From Julie’s agent FAQ: “In Notes/Domino 6, all From fields are based on the effective user of the agent. To enact pre-6 behavior, you must sign the agent with the server ID; the mail will then come from the server”.
This applies to Simple and Formula agents as well, beginning with R6. Well I have tried everything to get the mail to appear to come from the agent signer and no matter what I do the e-mail generated from @MailSend always comes from the server. Any ideas? I know I can do what I need to in LS but I shouldn’t have to, according to Julie’s article.
Don’t you need to set the field “Run on Behalf of” in the agent properties for it to send from another user? If the agent is scheduled, even if say, you signed the agent, the mail will still come from the server, because the server is the effective user in that scenario. However, if the server is running the agent on your behalf, you should appear as the sender.
At least that’s how I understand that it works. I could be wrong.
I tried using Run on Behalf Of with no luck. And, yes, the server doc is set to allow this. This shoud work according to Julie but I can’t get it to work. Was wondering is anyone else has. And I will state this again just in case someone else reads only this post: I know I can se LS but I shouldn’t have to.
I’ve never tried it with @MailSend, but I noticed that mail sent by an LS agent would originate from the server despite of Run on behalf of settings, if I use tell amgr run “database.nsf” ‘agentname’.
Also note the limited collection of valid name formats for Run on behalf of. If I remember right off hands, a Notes name must be supplied in fully hierarchical format using canonical notation. An email address requires the domain name to be appended.
All mail sending can be completely faked, it’s works 100% reliable but is “unsupported” by Lotus. Well, even the supported things don’t work 100%
When I want a specific e-mail address or sender to appear in a mail, I create a document in the server’s mail.box, and put in the “From” field the name of the sender, and in the “Recipients” field the name of the recipient, and the field “PostedDate” must be set too, else the mail might appear in “edit mode” on the recipient (depending on his mail database design).
However, I would never sign an agent with the server’s id, since there is no quarantee that the server has the correct readers access to the application which the agent utilizes (like Lotus uses “Lotus Development” signer id for everything). The only “right way” is to have a dedicated signer id, which is used to sign ALL designs, including agents. This signer id, must have also all readers field access to all applications. Also regarding SOX monitoring, it is needed that there is only one “trusted” signer (which is basically the end of the gate/process chain needed for SOX) who has all access to everything. No users or servers should be allowed to pass a SOX signer test, since the only ID which is not reported is the “trusted” process signer ID.
you’re missing the point Mika, I don’t want to “fake” anything. According to the FAQ this should work beginning with R6. Even though the agent is signed by me but run by the server it should show as coming from me since I am the signer. Even with Run on behalf of set it still appears to come from the server.
