Emails backed up "unable to start handling..."

gerber1

I am stuck, unable to send emails. This was all working fine in development/test and then in production when initially deployed. Then production developed trouble connecting to the SMTP server due to an apparent port requirement change (25 used to work, but now it must be 3535... It is GoDaddy). The simple setup change from 25 to 3535 worked in dev/test.

When I save the SystemSettings on production after changing the port from 25 to 3535, I see the message:

The Page at xxx.xxx.xxx:8080 says:
"unable to start handling incoming or outgoing e-mails due to the following exception outgoingEmail_L3WBCHEM01 (access denied)"

On the server, the log shows:
2015-04-28 14:10:31,541 ERROR -Unable to start e-mail channel for business space L3WBCHEM01. It is possible that SystemSettings object is not initialised.

This entry is then repeated 3 more times consecutively a bit further down after

I also see in the startup sequence:
log4j:ERROR setFile(null,true) call failed.
java.io.FileNotFoundException: system.log (Access is denied)

1) Am I simply having a privilege issue I need to resolve with what the server can access?
2) I had recently turned logging on and had to enable access to several props files and the like for it to save. Maybe some lockdown procedure was run on my hosted (VPS) server...
3) currently, no logging is enabled, per the checkboxes in the System Settings/Logging tab.

Other thoughts:
Does anyone know how to simply empty the email queue of all pending emails (wherever it lives that it survives a server reboot)? I can't really have all that accumulated junk start going out after it was otherwise resolved...
Does anyone know what "(access denied)" even means in the context trying to save SystemSettings? (It does save the SystemSettings changes, by the way.)

I have seen before what appeared to be backed up emails in the Java email handling, and I think this may be part of what is happening here.

Thanks in advance!

Roy

AwareIM 6.0 (Developer, build 2030)

customaware

Hi Roy,

I am not a server/network guy in any way shape of form so this is just a wild stab in the dark and at the risk of unbridled ridicule....

Have you opened port 3535 on your server?

gerber1

Thanks, Mark.

That did not help, but I did two things, and a release of the backed-up email followed:

1) I copied a BSV from a test app to my server for AUTO_DEPLOYMENT, 2) I changed the port used for email from 25 to 80. I believe the latter was what got it started again..

Now this problem is back. I will try changing the port again. I saw that someone else has had the same behavior and now alternates ports to get them flowing again.

I still wonder where these emails exist when this problem occurs. Is it some resource owned by Java? They are not in any table I have inspected. However, they are persisted somewhere because they survive a reboot of the machine. When the blockage is removed, I have seen emails from more than a week prior start sending. This can be embarrassing if you have already done something outside the application to remedy an issue, and then a week later an email goes out. I would like to be able to see them and maybe cancel some of them...

Support
There are enough issues around this behavior from enough users that some insight into how this works when everything is normal could be helpful in resolving it.

What is the following message actually telling us happened or didn't happen at a more detailed level?
"2015-05-14 09:55:45,563 ERROR -Unable to start e-mail channel for business space L3WBCHEM01. It is possible that SystemSettings object is not initialised."
I know it is not that there is no SystemSettings object, which was created months ago...

Why do I always see these messages in groups of four differing only by the thousandths of a second in the time?

Where and how are emails stuck behind this problem persisted?

Are there ways to recover stranded emails, so they can be addressed in some other way?

Thanks,

Roy