cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
428
Views
5
Helpful
3
Replies

Hyperflex cluster expansion issue.

corey.carroz
Level 1
Level 1

Peoples, 

 

We had a problem with a HX node and I needed to re-install ESXi on and and now I am trying add the node to the HX cluster (two nodes are in place today and am trying to re-add the third). 

I keep running into this error. 

 

Installing Software Packages on Storage Controller VM
failed in Task: '' with Error: 'Failed to Retrieve File Information of storfs-catalog Package Bundle'
 
Google turns up nothing on this error :'(   
If I SSH into the Hyperflex installer, is there a equivalent of a "journalctl -u" command that I can --follow the install on? I would like to see what its doing when it fails.  Also, all of the file system in installer seems to be hidden. If I SSH in and do a ls, I only see the backup folder.  Kind of at a loss as to how to find the log files. 
Thanks!
 
3 Replies 3

Steven Tardy
Cisco Employee
Cisco Employee

Are you using the same HX Installer version as the HX cluster?

Unfortunately the HX Installer logs are scattered in several files.

Nodes which were previously in the cluster can NOT be re-added using the HX expansion.

This is a workflow which Cisco TAC can walk through with you.

Open a TAC case.

Steven, 

 

Thanks for the response. We are using the same installer version as the HX cluster.  This node was a part of the cluster, and I did see the call out of nodes that were part of the cluster can not be re-added. However, I am not sure how HX would know if its a fresh ESXi install with all new IPs everywhere. Personally, I would rather be self sufficient  vs having to go to TAC since this doesn't look like a bug. 

Mind telling me what specifically is going on within HX that is preventing the re-adding of a node?  

 

Hyperflex Installer "Cluster expansion" is for never before seen, net-new, servers.

Since the cluster has "seen" these disks and motherboard before this is NOT a "Cluster expansion", but a "node replacement".

The "node replacement" workflow wasn't something HX Engineering put into the HX Installer so Cisco TAC support is required.

The server motherboard UUID and the disk UUIDs are already in the HX cluster database under-the-hood.

 

I agree that the order-of-events you've done is something which should be doable by the end-customers without Cisco TAC support.

I hear rumors of "node replacement" being added to Intersight so the end customer can self service this very common scenario.

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: