|Previous Top Next|
In case the server is not constantly on-line, it can periodically connect to Internet via dial-up for receiving the incoming mail and sending the outgoing post.
First of all it is necessary to configure the type of the Internet connection in "Options > Networking" section:
In our case you have to define that the connection to Internet will be via dial-up ("Use Dial-up Networking Connection") and select from the list one of the existing connections. If at the moment the server requires to go on-line the connection to Internet is already established, then there are several options:
|·||Don't use existing connections|
|·||Use only selected connection|
|·||Use any existing connection|
By default the settings entered during the connection configuration will be used, however you can set a specific user name, password and domain.
While establishing the dial-up connection different errors may occur. That is why the server is using the following strategy: it makes a series of dials and in case the connection could not be established, the server makes a break and produces another series of dials. Therefore there are several additional settings: pause between dials within a series, maximum number of retries of a series and pause between series.
As soon as the connection to Internet has been established the server sends the outgoing mail. In order to prevent a mail jam the server can go on-line depending on the amount of messages being queued or on the pending time limit.
Since connecting to Internet does cost a certain amount of money it becomes necessary to decrease that time period. Especially for that there are the following settings available: the automatic disconnect on the preset idle period of time or the forced disconnect upon the timeout.
As the server is not constantly on-line it cannot be responsible for receiving messages to all its domains. In order to keep the mail subsystem working properly it is necessary that someone else takes over the mail retrieval to our domains during that time. In most cases it's the provider's mail server. It receives messages, but does not process them. Instead it keeps them in a separate depository that may be presented as special e-mail account or server queue (spool).
If all messages are collected in a separate e-mail account then it is necessary to use the Remote POP (RPOP) technology in order to retrieve these messages. That means that our server has to connect to the provider's server via the POP3 protocol and receive all messages for that account. At the same time in order to identify the sender and the recipient the message headers are analyzed.
In case the server temporarily keeps the messages in its queue then the ETRN/ATRN technology has to be used. It means that our server connects to the provider's one via the SMTP protocol and sends a special command telling the server it is intending to retrieve mail from the given domain(s). After that, depending on which command was used, the provider's server either establishes a new SMTP connection with our server or the current connection out of the outgoing into an incoming one is transformed. Thus the mail is delivered as though our server had been constantly on-line.
Anyway you have to contact your provider and find out which method they can offer. It is recommended to use the second method, because the message header analysis does not always provide trustworthy information on the sender and on the recipient.