Active IQ Unified Manager Discussions
Active IQ Unified Manager Discussions
My customer reports, and I can validate, that an error condition occurring on the storage controller (e.g. full volume threshold crossed), will cause the Dataset to show an error status, even when the error is clearly not related to the dataset in question.
Example:
Questions:
Phil
Philip Bachman
Consulting Systems Engineer
N/A East - Northeast
Storage Management & Virtualization
NetApp
908.420.7800 Mobile
Philip.Bachman@netapp.com
http://www.netapp.com
Hi Phil,
The Resource Status is the worst possible status of all objects that are associated with the dataset(both directly and indirectly.)
Say for example if your volume v1 comes from filer1 and the filer1's global status is critical due to one of the volumes on the filer using 100% of the inodes
Event though V1 is not part of the dataset but volume v2 which is part of the dataset and belongs to filer1 affects the resource status of the dataset.
Infact the admin should be more concerned about the protection and conformance status.
Resource status is just to give an heads up that those which are not part of the dataset can cause the filer to go down or fail during CF takeover which will affect the dataset protection and conformance status.
Btw can you attach the resource staus window with the event you are mentioning ?
I guess the volume full is the global status non critical event of the filer.
Regards
adai
Adai,
I'll ask the customer for a screenshot.
The true issue is that the resource error sets the dataset status to an error state as well. This means you no longer have a true 'at a glance' status when you look at the dataset screen. You have to select each dataset to see if it is a resource error.
In a large environment, a single 'volume full' could put all the datasets that use that controller in an error state. Not particularly conducive to simplifying management.
Is there a way to change the behavior so the dataset status does not change to an error unless it is directly impacted?
Philip Bachman
Consulting System Engineer
Storage Management & Virtualization
NA East - Northeast
NetApp
908.420.7800
Please excuse typos or brevity, thumbed on my Blackberry.
Hi Phil,
The true issue is that the resource error sets the dataset status to an error state as well. This means you no longer have a true 'at a glance' status when you look at the dataset screen. You have to select each dataset to see if it is a resource error.
That is because the volume might be part of the Resource Pool.
Resource Status is about the status of the Resource Pool that is attached to Dataset Node.
Not just the members of the dataset that are from the resource pool, but for all, that are part of the resource pool.
More over Overall status is an OR values of Protection/Conformance/Resource Status.
Regards
adai
Philip,
I guess the issue is more about the default columns shown in the Dataset view? By default, it shows the "Overall Status" column, which is red if any of the filers that are used by a Dataset have an issue.Typically this column isn't that interesting for someone looking at the Dataset view as these are the backup/restore guys. They care more about "Protection Status" and "Conformance Status". You can add both these columns to the Dataset view by clicking the little icon next the the right-most columns header.
I would suggest to add the Protection + Conformance Status columns and removing the Overall + Resource Status columns (the Resource columns isn't shown by default anyway). That should give you a view Unfortunately the Management Console doesn't save these settings (yet)...
Hope that helps
..- Hendrik
Hendrik,
Thanks, that's a great answer! It will fix the issue for the customer - or would if we could save the view.
Phil
Hi Phil,
You cannot save the columns or in other words the columns selected or disabled are not persistent across log-in/log-outs in NMC.
Regards
adai
Sad. It should be an option.
Philip Bachman
Consulting System Engineer
Storage Management & Virtualization
NA East - Northeast
NetApp
908.420.7800
Please excuse typos or brevity, thumbed on my Blackberry.