HX 3.0 -> 3.5 upgrade fail " Checking ESX Host Version on Cluster"

[HyunJun Lee]

When upgrading from HX 3.0 to 3.5, the board firmware on the node was not upgraded in the middle, so the board firmware was manually upgraded. Next comes the phrase below. I need your help.

Checking ESX Host Version on Cluster Nodes with NVMe Disks

Internal error processing getPhysicalDisks: 'org.xml.sax.SAXParseException; lineNumber: 1; columnNumber: 1; Premature end of file.'

[KyoCode]

Hey, 

Just to confirm, are you upgrading to 3.5(2h) based on next doc?

Cisco HyperFlex Systems Upgrade Guide for Unsupported Cisco HX Releases

Did you run Hypercheck before performing upgrade? maybe there's a known issue impacting cluster 

https://bst.cloudapps.cisco.com/bugsearch/bug/CSCvj71618

hope it helps,

AIV

[Steven Tardy · Accepted solution]

See a few older TAC cases and internal CDETs with similar errors but no concrete "to get around this do X".

I would verify no disks are full (which HyperCheck will validate).
Check file /var/log/springpath/lshw-full.xml to see if the file is populated or zero bytes.
Also restarting services (or rebooting SCVMs one-by-one, waiting for storage cluster to heal) may also resolve the error.

[HyunJun Lee]

We urgently purchased the service and opened the Cisco TAC and solved it through the above measures.