Hi, 
I tried reproducing the issue like this on fresh installation:

1) modified the epp schema so underscores are accepted by apache

fred@fred-demo:~$ diff /usr/share/fred-mod-eppd/schemas/fredcom-1.2.1.xsd /usr/share/fred-mod-eppd/schemas/fredcom-1.2.1_original.xsd
34c34
<         <pattern value="[a-zA-Z0-9]([-_]?[a-zA-Z0-9])*"/>
---
>         <pattern value="[a-zA-Z0-9](-?[a-zA-Z0-9])*"/>


2) Updated the registry backend handle validation checker to accept the underscored handles


fred=> UPDATE regex_handle_validation_checker SET regex='[a-zA-Z0-9]([-_]?[a-zA-Z0-9])*' WHERE id=1;
UPDATE 1
fred=> SELECT * FROM regex_handle_validation_checker;
 id |             regex              |                        description                         
----+--------------------------------+------------------------------------------------------------
  2 | .{1,30}                        | handle length is at least 1 and no more than 30 characters
  1 | [a-zA-Z0-9]([-_]?[a-zA-Z0-9])* | valid characters according to CZ.NIC
(2 rows)


3) Created the contact and tried informing about it
REG-DEMO@localhost > create-contact --id=TEST_TEST --postal-info.name=test --postal-info.addr.city=test --postal-info.addr.pc=111111 --postal-info.addr.cc=CZ --email=test@test.cz --postal-info.addr.street-1=test
Command completed successfully (1000)

REG-DEMO@localhost > info-contact TEST_TEST
Command completed successfully (1000)

It seems that at least we can confirm that it is not an issue of the standard registrar workflow or the object handles themselves.

Taking a look at the database setup as you mentioned sounds like the way to go. Registrar interface uses these configuration variables for the db access:

fred@fred-demo:~$ /usr/sbin/fred-rifd --help | grep database
  --database.name arg (=fred)      database name
  --database.user arg (=fred)      database user name
  --database.password arg          database password
  --database.host arg (=localhost) database hostname
  --database.port arg              database port number
  --database.timeout arg           database timeout


If your core registry database, logger and fred-bsapp share schema or roles that might imply some namespace conflicts or access issues. On the standard fred-bsapp setup, the roles created for the application should be completely separate from the registry database. The application only stores it's own data in the database and uses the registry domain_blacklist api to update the registry instead.  So no need for the role to have access to the registry in any way.

As far as my limited knowledge about logger goes, I don't think it needs any registry database acess as well as all the logged data are received from application that performs the actual request.

Kind regards,

-- 
Jan Dobiášovský
SQA

CZ.NIC, z.s.p.o.
Milešovská 1136/5
130 00 Praha 3
_______________________________________________
fred-users mailing list -- fred-users@lists.nic.cz
To unsubscribe send an email to fred-users-leave@lists.nic.cz


 

On 9/4/26 17:21, Mario Guerra wrote:
And here is the Apache log (error.log):



root@fred26-epp:/var/log/apache2# tail -f  error.log
[Fri Sep 04 08:18:04.796401 2026] [core:notice] [pid 519:tid 519] AH00052: child pid 806 exit signal Segmentation fault (11)
[Fri Sep 04 08:18:04.797144 2026] [core:notice] [pid 519:tid 519] AH00052: child pid 807 exit signal Segmentation fault (11)
[Fri Sep 04 08:20:41.967887 2026] [core:notice] [pid 519:tid 519] AH00052: child pid 873 exit signal Segmentation fault (11)
[Fri Sep 04 08:20:41.968713 2026] [core:notice] [pid 519:tid 519] AH00052: child pid 874 exit signal Segmentation fault (11)
[Fri Sep 04 08:37:23.085034 2026] [core:notice] [pid 519:tid 519] AH00052: child pid 932 exit signal Segmentation fault (11)
[Fri Sep 04 08:37:23.085761 2026] [core:notice] [pid 519:tid 519] AH00052: child pid 933 exit signal Segmentation fault (11)
[Fri Sep 04 08:47:41.782942 2026] [core:notice] [pid 519:tid 519] AH00052: child pid 996 exit signal Segmentation fault (11)
[Fri Sep 04 08:47:41.783500 2026] [core:notice] [pid 519:tid 519] AH00052: child pid 997 exit signal Segmentation fault (11)
[Fri Sep 04 08:49:59.935576 2026] [core:notice] [pid 519:tid 519] AH00052: child pid 1056 exit signal Segmentation fault (11)
[Fri Sep 04 08:49:59.936126 2026] [core:notice] [pid 519:tid 519] AH00052: child pid 1057 exit signal Segmentation fault (11)
[Fri Sep 04 08:57:50.461327 2026] [core:notice] [pid 519:tid 519] AH00052: child pid 1115 exit signal Segmentation fault (11)
[Fri Sep 04 08:57:50.462234 2026] [core:notice] [pid 519:tid 519] AH00052: child pid 1116 exit signal Segmentation fault (11)

These happen when I type info_contact or when an equivalent XML is issued .

Just talked with Ricardo (I'm forwarding this to him).

It happens that in our case, the dev DB includes database roles for the BSA module and I tested a copy of that DB modified so the logs tables a re no more the the fred DB but NO additional roles. Well, Ricardo did include those roles and info_contact works well. Because these tests do not include the BSA part (the fred-bsapp module) which is not added to this version, I considered they were not needed

We are following this next Monday or Tuesday.

Thanks in advance for your help.

Mario Guerra - NIC-CR

De: Aleš Jarolímek
Enviado: Viernes, 4 de Septiembre de 2026 02:16
Para: A mailing list for users and developers of FRED registry system; Mario Guerra
CC: Andrey López; Ricardo Samper Jimenez; Michael Venegas
Asunto: Re: FRED 2026.1 report

To get complete overview, please send us logs from fred-rifd.log and
eppd.log. Probably those files are located in /var/log/fred-rifd.log and
/var/log/apache2/eppd.log.

Thanks

Aleš Jarolímek
L2 support

CZ.NIC, z. s. p. o.
Milešovská 1136/5
130 00 Praha 3


Dne 04. 09. 26 v 9:25 dop. Aleš Jarolímek napsal(a):
> Hello Mario.
>
>  Thank you for your summary, we're glad to hear from you.
>
> Regarding the "info-contact" case, could you please send us the
> fred-rifd log for that period? We'll analyse it. Thank you.
>
> Kind regards
>
> Aleš Jarolímek
> L2 support
>
> CZ.NIC, z. s. p. o.
> Milešovská 1136/5
> 130 00 Praha 3
>
>
> Dne 03. 09. 26 v 6:26 odp. Mario Guerra napsal(a):
>> For the most part, our tests with FRED 2026.1 has been successful as
>> follows:
>>
>>  *
>>     We use a copy of a development DB and modify it so the log tables
>>     in the DB FRED are eliminated
>>  *
>>     We configured the fred-app services properly
>>  *
>>     We have successfuly generated DNS tables (and signed them, BTW)
>>  *
>>     FERDA, whois, web whois and RDAP work as expected
>>  *
>>     Messenger, secretary and fileman are properly configured.
>>
>>
>> There is one curious problem with eppic (and it happens with our
>> homemade EPP client, too).
>>
>> eppic works mostly well, but there is one command that does not:
>>
>> root@fred26-app:~# eppic
>> 2026-09-03 09:59:15,956 WARNING  transport:connect:105 Verification
>> of the peer certificate is disabled.
>> NIC-REG1@192.168.81.203 > i*nfo-contact --id=CN_1*
>> 2026-09-03 09:59:18,244 WARNING  epp:send_request:163 Communication
>> with the server failed. Retrying...
>> 2026-09-03 09:59:18,245 WARNING  epp:logout:71 Logout failed.
>> 2026-09-03 09:59:18,287 WARNING  transport:connect:105 Verification
>> of the peer certificate is disabled.
>> 2026-09-03 09:59:18,530 ERROR    epp:send_request:166 Communication
>> with the server failed. Logging out and closing connection...
>> Traceback (most recent call last):
>>   File "/usr/lib/python3/dist-packages/eppic/epp.py", line 159, in
>> send_request
>>     return self._send_request(request, cltrid)
>>            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>>   File "/usr/lib/python3/dist-packages/eppic/epp.py", line 175, in
>> _send_request
>>     response = self._client.send(request, tr_id=cltrid)
>>                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>>   File "/usr/lib/python3/dist-packages/epplib/client.py", line 109,
>> in send
>>     response = cast(ResponseT, self._receive(request.response_class))
>>  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>>   File "/usr/lib/python3/dist-packages/epplib/client.py", line 121,
>> in _receive
>>     response_raw = self.transport.receive()
>>                    ^^^^^^^^^^^^^^^^^^^^^^^^
>>   File "/usr/lib/python3/dist-packages/epplib/transport.py", line
>> 158, in receive
>>     raise TransportError('Empty response recieved.')
>> epplib.exceptions.TransportError: Empty response recieved.
>> 2026-09-03 09:59:18,531 WARNING  epp:logout:71 Logout failed.
>> 2026-09-03 09:59:18,531 ERROR    app:run:184 Communication with the
>> server failed.
>>
>> If only happens with this particular operation. The other info
>> operations and the rest of commands, including creation of a contact,
>> work  perfect.
>>
>> And it seems that it happens on the backend, because the
>> corresponding XML generated from our page, which is a homemade XML,
>> equivalent to info-contact, fails too.
>>
>> As I said, info-nsset, etc. work well, too.
>>
>> What do you think?
>>
>> BTW, we have copied our dev DB and modified it (exporting or
>> eliminating the log tables, now in the fredlog DB). And that's it.
>>
>> Best regards.
>>
>> Mario Guerra - NIC-CR
>>
>>
>> _______________________________________________
>> fred-users mailing list -- fred-users@lists.nic.cz
>> To unsubscribe send an email to fred-users-leave@lists.nic.cz
>
>
> _______________________________________________
> fred-users mailing list -- fred-users@lists.nic.cz
> To unsubscribe send an email to fred-users-leave@lists.nic.cz