Hi Robert,

What exactly do you feel is awkward and obscure about Protection Manager's GUI?

Please elaborate.

Is it the concepts of Data Sets, Policies and Resource Pools or something more fundamental in the user interface design.

As the designer, I'd like to get more consrtuctivre feedback than blanket stements like "this sucks and needs a redesign".

Finally, you a wrong about hte target user for Protection Manager.

It was never designed for a dedicated backup administrator that uses NetBackup. It was designed for a storage administrator using disk to disk backup on Netapp equiment only.

On the last point:

I want to change the severity of some events: the errors and critical's are forwarded to our incident system. The problem is that some of those events are less /not important in our situation. It's no a lot of work to eliminate those alerts. 

You can do this using the "dfm eventtype" command. For example, I have changed the severity of an event from error to normal below:

trinity:/root# dfm eventtype list | grep Error

...

user-files-quota-full                         Error        userquota.files  

volume-full                                   Error        df.kbytes        

volume-quota-overcommitted                    Error        volume.overcommit

volume-space-reserve-depleted                 Error        volume.reserve-depleted

trinity:/root# dfm eventtype modify -v normal volume-space-reserve-depleted

Modified event "volume-space-reserve-depleted".

After the change:

trinity:/root# dfm eventtype list | grep volume-space-reserve-depleted    

volume-space-reserve-depleted                 Normal       volume.reserve-depleted

I was forgotten an other important one:

• support for GX

And this is no problem when this go in several steps. It can't be very hard to implement aggregate and volume monitoring. All the other things can come in a second or a third step.

Hi Bryan,

Last week, we had a user group meeting with OM as one of the topics. This are the some extra suggestions of this group  (the suggestions that  I already had mentioned are not repeated: Re: What features do you need added to Operations Manager?  and  Re: What features do you need added to Operations Manager?)

• It's a little bit confusion: you have two interfaces: the web interface and the full client interface for performance monitor, protection manger and provision manager and the web interface for all the general stuff.

• speed/performance of the tool can always better

• a good online tutorial

• for protection manager: the names he used (e.g. for his volumes) are probably very logical, but a normal person can't understand them. Why not used names that everybody understand, so that you without OM/PM very easy can see what they are. A lot of users create the relationships by hand and then import those relations in PM.

• for protection manager: there are always resources where it's normal that you don't want to protect them. I must be very easy to say that's ok that there is no protection for those resources.

HI Philip,

Your are completly right. We have indeed solved this on that way.

TX

Reinoud

Shelf firmware - can it report FW info? recommend where we should be?

If the reports could be given a average for the month on total ops per sec and latency.

Our monthly reports have total space and latencylun latency.  It would be nice if I could show the trend.

Thanks

OM reports should be able to provide a field for Volume/Qtree security information.

Identifying Items such as Vol SecurityStyle (Unix/NTFS) and NTFS ACL's would be very helpful.

I believe this information can be easily gathered with the fsecurity command but not available through OM.

Regards,

Robinson.

Here are a couple of minor items in Prov Mgr:

1. in the NMC gui, I'd like to have a 'refresh' button.  Sometimes when I delete provisioned storage the gui doesn't refresh to show the storage gone.  Usually adding storage works just fine but it seems like deletes don't.

2. In Provisioning Policies I'd like to see the ability to set a default qtree size.  I can set the quotas but not the default size of the qtree to be created.   We are looking at using this for a helpdesk so they can make user home dirs.  The quota is 100MB but the default size for the qtree keeps coming up as 20MB everytime they provision, so they manually need to change it.  Would like to make that part of the policy template.

3. A post-delete script for provisioning. We can have a post-script for provisioning storage, but when they go to delete storage there's no ability to run a script.

Thanks!

Hi,

I'm currently out of the office in training through Apr 09, I will have limited

access to email and voicemail during this time.

If this is an urgent support issue please call our support line at

1-888-4NETAPP. 

I'll return return any mails when I'm back in the office.

Regards,

-jenni

--

Jennifer Coopersmith

NetCache Sustaining Engineer

NetApp Global Services

NetApp

408.822.6908 Direct

jenni@netapp.com

www.netapp.com

I would like to see;

1. Tighter integration or even a merge of SANScreen with Operations Manager. They both do pretty impressive stuff independently, but together could be alot more effective at providing an end-to-end monitoring and management platform.

2.  Host filesystem utilization monitoring and stronger correlation with all the "storage side" utilisation information we report on. FSRM seems to go some of the way there but what I am looking for is a way to help customers use Thin Provisioning to maximise efficiency. Some are wary of Thin Provisioning because they can't get a good view of how much capacity their applications/files are really using and are afraid they will "get a run on the bank", run out of storage without being prepared and have applications go down.

Thanks,

Aaron

I would like to see the data warehouse functionality that the SANscreen team built adapted to work with Operations Manage and Performance Advisor.

This would give the user the ability to produce compelling reports of their own design using the data the DFM server collects.

Phil

Small correction for the above post. "Data Warehousing"  feature is NOT available in Operations Manager 3.7 but what is available is "Data Export" feature wherein user can export the Operations Manager server and Performance Advisor data. Also user can connect reporting tools like "Crystal reports" to the Operations Manager database using JDBC/ODBC connetions and create custom reports. The details of this feature are available in the TR metioned in the above thread.

Thanks

Ravi

Thanks to Andrew Miller for pointing this thread out to me

Been playing with Operations Manager quite a bit this week and sorting out some problems and reports for a customer. Come across a few limitations and areas for improvements, just a couple of bullet points.

• Button for email this report. I know you can schedule an email, but how about just a one off, "email this" option.
• Copy and Edit this report. Making custom reports are often closely based on another report, and it's not always obvious which counters and details another report is using, so being able to create a custom report based on an existing pre-built report would be very useful
• Make this report the default view. So if I have a favourite report I don't have to keep going into the section and then changing the view to see this. Yes I can setup favourites, but being able to change the default report for a particular view would be good.
• When creating a custom report you can change what the field is displayed as. But going back to a report you have to totally remove the entry to be able to change the name. Either being able to edit the display name, or make the removed field the selected one on the left so it's easy to update and re-add back in.
• Being able to filter the top x results. So if I schedule a report on quota usage, I might only want the top 50 users (rather than my customers 8000 getting emailed to them!).
• Being able to filter based on a threshold, but a custom threshold. So show my all quota's that are over 110% for example or show me all volumes that are over 500g.

The chargeback rate can be specified at a resource group level. So if you put all the volumes on FC disks in one group and all the volumes on SATA disks in another (for example) you should be able to achieve what you need. A CLI-based workflow for what you need to do is presented below:

[root@trinity ~]# dfm group create fcvols
Created 1 group.
[root@trinity ~]# dfm group create satavols
Created 1 group.

[root@trinity ~]# dfm group set fcvols chargebackRate=1
Changed annual charge rate (per GB) for group fcvols (2048) to 1.

[root@trinity ~]# dfm group set satavols chargebackRate=.5
Changed annual charge rate (per GB) for group satavols (2049) to 0.5.

-- Move all filers/aggregates/volumes to the respective group using 'dfm group add <group-name> <object-name>' --

You wish is our command.

Operations Manager 3.8, scheduled for release very soon, is deduplication aware and includes deduplication reporting and monitoring. Protection and Provisioning Managers include provisioning support for dedupe.

Phil

Philip Bachman

Consulting Systems Engineer

NA East - Northeast

philip.bachman@netapp.com

NetApp

908-420-7800

Please excuse typos or brevity,

thumbed on my Blackberry.

+1 for this....would be very helpful to see historical trending on deduplication savings both per volume and aggregate/system.

Hello,

Thanks for the response.  I would like to see the vlaue expressed as a percent so I can monitor the percentage over time.  Is this available today?  We want to see how our useage, new OSes, etc are affecting deduplication numbers. I would also like to be able to see all volumes at once on one graph ( this goes back to another issue with respect to more flexible and customizable built-in reporting.

Just a side note -- dedup is in NetApp Systems Manager (can enable/disable as well as view some basic statistics).

Hi,

Maintenance mode for filers in DFM would be nice!  When I am doing something with a filer that's planned I'd like to be able to tell DFM to not alert while that filer is down or whatever.

A feature that was present way back around v3.4 was that when you selected a filer, it showed several charts on one page that gave a good, quick view of overall system performance at a given point in time, not just a few like the current versions do.  I would like to see that again.

Dear Santa,

For my new Operations Manager I would like to see the following features (listed in no particular order of importance):

1. Configuration Management
   1. Ability to configure and enforce vscan options from Operations Manager (specifically 'timeout', 'abort_timeout' and 'mandatory_scan').  I can configure and manage my enterprise AV infrastrastructure from a single console, why can I not do the same for my enterprise storage?
   2. Ability to provide a reporting mechanism whereby Operations Manager configuration management inheritance and override settings have been implemented at different levels within the Operations Manager resource group tree.
   3. Ability to warn when an administrator is about to delete a configuration policy that is currently applied to groups within the Operations Manager resource group tree.  Currently active configuration management policies can be deleted centrally and Operations Manager will not warn whether they are in use or not.  Operations Manager should warn that the policy is applied and provide resource group applied and allow administrator to cancel action or proceed with delete
2. Performance and Capacity Planning
   1. Ability to report capacity in more than just GB.  Explain aggregate utilization in terms of IOPS but not just with the current IOPS reports available today as none of the current reports provide information on what the IOPS potential is of an aggregate.  Number of disks in aggregate (n) x IOP profile per disk type (p) yields Total IOP potential (t), for example.  Then current IOPS rate could be expressed in terms of IO utilization and allow for thresholds to be created so that additional disks can be added to aggregate to alleviate IO contention.
   2. Ability to alert on aggregate utilization growth rates.  In a thinly provisioned environment the focus on capacity management changes from the volume to the aggregate level. The same kind of reporting and alert tools that exist at the volume level are also required at the aggregate, for example:
      1. Aggregate growth rate abnormal
      2. Aggregate growth rate normal
3. Reporting (general)
   1. Ability to report on LDAP domains to which Filers / vFilers are bound (you added Windows [netbios] and NIS to 3.8 but LDAP is still not there
   2. Ability to report on which virus scanners are registered to each Filer
   3. Ability to report on which third party fpolicy servers are registered to each Filer
   4. Ability to provide more than just space savings reports when reporting on de-duplication
      1. De-duplication jobs that have failed and reason for failure (volume ran out of space)
      2. De-duplication jobs that are taking longer than the historical average de-duplication run time for a volume
   5. Ability to report on the disk shelf type, shelf modules and shelf firmware connected to the Filers monitored via Operations Manager
   6. Ability to capture and report on certain volume attributes that are not available today such as:
      1. sis enabled
      2. no_atime_update enabled
      3. convert_ucode enabled
      4. create_ucode enabled
   7. Quality of service reporting for a defined period (week, month, year):
      1. Downtime (system level, aggregate level, storage level, network level) as hours and percentage of period
      2. Degraded service (RAID rebuild, partial network failure)
      3. Reboots
      4. Crashes / Panics
4. Software Management
   1. Ability to provide a central repository for ONTAP, Shelf firmware, Disk firmware and system firmware within a customer environment.  Repository should provide change management functionality to demonstrate no one has tampered with the O/S or firmware images
   2. Combine software distribution and software management with the current configuration management functionality in Operations Manager.  Core configuration management exists today and Provisioning Manager extends that further but a gap still exists with software distribution
   3. Manage ONTAP licenses and licensed features centrally.
5. Use-ability
   1. Many of the reports contain numbers but there is little explanation of how these numbers are calculated.
      Maybe add add some kind of mouse over pop-up that explains how a number in a report was calculated.
      This becomes important when perform capacity planning tasks in a very complex storage environment were technologies such as the following are used:
      1. volume auto-grow (space management)
      2. thin provisioning
      3. deduplication
6. Security
   1. Implement true Role Based Access Controls (RBAC) for ONTAP within Operations Manager
      For non-system/non-scheduled tasks allow the ID of the Administrator logged into Operations Manager to be the credentials passed to the Filer to manage it.  A example of a system tasks would be a Protection Manager policy that deletes snapshots based on a retention period.
      All storage management tasks that fall outside this automated classification should use the ID of the administrator logged into Operations Manager rather than the hostLogin and hostPassword function.  Operations Manager is the object broker, ONTAP is the RBAC broker.
      ONTAP cannot be the RBAC broker if a static set of credentials is used to handle authentication between the DFM server and the Filer
7. Scaling
   1. Federated Operations Manager as opposed to multiple instances of Operations Manager configured to do certain functions in the enterprise environment

Too much?

Hi Richard,

Configuration Management

1. Ability to configure and enforce vscan options from Operations Manager (specifically 'timeout', 'abort_timeout' and 'mandatory_scan').  I can configure and manage my enterprise AV infrastrastructure from a single console, why can I not do the same for my enterprise storage?

This is available in 3.8 (in bug 332666).

Thanks,

Raj.

Interesting... I have seen other customers who have built scripted log monitors that pull their messages and syslog to a central location (the DFM server in this instance)  and then use parsing to highlight items of interest, your vscan.dropped.connection would be one example of messages they parse for and alert;

I was also asked about this precise functionality not 30 minutes ago...

Cheers
Rich

Upload buttons to send performance (maybe perfstat?) and ASUP data to NetApp so it can be used for Support and account management stuff

I would like the ability to script custom events that self-heal, like the built-in ones do.

For example when an aggregate-full alarm is issued, the aggregate-space-normal event will cause the original to be removed and the status set to green.

Right now I am writing my scripts to re-test all original events and manually delete them if the situation no longer persists, but this is really cumbersome.

Hi,

You'll see the Custom Zoom feature in the next release of PerfAdvisor;

Interesting points on the other requests

Cheers

Rich

Not sure wether this thread is still actual or not.

But as I already stated earlier in some posts, the biggest pain point is still the snapshot naming.

I would like to have:

- the choice between a "generic"-naming (_recent) and "unique" (timestamp)

- snapshot names without any blanks/spaces, cause they are a nightmare for every scripter

- no ":" in the timestamp, because this seems to cause problems for Windows previous versions (https://now.netapp.com/eservice/case/caseView.jsp?caseId=200121238)

Next I am still unhappy with the concept of choosing an "archive management group" and thus be obligated to create jobs for every retention I want.

For the interested the posts:

http://communities.netapp.com/message/12645#12645

https://forums.netapp.com/message/9426#9426

http://communities.netapp.com/message/10774#10774

DFM by design restrict access to the DFMGROUP DB only.

You can view the data exposed by Database Views.