I'm having an odd issue. I have a FAS2040 cluster. Controller1 will not send out autosupport information and returns the following error in the log when a test is ran:
Severity Event Triggerred Message info asup.smtp.retry Wed Mar 2 19:17:20 CST AutoSupport mail (HA Group Notification from netapp1 (USER_TRIGGERED (Test Autosupport)) INFO) was not sent for host (0). The system will retry later to send the message warning asup.smtp.fail Wed Mar 2 19:17:20 CST AutoSupport mail to 192.168.0.13 failed (Failed to transmit smtp asup) for message: HA Group Notification from netapp1 (USER_TRIGGERED (Test Autosupport)) INFO info callhome.invoke Wed Mar 2 19:16:56 CST Call home for USER_TRIGGERED (Test Autosupport)
I can do the same test from Controller2 and it works fine. Both controllers have identical Autosupport settings.
Anyone seen this behavior before? Any help would be greatly appreciated. I've wasted 8 hours today trying to get Netapp Support to help me as I also had this same controller failover for an unexplained reason today. Not sure if it's a hardware problem or the two are related.
I had the same problem. I was using an iis smtp server to relay the mail. My autosupport mails had grown larger they the relay server would allow. You can also turn logging on and see what is being thrown.
I though the same thing. Its worth investigating the smtp logs on the relay server. that is how I found the issue. If you are sure the autosupport is pointed to the correct relay server and it can reach it then the problem is mostly like on the relay server side.
We are facing with the same issues. The strange behavior is faced not only on one controller. Lots of issues are related with DoT 8.x. There is a common issues if you use Filer View to send Autosupport - nothing happens, but if you generate message using CLI it works fine. Has anybody ever seen that?
I finally figured this out late last night. Here's the deal.
The mailhost was configured to do a reverse lookup of the IP address that the SMTP conversation comes from. If it can resolve that name, it allowed the mail to pass. If it could not reverse lookup based on that IP source, it denied it. The Netapp was using vif0 to send the autosupport messages. It just so happened that I had a DNS name setup for one controller's vif0 interface (the one that was working), but not the other. I'm using vlans and I really didn't have the need to setup a DNS name for vif0 on the controller in question, until now!
Anyway, bottom line is that I needed to make sure that the IP address of the interface vif0 resolves back to a name. It's a security setting in the SMTP relay host. I finally figured this out by using the packet capture feature in Ontap and analyzing through Wireshark. Hope this helps somebody else out there.