Comment 13 for bug 1810972

We have syslog for each rack controller included in the logs attached in comment #8. There are no signs of disk issues - if there were disk issues I would expect other things to fail and there to be warnings/errors about the filesystem in syslog.

Also, we're only seeing this with initrd.

Where are the nginx logs for the rack controller? It would be great to see size of the response and response code for the initrd requests.