Ex: Re: weird java error - java.lang.reflect.InaccessibleObjectException?

[Paul B. Henson]

> From: Cantor, Scott <cantor.2 at osu.edu>
> Sent: Thursday, August 22, 2024 2:20 PM
>
> I just don't think that's ideal given how it impacts users so I've always avoided it.

Agreed; it's slightly disruptive, and I don't run around restarting it randomly for giggles... But it's really a trade-off between the amount of time I need to spend restarting when I need to and some number of users needing to re-click on something. And with my workload and time constraints the balance usually falls on a quick restart when necessary.

> I patch manually, but I would anyway because I don't think it's safe to have the
> signing key accessible at boot. I don't think unattended restart is acceptable
> given the sensitivity of that key, outside of an HSM.

We used to patch quarterly, and I did pull nodes out of the load balancer one at a time before I updated them. But now we have such a focus on "security" that all packages with issues documented (regardless of whether said issue would actually impact the specific deployment) are supposed to be updated within two weeks of the release of the issue, and now we automatically update half of the nodes each week. I'm not going to spend every Thursday night (our maintenance window) doing that by hand.

I think I've mentioned before my managements glee at their plan to deploy the idp in AWS behind an auto scaler, so it's not like somebody's going to be sitting around manually priming those.

It's ironic how all of the security requirements nowadays actually make our systems less secure; I used to have outbound firewalls strictly locked down, now everything has complete access to port 443 anywhere on the Internet so they can phone home to cloud scanners and random third parties like lovely crowdstrike have the ability to run arbitrary commands as root on our systems whenever they feel like…