Email is bouncing

Hey so far 24 hours of no errors and email is under heavy load (by adding that temp_errors = *) to the spam directive.

P.S. i am using a src rpm from da for exim 4.51 it was released a while ago.

see http://www.directadmin.com/forum/showthread.php?s=&threadid=8078

so yes i am using the stock da_exim-4.51 src rpm and by adding that directive so far it's been 24 hours with no errors or bounces due to this issue.
 
I too would like to see an official exim RPM for 4.51.

On the subject of SpamAssassin and spamd, I have to say that for us SpamAssassin lets through much more spam than it catches, and we've added a lot of rulesets.

We're much happier with SpamBlocker.

Jeff
 
well it happended again.

guess temp_errors = * didn't fix it.

is spam blocker the 3rd party app you released?

I may try that.
 
SpamBlocker is built in to the official DA exim.conf release (although we will be issuing a new version with more features and a few discrepancies between the actual workings and the documentation cleaned up, by the end of the weekend).

It should be already installed, unless you've done something to remove it.

To implement it you just copy the names of the domains you want to use it from /etc/virtual/domains to /etc/virtual/use_rbl_domains. You don't even have to restart anything.

Note that when you add a new domain to the server you should add it to the use_rbl_domains file if you want it to use SpamBlocker.

You can automate the function by removing use_rbl_domains and using a symbolic link to link it to /etc/virtual/domains.

That way all domains will automatically use SpamBlocker.

Be sure to read the comments in your exim.conf file first, though; you'll have to create a webpage for anyone to go to who gets accidentally blocklisted, so they can get whitelisted again.

Jeff
 
yeah it appears i am using it. It doesn't block as much spam as I would like :( a whole bunch still get through and spam assassin is still crapping out.

*sigh*
 
The default SpamBlocker install should block over 80% of the spam hitting your server; for us it blocks even more.

They key is to have all the domains in /etc/virtual/use_rbl_domains.

Jeff
 
jlasman said:
SpamBlocker is built in to the official DA exim.conf release (although we will be issuing a new version with more features and a few discrepancies between the actual workings and the documentation cleaned up, by the end of the weekend).

It should be already installed, unless you've done something to remove it.

To implement it you just copy the names of the domains you want to use it from /etc/virtual/domains to /etc/virtual/use_rbl_domains. You don't even have to restart anything.

Note that when you add a new domain to the server you should add it to the use_rbl_domains file if you want it to use SpamBlocker.

You can automate the function by removing use_rbl_domains and using a symbolic link to link it to /etc/virtual/domains.

That way all domains will automatically use SpamBlocker.

Be sure to read the comments in your exim.conf file first, though; you'll have to create a webpage for anyone to go to who gets accidentally blocklisted, so they can get whitelisted again.

Jeff
Click to expand...

What should the accidental blocklisted webpage be called and what should it contain for instructions to the blocklisted person?

I have done the symbolic link from /etc/virtual/use_rbl_domains to the etc/virtual/domains file. Anything else I need to do?
 
I know this is an old thread, but I am having this issue too. Was there any resolution to this problem other than disabling spamassassin?
 
I can't remember if this was the same issue but i think it had something to do with DA backup. During the backup process at some point my firewall would F up as well as getting spamd errors at random times etc.

Ever since i switched over to my own backup methods everything has been just fine.

So not sure if it's relevant or not but it i have yet to have a spamd problem using the latest exim and spam assasin (exim 4.53 and spam assasin 3.1.1).
 
I just upgraded to SA 3.1.1 I hope this solves my problem. I'm using exim 4.60. The error we're having seems to be a "timeout" error, I wonder if this line from the SA 3.1.1 changelog addresses this issue:

- bug 4696: consolidated fixes for timeout bugs
Click to expand...
 
The new version did help. I've just disabled SA for now. Thats too bad there isn't a solution for this, my customers like SA.

Thanks,
Dustin
 
gerrybakker said:
What should the accidental blocklisted webpage be called
Click to expand...
Anything you want :) .
And what should it contain for instructions to the blocklisted person?
Click to expand...
You can either set up a form which will automatically mail to you, or set up a whitelisted email address (search the exim.conf file for [email protected] to see what you have to change to whitelist the email address) where people can write you.

We have people send us the email address they want whitelisted, and the address they want it whitelisted for.

At the moment we can only whitelist senders for all domains on the server, but we don't tell people that :) .

You can see our site here.
You're welcome to copy from it, but you should NOT direct people to our site, and we can't give you permission to copy those two graphics at the top as we don't own the rights.

Jeff
 
For fun,

I tried to run a full system backup again. Yep, that caused my exim to become defunct. So something during the backup process is indeed breaking something for me along the way.

Then the problem seems to just worsen over time and I have you reboot.

Wish i could figure out what it is.
 
Can you analyze what time you run the system backup and exactly what time exim stops running?

Can you tail the backup log to see exactly what's being done, while tailing the /var/log/exim/mainlog in a different window, to watch for the exim failure?

Jeff
 
you can't time what time exim starts throwing up <defunct>'s.

cause it's random, and they go away. The only corelation is it happens once the system backup runs and continues to do so until reboot, randomly and on random accounts and emails (some clients complained of the 421 timeout error, myself i never had that problem on my email accounts).

last night i thought i would give it a try (cause i have been without issue using a custom rysnc script i made) for about 7 months.

and alas it started doing it and customers started complaining.

The root cause i think is it somehow someway breaks iptables, blocking communication from spamassasin to exim and they just start crapping out.

i think i will just try to find another backup solution since DA is using a 3rd party one anyway that is no longer supported by the developer.
 
I've just done some searching trying to find exactly what <defunct> means. My understanding is it means the process is finished running and hasn't been closed down yet by the master daemon.

If so, then I misread your other post.

Sorry.

Are you sure sysbk is no longer being supported?

I've learned a lot about it and can take it over if it's looking for someone.

I'm not sure but I may have been the person who recommended it to DA (may not have been, though; I really don't remember). I do know I used it before it was DA standard.

Do you know where it originally came from?

Jeff
 
I'm not 100% sure of the full meaning of <defunct> either but it's not good :)

Yeah i'm positive it's not: http://www.rfxnetworks.com/proj.php

made by the same guy who made APF firewall.

under "Old Projects (projects no longer maintained/updated; and/or on hold due to time restrictions)" is Sysbk.


Either way i think it's going to take me less time to implement another type of backup script rather than figuring out this anomaly that all these things have in common.

It seems to be redhat ES 3.0 specific at least...
 
vandal said:
I'm not 100% sure of the full meaning of <defunct> either but it's not good :)
Click to expand...
Not necessarily true. I've always been under the impression they fill up tables but otherwise aren't bad.

See a thread that agrees with me here.
Yeah i'm positive it's not: http://www.rfxnetworks.com/proj.php

made by the same guy who made APF firewall.
Click to expand...
Thanks; I didn't realize that. I'll think about to start maintaining it.

Jeff
 
Defunct processes aren't necessarily bad. That just means that a process called exit(), but the parent process is still running, and didn't call wait() or waitpid() to find out the exit status of that child process. They should go away when the parent process exits.
 
i had sysbk <defunct>
so i upgraded to 1.2, was 1.0 on my DA/FC3 new server
i will see if any problem again.

... Later after upgrade, same message with # ps x
 
Last edited:
You must log in or register to reply here.
Back
Top