We're seeing this too.  Horrible timing for us.

Yeah... glad we aren’t the only ones.

We will be investigating this.

Thanks Robert. I uploaded 2 dumps to my ftp account as well. Hopefully it’s easy to track down, SM going down every few hours is wonderful. 

@echoDreamz since in this case the server doesn't really "go down" or "crash" - what are you using (if anything) to monitor when the mail service gets hung up and stops responding?

We have prtg that goes ape shit and let’s us know. 

Robert, is it possible to tell SmarterMail to forget about Cyren? Or once it’s activated, that’s it...

Sadly, right now the Cyren subprocess just looks to see if it's licensed. We are adding code now to also check if it's enabled/disabled and kill the subprocess if it's disabled. We'll be making a custom build available so you can disable Cyren while we continue to investigate the stalling behavior.

I've sent out the custom build to Shaun and Chris. All this does is prevent Cyren from booting up unless your actually using it (antispam or antivirus options enabled), before this patch it would run if you were just licensed.

Nothing has jumped out at me on why it's stalling but this build should get you guys stable while I figure out the stalling behavior.

I'm having trouble reproducing this. So I'm gonna throw some questions down that can help me narrow down what's causing issues.
-Did you do specifically antispam, antivirus, or both?
-How immediately does the stalling appear after enabling Cyren? (Chris mentioned 4-5 hours)
-When it does stall how does it stall? Do protocols freeze? Does the spool still flow? Does webmail stop responding?
-Is there anything in the windows event log for MailService_SubProcess.exe?

We simply activated a trial of Cyren's anti-spam and turned it on. A few hours later, SM stalled and hasnt stopped stalling since. It's regularly stalls at around 3-5 hours. As rule I've just been recycling the service every 3 hours so I dont hear PRTG go nuts.

When the process stalls, everything quits. all protocols, webmail, everything, disk activity, CPU usage everything goes dead.

Nothing in the event logs at all, nothing crashes, the process just throws its hands up and quits. It's sudden too, there is no "looks like SM is about to give out", it goes from running to immediately dead.

Scratch that, it looks like webmail completely drops first, resetting IIS does nothing, you just get a "waiting..." from the browser and nothing else happens. Then a few minutes later the entire SM process goes out and SMTP, POP, IMAP, XMPP all die.

I made a note in this thread here (https://portal.smartertools.com/community/a92940/beta-7355-user-reporting-they-cannot-login-cannot-impersonate-either.aspx ) that perhaps these two issues are somehow related.

Just wanted to update this thread. I don't believe this is what is causing the stalling based on dumps by Chris. Shaun, I believe you are correct. See that thread for an update with what I found.

Good afternoon, 

We are also experiencing locking issues since updating to version 7355 last Saturday. Simply at the end of each day the service stops working and we need to restart the cloud. 

It did not happen with previous versions.

It's been over 24hr now since we installed the build provided by Matt. SM has not stalled once and webmail users have reported no slowdowns. It appears the issue may have been resolved.

We installed version 7360 yesterday at 7:00 pm, so far it hasn't crashed.

This issue or atleast the issue I diagnosed based on input from Shaun and Chris related to how we load mailbox data into memory and it was an issue with how we locked code. While it started off as Cyren, based on what I diagnosed Cyren had nothing to do with it. I had marked this thread as not a problem and this was my response.
"Just wanted to update this thread. I don't believe this is what is causing the stalling based on dumps by Chris. Shaun, I believe you are correct. See that thread for an update with what I found."


Maybe you are seeing some unresponsive behavior with Cyren but I don't believe it is the same behavior that prompted the creation of this thread. Could you iterate more on the behavior you are specifically seeing, I can re-open this topic. Chris was seeing web interface lock up and stop working, his protocols and clients stopped connecting, the entire SM stack was seizing up, is this what your seeing?

We are using Cyren as well and since Matt's build (and the current beta release), we've had no stalls, no performance issues whatsoever. Just a few complaints of email formatting issues and of course getting MAPI ironed out and SM is in a good place for us.