Undocumented network traffic change in Nimble OS 5.2.1.600 to 5.2.1.700

[JulezMBI]

I just wanted to verify this behavior we're seeing on our new array.
We purchased a pair of new HF60s to replace our aging CS500s. I was working through the setup today and toward the end I was going to finish by updating from 5.2.1.600 (what it shipped with) to 5.2.1.700 (what was available for download).

I configured everyting short of the authenticated SMTP address and started to work on that. I had some issues with the account, so I decided to skip it and insead finished the firmward update. While finishing the firmward update I figured out the issue was related to an SMTP block we had in place on our firewall for the management network of the new HF60. I rectified this issue and tried again. I still wasn't getting any email alerts.

Back to the firewall, I notice NTP and SMTP traffic no longer coming from the IP of the new HF60. Instead, after hunting around for awhile, it's now coming from the Diagnostic IPs, instead of the management IPs. We allowed the SMTP traffic through from that address and we successfully recieved email alerts.

I went back to the release notes for 5.2.1.700 and didn't see any indication that this change was being made. Is this a bug? Or is this an undocumented change?

[SprinkleJames · Accepted solution]

Hi JulezMBI,

It has long been the case that you should allow SMTP relay from not only the group management IP address, but also the controller support/diagnostic IP addresses. SMTP traffic can come from any of these IPs, as well as remote support and other types of management traffic. There are several KB articles on InfoSight that mention this about SMTP relay.

I do wish they would make this more clear in the admin guide, with regard to SMTP traffic specifically.

[JulezMBI]

Thanks for the reply!
This is interesting, I've been using Nimble arrays since the CS220s came out, well before RBAC was introduced and I had never seen this mentioned, nor did any of our SE's ever mention this to me.