I am scared to upgrade....

I don't want to scare people about it, it's maybe something related to our setup but downgrading fixed it right away though.

No issues here on two servers, but they are fairly low volume.

Two servers here, no problem reported...

Our SMTP and XMPP stopped working. Seems it has something to do with the SSL certificate and not providing encrypted connections. I have a ticket opened with ST. We are doing an emergency roll back.

The issue appeared to be TLS not working for SMTP and XMPP. Rolling back to 9273 fixed it. But we also ran into an issue where our whitelisted IPs had SMTP Auth Bypass turned off (previously enabled). Not sure if it was during the upgrade or roll back.

Chris,

The SMTP Auth Bypass was likely lost during the rollback. While preparing some changes for the coming High Availability release, we realized that SMTP Auth Bypass settings were being stored in two different ways that were not always being properly kept in sync. With this update, the two settings were consolidated. The loss of  SMTP Auth Bypass after rolling back is a result of code looking at the setting that is now deprecated.

And to follow up a bit -- Chris has a ticket open with us and we were able to get a config only copy of one of his gateways. So far, we can't replicate the issue when upgrading to 9280, but we'll keep testing. We also reached out to Sebastien, but have not heard back. 

If anyone else sees this issue, please open a ticket. 

After updating to 9280 services are responding, but on two SmarterMial servers that I have updated so far, SMTP response times have gone from 15 msec to 1,034 msec

After rolling back to 9266, SMTP response times are back to normal.

Hi Bruce, I checked our PRTG monitoring in house for some of the test servers we run and I'm seeing some of the same behaviors here:

The downtime highlighted in red/white there is when I upgraded the machine to the test build I believe, so this would align with what you're seeing as well. As you can see this server usually has a very low response time so that is a pretty significant jump. I see the same jump in response time on another test server around the same time so I'm going to get a task written up to have development investigate this further. 

Thank you, Kyle, for letting me know you were able to replicate the same behaviour so that development can take a look.

We're triaging this now, its related to upcoming PROXY protocol support. This is not necessarily a surprise but how it's showing up on monitoring tools is not something that showed up initially. Latency looks bad but we are getting the connection, buffering it, and handling it otherwise normally. This will affect all the TCP based protocols and shouldn't cause any other issues (minus the delay).

Regardless I'm looking at getting this behavior fixed.

@Bruce the delay you were seeing on SMTP should clear up in the latest public release build :)

Thanks, Kyle, build 9287 does looks to have fixed the issue

Hello, I can confirm that with build 9287, the original issue I posted here is gone. No more timeouts on TCP services after upgrade.

Thanks!