Dear Micha,
Am 12.01.26 um 22:18 schrieb Micha Lenk:
[…]
looking closer into the details you provided and after some internet
research on other occurrences of the gnutls_handshake error
(GNUTLS_E_PULL_ERROR = -54), I came to the conclusion that the 10s delay
between the following two log messages:
3:2025-12-30 23:15:47:aqhbci jobvpp.c: 112: Clearing responses
3:2025-12-30 23:15:57:gwenhywfar syncio_tls.c: 1353: gnutls_handshake: -54
(Fehler in der Pull-Funktion.) [fatal]
... most likely is explained best by a communication timeout with the
server. There could be many reasons for such a timeout, typically
outside of the Gwenhywfar or GnuTLS code. Fixing them means fixing some
network issues, like a resolving firewall issues or similar. But without
any further information from you, I can only guess. So, I'll stop here.
As you didn't provide more information within the last week, I'll close
this bug now. We can easily re-open the bug if you provide the requested
information before the bug is archived in 28 days.
Thank you very much for your analysis. Indeed, that seems to be it. I
was abroad, and my guess is, that the Berliner Volksbank server
https://fints2.atruvia.de:443/cgi-bin/hbciservlet
has some firewall rules to prevent access from there. Or the hotel Wi-Fi
had an issue, though I did not experience this with other sites.
Kind regards,
Paul
