Problems after restarting NXRM via Docker, Database seems to be loading, but many other problems in service


#1

Hello,

I’ve been using the Docker build of NXRM (3.12.0) with a persistent disk store and an S3 blobstore. In the past I’ve been able to shutdown the instance, and everything comes back just fine.

However last restart, strange behavior is being encountered. I see in the logs that various repos/settings I have configured are being read which strongly suggest that the db is being loaded, however when browsing via the UI, I’m only allowed to authenticate, but have no authorization, no admin screens, no repos, just barebones.

I have confirmed that the diskstore still has permissions for UID 200.
I have confirmed that authentication fails for a non-existing user.
I have confirmed that the IAM role assigned to my machine has policy in place for the S3 bucket.
Upon trying to browse directly to a repo, non-existent ones continue to return 404, while existing ones show a dialog to use the HTML index.
I was only able to coax an error into the log by attempting to use the HTML index. Per the stacktrace, there is an NPE thrown on this line (I did confirm the branch against my docker image and the trace itself. Reading this code, it appears that an authentication is returned, but no authorization.

Any pointers or things to look at to move forward to rectify? I do have a snapshot of the volume from after the last shutdown, as well as logs from prior to the shutdown.

Appreciated,

Adam


#2

I’ve occasionally seen issues were my browser gets confused about the session state. I usually blame myself because I’m testing different versions and have multiple instances running at the same time and frequently forget to hit the logout button.

I’d first try logging in with a different browser, then perhaps use your browser’s dev tools to clear the cache and application state.


#3

Won’t lie, I was doubtful on that answer… but things seem to work in an incognito browser. Thanks!