Creating the aggregates
Tiering - clinking on the tiering button configured my aggregates for me. It didn't ask if I liked it. It didn't even ask if I wanted it to do it. It just created them. OK,I won't argue the config it created was less than ideal but it removed all control and may not have been representative of what I wanted. +1 for working, -2 for not giving me feedback that we were proceeding without acknowledgement.
>Configuring the NAS (SMB/NFS) SVM
It could not configure the SVM as all of the ports were in their own separate broadcast domain. (This is real insanity!)
Go to the CLI, remove the half dozen broadcast domains that had been arbitrarily created and manually create my ifgrp, vlan.
Goddamn -> it (ONTAP! Don't bring this craziness to my CLI too!) put those in their own separate broadcast domains too! OK, remove those from the arbitrarily created broadcast domains and create a rational mgmt and data. I manually add the ports to the broadcast domain and the GUI still can't create a NAS (SMB/NFS) SVM. -1 for the GUI failing, -1 for the CLI including stupidity too.
> Configuring the FC SVM
It is capable of creating a FC domain without issue. YEAH! +1
So all of this work and the new hotness failed miserably and without mercy at almost every single step.
I want to add to this, I recently sat through a number of NetApp courses - NFS, SMB, Performance & Security compliance.
Do you know WTF I saw? (besides ancient information going back to 9.1I) I saw ancient classes that had not been updated in YEARS, but more so I saw a progression of GUI interfaces from NetApp that ALL worked differently. The poor instructors that had to deal with the BS were apologetic. "This may not represent what the GUI on your system looks like...". I saw insanity. pure chaos. A company that can't figure out wtf it is trying to do with its GUI. Why does every single release from 9.1 on have a new interface with things in different places? It smacks of a complete lack of imagination, preparation, planning & managerial direction as well as a company flinging every piece of (GUI) poo against the wall as if that would make things somehow better.
NetApp needs to get their game together and that does not start with telling customers "It's ok, we've run this past a few mediocre people who can't tell a disk from a ram stick and they liked it.!". You guys need to get a group together and really work through this. You need to design a single interface that really works and provides a "single-pane-of-glass" interface that is relatively consistent from one release to the next. It would be great to have a consistent interface along with a "What's new!" page that covers WTF is new, but it otherwise operationally consistently and works well across releases. You want to keep the interface fresh or some other ridiculous BS - OK, theme it. I want a dark theme, a light theme and a "it actually works theme." I can tell you that right now - that "dog don't hunt".
This thread has really made me think twice about deploying a new system in a few weeks with OnTAP 9.8P1. I'm thinking we will be better off on 9.7 for now until (hopefully) NetApp listens to their customers and cleans up the new interface. I've been complaining about several aspects of the 9.3+ interface for years now - in particular the fact that creating SnapMirror jobs forces you to add an extension to the destination volume even if you don't want one. Sounds like 9.8 is a whole lot worse. Thank you to all for exposing the issue for those of us not on 9.8 yet.
I spent ages looking for LUN management to put a couple of LUNs back online. Eventually I gave up and had to use the CLI.
The greatest simplification would be during installation if it asked if you wanted to use all storage in one area(aggr). Instead of dividing it between the two controllers automatically. For a small installation this default is a headache for a novice installer as the process to recombine the storage into one area is complicated and time consuming.
Thanks for the feedback. You should be able to bring offline LUNs back online again in the GUI. Please see a screenshot below. The menu can be accessed by clicking on the 3 dot button next to the LUN name.
The LUN which I could not bring online was a boot LUN which had developed a "space error" status after a controller restart. There was no "online" option when I clicked on the 3 dots, but after bringing it online in the CLI it was ok.
Thanks for your feedback. I can see an online option for another LUN which is simply offline.
However there seems to be no offline option for LUNs which are online.
This is the exact procedure which I regard as being complicated and time consuming and further complicated by the number of data partitions depending upon RDD or RD. It would be much nicer if it could be replaced with one question during setup.
Shocking downgrade to the UI! Most of the routine, daily management tasks now have to be relearned as only seem to be available via CLI.
Can't even add a spare disk to an aggregate as far as I can see. What was a 2 click UI process is now CLI. Resizing a volume - why remove the right click and resize volume wizard? It gave a simple process, the current sizes for data and SSR, allowed the user to model the changes before applying.
I find it hard to believe any user feedback was received on this before deploying. Have advised other engineers to stick with earlier Ontap versions.