Our team is very interested in the questions posted on System Manager functionality and your experience with System Manager. Thanks for taking the time to share. Please keep the questions and feedback on your experience working with System Manager coming.
I have the following filers: FAS3170 on DOT 8.0.1, FAS3240 on DOT 8.0.1P3, and FAS2020 on DOT 184.108.40.206. On all three, when using SM 2.0R1 to configure snapshots on my CIFS volumes, all of the retention settings are set to 0 days even though they are really set to 52 weekly, 198 daily, and 0 hourly snapshots. If I change them using SM 2.0R1 and go back into Configure Snapshots for the volume again, they are all back to 0. When I go to configure snapshots in SM 1.1R1, the retetion settings are there correctly. This issue was in the beta too.
I'm now using the latest System MAnager 2.0R1 build and I noticed that the NDMP management section has been removed. I know that it was there in the previous beta release, will this functionality make it to the production version any time soon?
I've noticed that when clicking the Storage Efficiency or Snapshot Copies tab for a large volume it takes a while for the information to appear, this also seems to load the cpu on the controller while the this "information load" is in process it rises to 100% from about 35%.
Running snap list "volume" or df -S doesn't result in this extra cpu load and returns the information instantaneously.
When creating a new volume or qtree the default suggestion by System Manager is a volume name with the current time and date in it. I don't see any usecase for a naming like that, so it doesn't help the administrator. I suggest showing an empty field so the administrator doesn't have to delete the field content before typing the real vol name. Alternatively use the same bahavior as FilerView, where naming like vol0, vol1, vol2, etc. is suggested as default. This could be usefull to some at least.
The columns in i.e. the vol view should "make an effort" to make better use of the screen . Let's say the administrator has a 27 inch screen and all columns can be expanded to be able to show all content of the fields... System Manager doesn't automatically do this. Furthermore if the administrator makes the effort to expand all the columns... when he logs in the next day... he has to do it again... so the state of the columns should be saved.
For example in the vol view the bottom window is not resizable all the way (up/down). The Administrator should be able to resize i.e. the list of snapshots all the way up.
If the administrator reloads the page in the browser i.e. by accident he should either see what he had on screen before or at least be warned by a pop-up message that he's about to jump to the beginning. F5 is effectvely the same as restarting the application. Intuitive would be that the current page would be reloaded.
The good: it's much, much faster than System Manager 1.1, so thank you for that. Also, there seems to be more control built in. I look forward to the expansion of System Manager to the point that I no longer have to switch between it, the command line, and FilerView. Also, it is now possible to add multiple initiator groups to a LUN, so, again, thank you.
The bad: I am using V-series filers, and it is now impossible to take ownership of an array LUN from within System Manager. Previously, I would go to an unowned LUN, right-click and select Take Ownership, but that option no longer exists.
So far System Manager 2.0 definitely seems greatly improved over 1.1. I definitely appreciate the improved performance and the lack of MMC crashes. I also like the new separation of CIFS Shares and NFS exports in the GUI, with the preview pane of Share Access Control. That'll save me a lot of time, thank you!
The only problem I'm having right now is managing a SnapMirror from one filer. If I open up my v-Series cluster and navigate into a particular filer, down to SnapMirror, I get the error "<Snap Source Host> is not a managed system. Add it through the main page to make it managed system." I'm not sure what to do with this, since the filer it's complaining about is, in fact, already added to the Home page in System Manager and can be fully connected and managed. The oddest part is that I can view and manage SnapMirrors on the other node of the same v-Series cluster. I can even manage the same SnapMirror from the source filer in System Manager without error. I only get this error trying to view and manage the SnapMirror from the destination filer.
What are your recomemndations for best performance with System Manager 2.0R1? Judging from the public beta discussions, it looks like you recommend Firefox over IE, and if you have to use IE then IE 8 over 7. I think there was also a recommendation to upgrade the JRE.
Is this correct and is there anything else you could recommend? Thanks!
Most user who has FilerView experience and like the Read-Only access feature for selected user account. Most feedback of those user talking of System Manager, they would like to know if System Manager will has this similar feature for the view only user - e.g. operator.
Using NetApp System Manager 2.0R1 with IE8 and Windows 7 x64 edition we seem to have performance issues commiting changes to the NetApp controllers, with some changes taking several minutes with OnTap 8.02. Are there any recommended tuning steps at the client, controller, or network infrastructure that should be considered to improve performance or is this a known issue with this release?
A referesh doesn't seem to occur in many areas when switching between secions in the navigation pane, requiring one to press the "Refresh" button in the right pane. Example: SnapMirror status isn't updated for new entries added. Is this refresh issue going to be addressed in future releases?
Using the same network, we tested other notebooks with Windows XP, Windows 7 and all Windows systems have slow NetApp System Manager 2.0R1 read and write performance with IE 8, FireFox, and Safari. One interesting fact is using a Mac OS 10.6.8 and the OnCommand System Manager 2.0 R1 performance is dramatically improved over any Windows system we've tested on the same network.