Rspamd 2.1 Update request

Feature
21

do you plan to solve the problem with the new version 2.2?

1 Like
22

I have no problems with rspam2x

23

now I have updated NethServer from “software center” Updates available 

has installed version 2.2 of rspamd and “QuarantineAccount” it works :wink:

but when rspamd service start I get these errors :frowning:

image
image.png1431×193 8.86 KB

24

Loghelper is obsolete and I deactivated it with the rpm from testing

25

The clamvirusfail symbol should be inspected, it is fun because it should works. Do you have the same errors if you restart the service several time?

if you want to inspect deeply you can use less /var/log/maillog and to use / to sort by a string (N and n to move in the file)

26

Restart service rspam:

image
image1418×186 8.8 KB

image
image1495×452 21.7 KB

image
image1550×53 4.47 KB

1 Like
27

rpm -qa | grep 'nethserver-mail'

rpm -qa rspamd

28

image

29

I downgraded to rspamd-2.1-1, the message is no more present in log, I will report to upstream

1 Like
30

even if there are messages I can still leave 2.2, do you work well?

31

We do not use all versions, some have some eastern eggs, play as your own risk :slight_smile:

32

interesting feature with rspamd2

178466:Nov 30 11:17:36 prometheus rspamd[5999]: <c9d0bf>; proxy; rspamd_stat_check_autolearn: <DUJZO36W-23XY-3FVU-JQQF-WG8KAKDTE0GY@facilement.site>: autolearn spam for classifier 'bayes' as message's action is reject, score: 27.49

autolearn spam and ham

if score < 0 -> ham
if rejected -> spam
33

The module reputation (it replaces ip_score) is a good major improvement, I see it in action in logs, the score is added dynamically following several conditions

surbl has been obsoleted and replaced by rbl

34

Hola todos, hoy tenemos rspamd v2.2 para probar, por favor yo necessito tu ayuda

yum install nethserver-mail-server nethserver-mail-common nethserver-mail-filter --enablerepo=nethserver-testing

la QA se encuentra en github

Rspamd v2.2 is ready for the QA, please we need your tests, you can find the QA at the github page


gracie, gracias, thank you, merci

2 Likes
35

I am now using Libraesva antispam, so I temporarily disabled Rsapmd, :slightly_frowning_face:
can I still try the update even if disabled?

36

In the postfix configuration we set the linux socket to use rspamd, if rspamd is stopped I bet that postfix will do a soft reject (try again later)

37

disabling Rspamd postfix seems to work,
Nethserver receives mail from Librasva and delivers it to mailboxes, can you check if it is correct? or what should I change?
if I need it I open a new post, I wouldn’t want to go out of topic here

38

Yes you should open a new thread to explain what you did to switch, eventually why, it could be nice

39

by choice of the company management they chose Libraesva

40

Did some tests on my home mailserver and it looks good so far. Thanks for the great work! :clap:

  1. antivirus (tested with eicar string)
    • virus rejected
    • clamd down, soft reject works
    • clamd down, sending as authenticated user from webmail works
  2. reputation is there
  3. bayes filter
    • log entries for autolearn and bayes tokens are there
    • HAM/SPAM counts up correctly
      • from moving mail to junk folder
      • by uploading spam/ham in rspamd UI

Not checked: Count up of ham/spam on autolearn as the mails have been learned already:

Dec 11 08:07:50 server2 rspamd[6291]: <e7d1b7>; proxy; rspamd_stat_check_autolearn: <7C7A1121.2FABD67C@cnt-grms.ec>: autolearn spam for classifier 'bayes' as message's action is reject, score: 41.70
Dec 11 08:07:50 server2 rspamd[6291]: <e7d1b7>; proxy; rspamd_task_process: learn error: <7C7A1121.2FABD67C@cnt-grms.ec> has been already learned as spam, ignore it
Dec 11 08:07:50 server2 rspamd[6291]: <e7d1b7>; lua; neural.lua:487: cannot learn ANN tRFANNACDF6FD7A287B20A260: too many spam samples: 2

I’ll report as soon as autolearn counts up.

4 Likes