Zimbra Forums

Hello Sir,

I did implemented strong password policy. cbpolicyd was configured prior to that security measure and it is continued to avoid intentional bulk mails like wishes, etc.

Hello Sir,

Have you tried for milter and cbpolicy in zimbra along with rspamd. Any positive result.

Regards,

Sent from my Redmi Note 4 using Tapatalk

bunny wrote:Have you tried for milter and cbpolicy in zimbra along with rspamd. Any positive result.

Sorry about the late response, I've been slightly sidetracked. Yes, I did try those and in a short test I didn't have any problems and I don't see why rspamd would have any effect on those features.

phoenix wrote:

bunny wrote:Have you tried for milter and cbpolicy in zimbra along with rspamd. Any positive result.

Sorry about the late response, I've been slightly sidetracked. Yes, I did try those and in a short test I didn't have any problems and I don't see why rspamd would have any effect on those features.

Nice to Know that it worked. Is there any change in configuration of all the three ie., Rspamd, milter and cbpolicy.

Sent from my Redmi Note 4 using Tapatalk

No changes, this was on the Centos& test server that had a standard ZCS 8.8.7 and rspamd as per the instructions from the first post. After that I installed the additional ZCS server and configured them as per the documentation and it worked without any problems that I could see.

NV think I solved it

sangamc wrote:NV think I solved it

Did I miss you posting a question/problem? If it was a problem could you post the solution for others to read?

BTW, it's good to hear your experience with rspamd is positive and I'm glad it's working well for you. Have you also seen there's a new 1.7.0 version been released?

Thanks for this, aside from a couple tweaks that were needed, this was what I needed to get rspamd up and running. Lets talk about spam training though. This script only trains for the system account correct? It doesn't train rspamd per user? That is, each user having their own spam/ham learns that only impact their mailbox. I know DSPAM does per user learning and I"m assuming rspamd does too. I don't see this script telling rspamd what user it si processing the messages for though or am I missing something?

Also, the script parses every single message over and over, all of them, eve messages learned 500 times previously. Perhaps this should be done based on timestamps of when the script was previously run so it doesn't need to do so much work? Does zmspamextract include correct timestamps of when the mail was received when it dumps spam?

Anyway thanks for this howto, it really helped a lot!

This was just a rework of the original. Zimbra training script to mimic those functions of ZCS. There is no individual user account training, when a message is sent to the Junk folder (or marked as Not Spam) by a user a copy of that message is sent to the system spam/ham folders. The zmtrainsa script then only needs to read the system spam/ham accounts.

Hello Sir,

Last week I migrated my new zimbra server 8.8.7 with rspamd, cbpolicyd and milter services. Till today it is working without any problems.

Thanks for all the support.

Thanks and Regards