Re: Best way to use Spamasassin on distributed mtas

2017-05-15 Thread Axb 

On 05/15/2017 03:17 PM, Julian Kippels wrote:

Hi,

what is the best way to use spamassassin and especially sa-learn on
distributed servers? Can I just set up a nfs-share with the bayes
databases and mount it on all servers? Would it be possible to use
sa-learn on several different servers concurrently with this or would I
end up having race conditions?

Thanks in advance
Julian



For Bayes, to avoid performance issues, file locks, etc, I'd recommend 
using the Redis backend instead of a NFS share.

Best way to use Spamasassin on distributed mtas

2017-05-15 Thread Julian Kippels 
Hi,

what is the best way to use spamassassin and especially sa-learn on
distributed servers? Can I just set up a nfs-share with the bayes
databases and mount it on all servers? Would it be possible to use
sa-learn on several different servers concurrently with this or would I
end up having race conditions?

Thanks in advance
Julian