VLDB not consistent after a forced MetroCluster switchover…

After a forced MCC switchover and a successful switchback I had the following error on a customer MCC in the event log:
6/21/2017 09:56:49 cluster01-01 ERROR vldb.aggrBladeID.missing: The volume ‘cluster02_svm1_root’ is located on the aggregate with UUID ‘3b61c6fd-b6d7-4d5f-9168-81b12632581c’ whose owning dblade UUID ‘cce2f293-3bd4-11e7-85ef-a1f0d21750d7’ does not exist in the Volume Location Database.

This is an open internal bug from NetApp and can rarely happen if the VLDB has been updated without the knowledge of the other cluster:

It can be solved by executing the following command on the cluster which hosts the affected SVM:
cluster02# metrocluster resync vserver -cluster cluster02 -vserver cluster02_svm1

Check the status with:
cluster02# metrocluster vserver show
cluster01# set diag; debug vreport show; set admin

Does anyone had the same problem recently? Thanks for commenting that.


NetApp HCI is here…

Normally I don’t do postings about Product Releases or New Features of a product but this one I really want to share with you because this is a huge change to the world of HCI (Hyper Converged Infrastructure).

On 5th June NetApp announced the new HCI product which integrates into the NetApp Data Fabric.

The NetApp HCI blocks ar based on VMWare vSphere Compute Nodes and SolidFire All-Flash Storage Nodes. Each block has a total of 4 Nodes.

But what does it make special and what’s the killer feature of this product?

  • You will get all the SolidFire advantages like guaranteed Performance, world-class QOS, vVOL integration, and so on…
  • You will get full Storage power with the All-Flash HCI…
  • Mix and Match Compute and Storage Nodes matching your requirements…

  • Easy management via vCenter Server Web Interface…
  • Easy setup – The appliance is setup in about 30 minutes…
  • IMHO the real killer feature is the integration into the NetApp Data Fabric like SnapMirror and StorageGrid…

As I already mentioned each block has 4 Nodes. You can also unite multiple boxes with the 25GbE/10GbE links and get more Storage/Compute power. The Datastores are attached with iSCSI. If you want to have the familiar OnTap feeling, just deploy an OnTap Select instance and you will get the whole bunch of CIFS, iSCSI and NFS power for your environment.

What can we expect next?

  • We will hopefully get more and more Hypervisors
  • NetApp HCI SnapMirror integration with OnTap 9.3
  • Integration into SnapCenter 4.0

I’m really excited about my first use of this product and I’m #DataDriven…

Changing raid type RAID-DP to RAID4 and vice versa

I just needed to change the aggr raid type of the root Aggrs for an AFF MCC from RAID-DP to RAID4. This is an easy step:

storage aggregate show-status -aggregate AGGR0_NAME
storage aggregate modify –aggregate AGGR0_NAME -raidtype raid4

And verify:

storage aggregate show-status -aggregate AGGR0_NAME

The same can be done the other way:

storage aggregate modify –aggregate AGGR_NAME -raidtype raid_dp

or with RAID-TEC

storage aggregate modify –aggregate AGGR_NAME -raidtype raid_tec