Provisioning doesn't work anymore after adding Disks to an aggregate

[brauntvr2swiss]

Hi All

Following Problem:

- Two Aggregates (aggr3/aggr4) have increased by adding Disks

- This 2 aggregates (aggr3/aggr4) are member of a Ressourcepool.

- Today We have put a new Qtree in a Dataset which has attached this Ressourcepool. While

  Provisioning want to create a Mirror Volume following "Nonconformant" Error occurs:

-When I remove the 2 aggregates,which have increased (aggr3/aggr4) from the Resourcepool, then Provisioning works without

Problems:

On this Nearstore System I have to add manually a dfpm reslimit after a Ontap Upgrade (8.1.1):

U:\>dfpm reslimit get 1063

Id 1063

ONTAP Version 8.1.1

Product Model FAS3160

Availability None

Maximum number of FlexVols per storage controller

Maximum CPU utilization threshold of storage controller

Maximum Disk utilization threshold of an aggregate

Maximum Deduplication size of a storage system model and ONTAP version (in GB) 51200

Can someone explain me this behaviour....

Thanks a lot and regards

Thomas

[adaikkap]

Hi Thomas,

     I think I got you problem. If I understand correctly you are using OCUM 5.0/5.0.1/5.0.2. Traditionally when we want to a create a mirror relations, the destination volume is created to the size of the aggr with guarantee set to none.

I suspect that in your case the aggr3/4 is 64 bit aggr ? Is it so ? If that's the case then the reslimit doesn't have the max size because of which the conformance fails.

Pls follow the workaround in the bug 591117 and also add your case to this bug so that its prioritized to a release sooner.

http://support.netapp.com/NOW/cgi-bin/bol?Type=Detail&Display=591117

Also you can avoid this by upgrading to 5.1 where we support Dynamic sizing for Mirror destination volumes as well, where instead of creating aggr sized mirror destination volume we only create it to the size of the source and before every update job check for the source volume size and grow or shrink the destination volume accordingly.

Regards

adai

[brauntvr2swiss]

Hi Adai

We will upgrade soon to 5.1. We have to test the ha-config-check.exe before upgrade, because its not anymore integrated in DFM and ConfigAdvisor is not yet supportet with metrocluster :-(

Yes this aggregates are 64bit.

Following  the workaround:

1)     Identify the resource limit Id for the ONTAP version and Platform.

      #dfpm reslimit get

2)     From the above output identify the reslimit Id and update the resource

limit of the 64bit de-dupes for that particular platform and ONTAP version

      # dfpm reslimit set <reslimit> maxDedupeSizeInGB=<value>

But I have did this already I think:

U:\>dfpm reslimit get 1063

Id 1063

ONTAP Version 8.1.1

Product Model FAS3160

Availability None

Maximum number of FlexVols per storage controller

Maximum CPU utilization threshold of storage controller

Maximum Disk utilization threshold of an aggregate

Maximum Deduplication size of a storage system model and ONTAP version (in GB) 51200

Which correspond with our Nearstoresystem

NetApp Release 8.1.1P1 7-Mode: Tue Aug 21 16:54:32 PDT 2012

Model Name:         FAS3160

And the dedupe Limit from the Documentation is set correct:

FAS3160 ONTAP 8.1.x 50TiB

I dont understand, what is to do know. Have I to increase the dedupelimit from dfpm reslimit to 74.5T (aggr size)?

I have found another hidden option called pmAutomaticSecondaryVolMaxSizeMb. Can I set this temporary to fe 40TB till we upgrade to 5.1?

Greetz and Thanks

Thomas

[adaikkap]

Hi Thomas,

     Can you set the dedupe limit to 80TB in your case and try this ? The other hidden options will set the max size for all future volumes to be provisioned or re sized to to 40TB.

Regards

adai

[brauntvr2swiss]

What do you think is the more prober Workaround:

Dedupe Limit to 80TB

or

pmAutomaticSecondaryVolMaxSizeMb to 40TB

or

it doesn't matter

IMHO: My preferred Workaround is the option pmAutomaticSecondaryVolMaxSizeMb

regards

Thomas

[adaikkap]

Hi Thomas,

     I am still leaning towards Dedupe Limit to 80TB as it only affects the controller of type 3160 and those running 8.1.1 but setting pmAutomaticSecondaryVolMaxSizeMb to 40TB affects all volume resize or creation with protection manger to 40TB.

In fact upgrading to 5.1 will solve all of this.

Regards

adai

[brauntvr2swiss]

Hi Adai

Now I have tested the dfpm reslimit set, but it doesn't work with this setting (Can't conform Dataset):

U:\>dfpm reslimit set 1063 maxDedupeSizeInGB=81920

Modified resource limit (1063).

U:\>dfpm reslimit get 1063

Id 1063

ONTAP Version 8.1.1

Product Model FAS3160

Availability None

Maximum number of FlexVols per storage controller

Maximum CPU utilization threshold of storage controller

Maximum Disk utilization threshold of an aggregate

Maximum Deduplication size of a storage system model and ONTAP version (in GB) 81920

With the hidden Option it works:

U:\>dfm options set pmAutomaticSecondaryVolMaxSizeMb=41943040

Changed auto-provisioned secondary volume max size, in megabytes to 41943040.

I think its not a Problem with the dedupe Limit, because the Nonconformant Event say:

"select a size of at most 44.9TB for the new Volume"

This is the size of the aggr before adding Disk and not the dedupe limit of 50TB (defined with dfpm reslimit set)..

I think I leave this hidden option and the next step is upgrading to 5.1..

Thanks for your help and regards

Thomas

[adaikkap]

Hi Thomas,

     Thanks for your confirmation. So its looks like a different issue. I suggest you to upgrade to 5.1 or when you upgrade revert the hidden or turn it off.Now what makes me think is why is conformance still suggesting the old aggrs capacity and not the new capacity. Also is your source volumes dedupe enabled by any changes ?

Regards

adai

[adaikkap]

BTW before adding disk was this aggr 64bit or 32 bit ?

Regards

adai

[brauntvr2swiss]

Hi Adai

Yes Dedupe is enabled and the aggr was from beginning a 64bit aggr...

Which value I have to set for disabling the hidden Value pmAutomaticSecondaryVolMaxSizeMb (disable/off/0).

I wanna test with a temp Dataset without Dedupe to see what happen in this case....

regards

Thomas