I'm almost out of ideas mate, there might be something customised about the config that it doesn't like, a particular version of MySQL or php config that is causing enough weirdness to break it.
I recently ran upgrades from our long running 1.7 install all the way to 1.10, and it worked, I tested it several times before running it in prod, it took hours to process, many gigabytes of data..
And mine has many plugins and customisations.. so if anyone was going to experience it breaking, it would have been mine. Which is what's confusing about your issue.
Your kernel is pretty old compared to mine, which shouldn't be an issue, and I've not tried an EL server in years.. are you running CentOS? Or RedHat?
Perhaps a script is assuming that a module is installed that isn't, maybe you can try transferring a database backup to an Ubuntu VM and try the upgrade again. A bit harsh..
Have you recently upgraded PHP or MySQL as well?
CentOS, hmm, it might be an selinux thing, or even permissions of Apache itself, the upgrade process creates a few temp files that it might be failing to do if it can't write to the application directory.. that's not great actually. Hopefully it's not that.
I wonder if it's a tmpfs storage issue, how much tmp do you have free?
What else is running on the box? Is it your server or hosted? You're not running out of inodes or something weird? Hmm, there's loads of possibilities, but figuring out exactly what is going wrong is challenging.