While upgrading. the system will go into maintenance mode.

And the maintenance mode will show "upgrade NMA has errors"

as in the image below.

This symptom is seen if having a Space cluster and there is no sufficient disk storage available in /var/log direcotory in one of the nodes.

PLEASE NOTE: (when there is just one node of Space and /var/log" is at 100% usage WEB GUI couldn't have been accessible while firstly pushing the upgrade)

To fix this the next steps are needed.

1. Login to Junos Space CLI
2. use "df -h" command to check partitions and storage available
3. if "var/log" directory is at 90% or above you will need to delete some old log files to understand the type of log files check https://www.juniper.net/documentation/us/en/software/junos-space22.2/junos-space-monitoring-troubleshooting/topics/concept/log-files-overview.html, contact JTAC support in case of any doubt with the files you want to remove. 
4. Once you clear disk storage in "var/log" in all nodes, another upgrade attempt can be done and NMA libraries error should not been seen again
5. If upgrade is successful after clearing storage you can continue regular operations on WEB GUI
6. (optional) For future preventive or maintenance steps check if you have a purging policy to have the Space server deleting old log files as seen in https://www.juniper.net/documentation/us/en/software/junos-space21.3/junos-space-workspaces/topics/concept/purging-policy-overview.html

To avoid "/var/log" directory being 100% usage

Overview Purging policy https://www.juniper.net/documentation/us/en/software/junos-space21.3/junos-space-workspaces/topics/concept/purging-policy-overview.html