Rspamd not saving updates to rules

[Greg]

Using latest 2.0.20 version, I go to "Symbols" page in and try to update the value for MULTIPLE_UNIQUE_HEADERS from 7 to 3.

After a few seconds after I click "Save" I get a red notification error saying:

local > Save symbols error: timeout

It seems I can't update the rule scores. :-\

[SB]

Could this be timeout in connection to redis? Low resources (memory)? Maybe restart container and let it load into memory, then try again?

[Greg]

Redis appears to be running and the redis stats show up in the admin UI.

I do not think this has anything to do with low resources, as the container has more than enough memory and disk space.

Here is my relevant section from the docker-compose.yml:

  poste:
    build: ./services/poste
    restart: unless-stopped
    hostname: $POSTE_HOSTNAME
    cap_add:
      - NET_BIND_SERVICE
      - CHOWN
      - SETGID
      - SETUID
    ports:
      - "${POSTE_LISTEN_IP}:25:25"
      - "${POSTE_LISTEN_IP}:80:80"
      - "${POSTE_LISTEN_IP}:443:443"
      - "${POSTE_LISTEN_IP}:465:465"
      - "${POSTE_LISTEN_IP}:587:587"
      - "${POSTE_LISTEN_IP}:993:993"
      - "${POSTE_LISTEN_IP}:995:995"
    networks:
      - secondary_network
    logging:
      driver: fluentd
      options:
        tag: "{{.Name}}.log"
    environment:
      - DISABLE_CLAMAV=TRUE
      - POSTE_LISTEN_IP
    volumes:
      - ./cache/localtime:/etc/localtime:ro
      - ./vol/poste:/data:Z

The Dockerfile is nothing special:

FROM analogic/poste.io:2.0.20@sha256:a4b2900c1e03b61bb9f6f4358d0d7f48fd5cdb4a7ffd230819fc3845dd372138

RUN apt-get update && apt-get install -y less && rm /etc/cron.d/freshclam

COPY configs/overrides/ /

The only override is a minor change to file 23-nginx.sh:

#!/usr/bin/with-contenv bash

rm /etc/nginx/sites-enabled/default 2> /dev/null
if [ "$HTTPS" == "OFF" ]; then
    cp /etc/nginx/sites-enabled.templates/no-https /etc/nginx/sites-enabled/administration
    echo -e "\t\t\033[1;33m* setting NGiNX to run without TLS certificate\033[0m"
else
    cp /etc/nginx/sites-enabled.templates/https /etc/nginx/sites-enabled/administration
    echo -e "\t\t\033[1;33m* setting NGiNX to run with TLS certificate\033[0m"
fi

sed -i "s/__HTTP_PORT__/$HTTP_PORT/g" /etc/nginx/sites-enabled/administration
sed -i "s/__HTTPS_PORT__/$HTTPS_PORT/g" /etc/nginx/sites-enabled/administration

if [ -n "$POSTE_LISTEN_IP" ]; then
  sed -i "s/;public_ip=N.N.N.N/public_ip=$POSTE_LISTEN_IP/g" /opt/haraka-submission/config/smtp.ini
fi

if [ "$APP_ENV" == "dev" ]; then
    cp /etc/dev/admin.conf /etc/php/7.2/fpm/pool.d/admin.conf
    cp /etc/dev/administration /etc/nginx/administration
fi

The host machine is running Fedora.

[SB]

Have you made any other changes to RSPAMD, other than MULTIPLE_UNIQUE_HEADERS?

Does this occur if you try to change other options?

Can you save the 'Configuration' page/tab in RSPAMD?

Have you selected the correct server to edit (if more than one option in the dropdown) ?

Are you able to open the configuration file and check it is displaying the line as it should?

[SB]

I have just tried to edit this, and I don't get a time out, but it doesn't seem to save the value.

[Greg]

Have you made any other changes to RSPAMD, other than MULTIPLE_UNIQUE_HEADERS?

No.

Does this occur if you try to change other options? Can you save the 'Configuration' page/tab in RSPAMD?

Nope, that's broken too.

Have you selected the correct server to edit (if more than one option in the dropdown) ?

Yes.

Are you able to open the configuration file and check it is displaying the line as it should?

How?

Also, I don't know if this is normal, but after restarting the container all of RSPAMD's stats are reset to 0. Meaning it says "0 scanned, 0 greylist, ... 0 learned", etc.

I don't know if that's how it's supposed to be or if this is an indication that it's not able to save anything at all, which would be a very serious bug.

[SH]

fix #518 remove nginx auth subrequest body and length

→ <<cset 10ced3cfe988>>