The only supported way to change a spare into a storage node is to re-kickstart the node. The node add workflow expects nodes to be freshly kickstarted and it will fail if one of the nodes being added is configured as a spare.

Are there plans to change that ? Or is the spare node only supposed to replace a bad node ? Most of the instances I come across when having to promote the spare node are capacity issues, so being able to add the spare with the node add workflow would be very convenient.

It's being considered for a future release of the workflow but no final decision has been made.

considering the fact that the workflow method is no longer used to add nodes to a 6.1 grid, what is the supported procedure to convert a spare node to a data node? I am EMC certified to perform node adds. Thanks.

Unfortunately the only supported procedure for converting a spare node to a data node is to re-kickstart it.

Thanks for the information, that is what we will do, rekick and then configure as a storage node.

Mark Nixon | Arraya Solutions

Certified EMC Expert

Senior Solutions Engineer

523 Plymouth Road, Suite 212, Plymouth Meeting, PA 19462

Cell: 610-368-5167 | Support: 610-684-8645

“Do not let the perfect be the enemy of the good.”

Understood, I worked at EMC for 9 years so I am familiar with the policy, my new company has premium support and we have an approved qualifier to add the spare node, which has actually already been completed. Thanks for all the responses.

Mark Nixon | Arraya Solutions

Certified EMC Expert

Senior Solutions Engineer

523 Plymouth Road, Suite 212, Plymouth Meeting, PA 19462

Cell: 610-368-5167 | Support: 610-684-8645

“Do not let the perfect be the enemy of the good.”

Is the following step, to determine that there is no existing data, not suitable anymore?

[   ] Verify that there is no existing data on the new nodes by typing: #: ls /data0? The following appears in the command shell: /data01: avamar cur home lost+found pool /data02: lost+found /data03: lost+found The example above is from a 3.3 TB node.  The /data01 should have five directories: avamar, cur, home, lost+found and pool. For 1TB and 2 TB nodes, the /data01 will only have 3 directories, avamar, home and lost+found.  The other data partitions should have only the lost+found directory.

Thanks.

The only officially supported method for adding a spare node to a system is to re-kickstart it first. The reason for this policy is that we have had issues in the past where a node has been replaced and reconfigured as a spare, then somebody has tried to add that node into the grid. If there is existing data on a node when performing a node add with it, really bad things can happen.

Does this apply to Gen 3 nodes as well? I turned a spare into storage back in 2012 and the proc had nothing in it about rekicking the node. Apparently that's the case now, what changed?