DirectAdmin 1.690

Is Exim 4.99 now working correctly? I see no exim.conf changes, what was the issue with the previous release? Or did Exim release another build?
 
I see no exim.conf changes, what was the issue with the previous release? Or did Exim release another build?

Exim crashed, so there probably was a bug in the exim itself. You might review the patch for more details:

Bash:
# ls -la  /usr/local/directadmin/custombuild/patches/exim_4.99_hold_segfault.patch
-rw-r--r-- 1 root root 12979 Dec  4 06:40 /usr/local/directadmin/custombuild/patches/exim_4.99_hold_segfault.patch
 
Hello,
I tried it and unfortunately I can't find the “checkbox” to activate it. I set it up according to the documentation, but now, what can I say? “Bad Gateway” and the constant attempts to start.

I manually changed all users back and terminated the service.
For users who are not quite so “knowledgeable,” the documentation has been very helpful until now...
 
Hello,


What are you trying to activate? And what is not working?
"Note: The isolated FPM is activated only for the user accounts that use jailshell (the Jailed check box in the user configuration page)."
 
Hello, i've a custom /usr/local/directadmin/data/templates/custom/cust_openlitespeed.CUSTOM.8 with webmail.domain.xx set as default for Roundcube.

After latest DA/Litespeed update any webmail.domain.xx still work but "Let's" is not yet generated on these subdomains, this is also from
/usr/local/directadmin/conf/directadmin.conf:

letsencrypt=1
letsencrypt_disable_renew_after_renew_failure=1
letsencrypt_list=www:mail:ftp:webmail
letsencrypt_list_selected=www:mail:ftp:webmail
letsencrypt_multidomain_cert=3

But discarded in SSL request.
Thanks
 
Hello, i've a custom /usr/local/directadmin/data/templates/custom/cust_openlitespeed.CUSTOM.8 with webmail.domain.xx set as default for Roundcube.

After latest DA/Litespeed update any webmail.domain.xx still work but "Let's" is not yet generated on these subdomains, this is also from
/usr/local/directadmin/conf/directadmin.conf:

letsencrypt=1
letsencrypt_disable_renew_after_renew_failure=1
letsencrypt_list=www:mail:ftp:webmail
letsencrypt_list_selected=www:mail:ftp:webmail
letsencrypt_multidomain_cert=3

But discarded in SSL request.
Thanks
Please check and try to update this file:

/usr/local/directadmin/data/templates/custom/cust_openlitespeed.CUSTOM.8

as it is described here (STEP 2):

https://docs.directadmin.com/webser...set-up-webmail-example-com-with-openlitespeed

because 68 line was modified there after release of DA version 1.690:
include /etc/openlitespeed/httpd-alias.conf
 
Nginx_Apache mode + PHP-FPM - Debian 12 - everything is working OK

As per 1.690 docs those are executed:

da config-set isolated_fpm 1 --restart
da build php

When switched to Nginx Templates

We get 502 bad gateway - error log says

2025/12/10 10:30:33 [error] 1051138#0: *84093 connect() to unix:/run/php-fpm/php84-1029.sock failed (111: Connection refused) while connecting to upstream, client: IP.IP.IP.IP, server: site.com, request: "GET / HTTP/2.0", upstream: "fastcgi://unix:/run/php-fpm/php84-1029.sock:", host: "site.com"

Tested - without template (enabled only) with WordPress Template and Wordpress FCGI Cache

Everything starts to work ok when Nginx Templates get disabled

P.S. I use Jailed Forced on the server
 
Last edited:
Thanks @DanielP. We were unable to reproduce the problem on our test servers. Could you please open a support ticket, so we could check it out on your server?
 
Did something change in 1.690 with DNS or how multiserver sends dns commands in any way?
Maybe the "Remove duplicate records when deleting a DNS record fix" changed something?

It seems we (see Directslave thread) are experiencing odd behaviour of Directslave. The domain .db file are suddenly not created anymore, domain names are put in the directslave.inc file, but on removal of the domain from the DA server, totally nothing is deleted from the DS server anymore. As if the delete command isn't send over anymore via DNS or something.

Might not be related with DA but it's just odd this we discovered this happening after the 1.690 update or slightly earlier.
 
Did something change in 1.690 with DNS or how multiserver sends dns commands in any way?
Maybe the "Remove duplicate records when deleting a DNS record fix" changed something?

It seems we (see Directslave thread) are experiencing odd behaviour of Directslave. The domain .db file are suddenly not created anymore, domain names are put in the directslave.inc file, but on removal of the domain from the DA server, totally nothing is deleted from the DS server anymore. As if the delete command isn't send over anymore via DNS or something.

Might not be related with DA but it's just odd this we discovered this happening after the 1.690 update or slightly earlier.
Could you confirm that it works fine if you downgrade your DA version to 1.689 (stable release) or even older one, like 1.688 ?
To get 1.688, you can use such command:
Code:
da update 88922683f8e8480d51fb8eebf9f57135cdf23902

Also you can try contacting developers of DirectSlave to check the problem closer on their side:


If they find something related to DA, then our developers will check that part closer.
 
or do it via DA Panel from here:
Oke going back makes some differences.

With 1.689 stil the .db file is not created on my Directslave, but on removal, the domain is removed from the directslave.inc file, which is not done when using 1.690.
Also now there is an error in the DS error log which was not present before:
Cant' delete zone file, /etc/dsnamed/secondary/testdomain.nl.db not found.

Going to test 1.688 now.
Oeps, that does not work, tried your command but get this result:
Code:
da update 88922683f8e8480d51fb8eebf9f57135cdf23902
updating to directadmin - - 88922683f8e8480d51fb8eebf9f57135cdf23902
directadmin - - 88922683f8e8480d51fb8eebf9f57135cdf23902: does not exist

try contacting developers of DirectSlave
Will do that but might be hard to reach as he's from Ukraine which is at war. He might be soldier.
 
@exlhost have dnssec-validation yes; in my Directadmin, but I just checked and looks I do stil have dnssec-enable yes; in named.conf of the Directslave server. That did not give issues before.
I now removed that line from the DS server and restarted named. Now it only has dnssec-validation auto; in the named.conf but same thing happens as with 1.690, not even the line in directslave.inc is removed.
This line gets added to the directslave.inc file:
Code:
zone "eventesten.nl" { type slave; file "/etc/dsnamed/secondary/eventesten.nl.db"; masters { 88.xxx.xxx.xxx.xxx; }; max-retry-time 300; min-retry-time 300; masterfile-format text; };
but that zone file is in fact not created by DS. This line gets deleted on deleting the account/domain from the DA server, but not when using 1.690.
Edit: didn't remove the line now either but with error about .db not found. Very odd.
Edit2: zone file not getting created was my fault.

I contacted Roman, but it's unclear if he's able to answer or help.
Does everything work fine for you in DS?
 
Last edited:
My previous post is due to the editting a bit unclear. The non creation of .db files on the DS slave was my fault and is fixed.
Tested with 1.690, problem occurs, line and .db is added but on removal no erros on the DS servers.
Tested with 1.689 looks like everything is working as designed after removing the dnssec-enable yes; statement.
I will try with the other server and report back by editting this post.

@romans Edit Tested with other server too. 1.690 issue with deleting (does not delete things). With 1.689 everything is working flawlessly. So version change to older 1.689 does fix things.
Edit2: Not 100% I just discovered creating and deleting works flawlessly in 1.689, but changing records does not work at all.
 
Last edited:
Back
Top