Hi Artem,
could you perhaps capture the traffic with tcpdump/Wireshark and
investigate what is going on in the wireformat level?
Libor
On 30. 01. 25 17:01, Артём Палецкий wrote:
Hello!
I have an issue.
Knot is configured as a secondary server, and when receiving a zone, a
"trailing data" error occurs, preventing the zone from being loaded
from the primary server.
```
Jan 30 11:03:40 hostname knotd[5407]: info: [
domain.com
<http://domain.com>.] refresh, remote
50788646-db98-4caa-b26e-95b30a470796, address 1.2.3.4@53, failed
(trailing data)
```
The same warning appears when using the `kdig` utility:
```bash
kdig @1.2.3.4 <http://1.2.3.4>
domain.com <http://domain.com> AXFR >
/tmp/domain.com <http://domain.com>
;; WARNING: malformed reply packet (trailing data)
;; WARNING: malformed reply packet (trailing data)
```
The issue occurs specifically with large zones. If the zone requires 2
messages to be received (e.g., `Received 32720 B (2 messages, 442
records)`), one warning appears. If it requires 3 messages (e.g.,
`Received 49083 B (3 messages, 878 records)`), two warnings appear.
However, if I place this zone (`/tmp/domain.com <http://domain.com>`)
into `/var/lib/knot` and then execute:
```bash
knotc reload
knotc zone-refresh
domain.com <http://domain.com>
```
Knot successfully loads the zone.
Unfortunately, due to confidentiality, I cannot share the contents of
the zone. Additionally, I do not have precise information about the
software installed on the primary server. However, if BIND is used as
the secondary server, there are no issues. A regular `dig` command
also does not return any errors.
Is there any way to make Knot ignore the "trailing data" error and
successfully load the zone?
Thank you for your help!
--