EXIM does not function properly after reboot. Does not send mail. BUG.

It is a good idea. But I only have couple personal licenses, so they wont react...
 
And I don't have any issue with it, so little sense to put in one either. Maybe anybody else having the issue and a normal license?
 
I'll keep an eye out. The only time I've experienced emails sitting in queue for a while for no apparent reason was with blockcracking enabled, seems to be part of how that functions. I decided I didn't like that. You removed it and it still happens, so not related. I am interested to know though, if there are any errors in the logs if you grep the exim ID of an email that has been sitting in queue.
 
Can you please explain how to downgrade back to 4.93 ? I have CentOS 7.x with Exim 4.94 and have no more incoming emails ... Need to get back to 4.93 kind of fast :)
 
Sure.
You can do it via SSH terminal or via CustomBuild (just adjust couple files).

Open CustomBuild. Choose Customize versions.
Find "exim". Write 4.93 instead of 4.94. I also removed MD5. SAVE.

Also you need to change the exim.conf version as well:

Find exim_conf_45. Change version to the one that was reinstalled when you changed and installed exim 4.93 just previously (most likely it will be 4.5.23 - that was for me). Delete MD5. SAVE.
Now you wont see the update notifications for these customised versions.

Restart exim. Thats it I guess.
 
Any news here? Exim is not releasing new version for a long time now. I assume the issue still exists in 4.94. I really would like to be able to use newest version of exim, but if the issue is still there, then I cannot. DA did not react at all....
 
Probably needs some more investigation because we got Centos 7 servers and Centos 8 servers which do not have this issue, all running the latest Exim version with the latest exim.pl and spamblocker exim.conf version.
Is this only happening on personal license or also full licenses?
 
Any news here? Exim is not releasing new version for a long time now. I assume the issue still exists in 4.94. I really would like to be able to use newest version of exim, but if the issue is still there, then I cannot. DA did not react at all....

We’ve deployed it across the whole fleet without experiencing this issue. You may end up having to ticket DA support and giving them a test subject to examine.
 
I have centos 7 and 8. The latest exim and config no issues sending mail. If the Smtp server isnt sending mail on reboot did we check
Code:
systemctl enable exim

Just for simplicity's sake.
 
Hello!
I think it's time to give one more try for Exim 4.94.
Its been a while now, since I downgraded to 4.93, I have decided that tonight I will try to check if this strange issue is still present.
Keep fingers crossed for me... :)
 
NOPE. Same shit.

updated OS
rebooted
removed customised versions of exim and exim.conf
updated exim
updated manually exim.conf via “Update Software Configuration”
checked mail sending - works
rebooted
checked mail sending - just keeps circling "Email Sending..." in RoundCube, nothing happens until it shows SMTP Error and timout for request. Puts email into Email Queue...
Tried exim restart - takes about minute and a half to do that, like it was before. Then it sends emails in the queue and sending works. Until another system reboot...

No idea what this FRIKIN' problem is...
 
Last edited:
Als your versions installs has same problem?
Otherwise if no look at diferences.

Ports, tls versions that are supported om that box firewalls whatever.
hello!
I doubt that ports have something to do with it, since 4.93 version is working perfectly fine. TLS did not change either, nor the firewall.
I am using other system just as a DNS server, no mail there to compare...
 
hello!
I doubt that ports have something to do with it, since 4.93 version is working perfectly fine. TLS did not change either, nor the firewall.
I am using other system just as a DNS server, no mail there to compare...
Not in my head now but something with versions yes or no tls or something els and config / setting default for that are diferent, of not working depending in os and version OS
 
So, I have the exact same issue, but I also know how I can prevent it, but I don't know the consequences of it.

At my end it seems to have to do with the /etc/virtual/domains file. That file is/should be linked to the ppl that choose to mail locally of externally. When Externally, the domain should go out of that file.

Now, for the admin and hostname of DirectAdmin box, it seems different. When setting up DirectAdmin it regs it as domain in that file. When I remove it, and reboot the box it works.
But I don't want to remove that file, since those "sub" domains DA adds, are ok to mail locally.

So I am really curious if you got the same case. Your server name/hostname, I assume you got a FQDN on it like vps.myserver.net.
Is that hostname, also in your /etc/virtual/domains file?

If yes, could you test to remove it, reboot your box, and check results?
 
So, I have the exact same issue, but I also know how I can prevent it, but I don't know the consequences of it.
I wouldn't advise to use your solution. The hostname should be present in the /etc/virtual/domains file.

It's an Exim issue as you could read and the solution is to downgrade to Exim 4.93 for those experiencing the problem.
Removing the hostname from that file might get you into trouble with system mails (and maybe other things).

I have 3 servers running (Centos 7 and 8) with 4.94 and the hostname in that file and no issues at all.
It must be somethng different imho.
 
I wouldn't advise to use your solution. The hostname should be present in the /etc/virtual/domains file.

It's an Exim issue as you could read and the solution is to downgrade to Exim 4.93 for those experiencing the problem.
Removing the hostname from that file might get you into trouble with system mails (and maybe other things).

I have 3 servers running (Centos 7 and 8) with 4.94 and the hostname in that file and no issues at all.
It must be somethng different imho.

I never had an email failing from Exim since I did that, also no system mails.. I am curious tho, if he has the same. Any pointers would be nice, and it does not harm to test it.
 
I wouldn't advise to use your solution. The hostname should be present in the /etc/virtual/domains file.

It's an Exim issue as you could read and the solution is to downgrade to Exim 4.93 for those experiencing the problem.
Removing the hostname from that file might get you into trouble with system mails (and maybe other things).

I have 3 servers running (Centos 7 and 8) with 4.94 and the hostname in that file and no issues at all.
It must be somethng different imho.

So after your advice, and a good night of sleep, I was rethinking what to do to solve this.
I added back my hostname into that file, and now only got the the reseller domain in it (this is a fictive one, not being used) and the hostname only. The main domain is out, since I dont that domain to mail locally, but deliver the mail to my O365.

At reboot it hang again, then I thought about putting a 30 second delay in starting Exim, and somehow that seems to work?!

I edited /etc/systemd/system/exim.service
and put a ExecStartPre=/bin/sleep 30 in there.

I rebooted several times to make sure im not dreaming, but on my end I think this is an Ok solution.

What are your thoughts about it?

---
Regards,
Dennis
 
Back
Top