Hi Felix, > Do you know for which queues ActiveMQ and for which RabbitMQ is used?
Actually, it is weird to mix the use of both. It should usually be either one of them. Can I know which James app you are using? > Do I have to configure a dedicated RabbitMQ server for persistence? I think you can still have persistence with ActiveMQ. > My deployment currently uses an in-memory RabbitMQ which does not have persistence, I guess. The "in-memory RabbitMQ" term seems weird to me. > Is there documentation anywhere on how to configure the queues for persistence? Just make sure your ActiveMQ/RabbitMQ data is persisted is ok. e.g. for ActiveMQ: `var/store/activemq/` (Relative to James’s working directory) e.g. for RabbitMQ: `/var/lib/rabbitmq/` Quan On Wed, Sep 16, 2026 at 4:46 PM Felix Auringer <[email protected]> wrote: > Hi Quan, > > thanks for your reply! > Good to know that I misunderstood the dequeuing in the code and it > actually works like I would like. > Do you know for which queues ActiveMQ and for which RabbitMQ is used? > Do I have to configure a dedicated RabbitMQ server for persistence? My > deployment currently uses an in-memory RabbitMQ which does not have > persistence, I guess. > Is there documentation anywhere on how to configure the queues for > persistence? > > Best regards, > Felix > > On 16/09/2026 11.30, Quan tran hong wrote: > > Hi Felix, > > > >> If I understand the James code correctly (please correct me if I'm > > wrong), the processors completely take emails out of the queue (meaning > > they only exist in memory?) and reinsert them later if necessary. > > This would mean that a crash / server restart while an email is processed > > would mean that the email is lost with no possibility of informing the > > client. Can I do something about that? > > > > Dequeueing does not permanently remove a mail from the mail queue (yet). > > James keeps it unacknowledged while processing it and > commits/acknowledges > > it; failures cause rollback/nack and redelivery by RabbitMQ. A James > > restart during processing, therefore, would not cause message loss. > > > >> Can I configure persistence for the email queues? I would like to have a > > system which I can restart while emails are queued without them getting > > lost. > > > > I think for RabbitMQ, James set those mail queues to be *durable* > already. > > You just need to make sure your RabbitMQ data path is persisted (in case > > you use Docker). > > > > I hope it helps. > > > > Regards, > > Quan > > > > On Wed, Sep 16, 2026 at 1:54 PM Felix Auringer <[email protected] > > > > wrote: > > > >> Hello together, > >> > >> I have some questions regarding the email queues in James. > >> > >> As far as I understand, there are two processes where an email queue is > >> necessary, receiving email and sending email. > >> In my head, the ideal process for these processes looks like this: > >> > >> - After the client has sent all commands, the email is put in a queue > >> (spool / outgoing). > >> - If and only if the queue operation succeeded, James tells the client > >> that its email was received successfully. > >> - The spooler/deliverer takes an email from the queue and processes it > >> (mailetcontainer / remote delivery). > >> - If and only if the email processing was successful (persisted in > >> mailbox/accepted by remote server), the email is removed from the queue. > >> > >> If I understand the James code correctly (please correct me if I'm > wrong), > >> the processors completely take emails out of the queue (meaning they > only > >> exist in memory?) and reinsert them later if necessary. > >> This would mean that a crash / server restart while an email is > processed > >> would mean that the email is lost with no possibility of informing the > >> client. Can I do something about that? > >> > >> According to the docs (and logs during startup), James uses two > different > >> message queues: > >> > >> - ActiveMQ (https://james.apache.org/server/config-activemq.html): > >> Manages at least the queues 'spool' and 'outgoing' according to the > docs. > >> According to my logs, it uses KahaDB as persistence adapter. > >> - RabbitMQ (https://james.apache.org/server/config-rabbitmq.html): Used > >> for distributed MailQueue and distributed event system according to the > >> docs. It is also running in my single node deployment (as in-memory > version > >> because I did not configure anything). > >> > >> Which of them is used for the processes described above? > >> Can I configure persistence for the email queues? I would like to have a > >> system which I can restart while emails are queued without them getting > >> lost. > >> > >> Best regards, > >> Felix > >> --- > >> Gesellschaft für interkulturelles > >> Zusammenleben gGmbH (GIZ) > >> Felix Auringer > >> IT > >> Reformationsplatz 2 > >> 13597 Berlin > >> > >> Tel: 030/513 0100 00; Fax: 030/513 0100 09 > >> www.giz.berlin; [email protected] > >> > >> Amtsgericht Charlottenburg HRB 200872 B > >> Geschäftsführerin: Dr. Britta Marschke > >> > >> --------------------------------------------------------------------- > >> To unsubscribe, e-mail: [email protected] > >> For additional commands, e-mail: [email protected] > >> > >> > > > > --- > Gesellschaft für interkulturelles > Zusammenleben gGmbH (GIZ) > Felix Auringer > IT > Reformationsplatz 2 > 13597 Berlin > > Tel: 030/513 0100 00; Fax: 030/513 0100 09 > www.giz.berlin; [email protected] > > Amtsgericht Charlottenburg HRB 200872 B > Geschäftsführerin: Dr. Britta Marschke > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > >
