Implement the Multiple Identity Setup

Now that the test of Multiple Identities was successful, it's time to implement. (This is going to be a documentation post - no interesting plot twists. Non-tech heads may wish to skip this post.)

Saturday

7:13 pm

Modify the /etc/fetchmailrc file to point the messages from all of my accounts to one user.

7:16 pm

Restart fetchmail so that it reads the new configuration.

fetchmail -q
fetchmail -d900

7:19 pm

Send some mail to each account, wait 15 minutes and see where the messages are routed.

7:50

Not working yet. The message to the main account was received fine. Two others are going to their original account, and two more haven't appeared anywhere.

I'm wondering if the service isn't really stopping. After issuing the -q command, the process is shown as dead with the subsys locked. I'm going to reboot the system (I'm such a Windows guy!)

Obviously, I need to learn a little more about how to restart or kill a daemon process.

8:05

Dog walk and web surf break

9:01 pm

It seems to be working now - except for one. The messages for that account are not being picked up off the POP server. There's nothing different in /etc/fetchmailrc. In fact, two of the other accounts are on the same server, and they are working fine.

9:24 pm

I stopped and restarted fetchmail again - this time using the --syslog option so that I could see potential errors in the log. As soon as I restarted, the abandoned messages appeared - still in the wrong folder. In the error log, all of the other accounts at that server had "socket errors". Maybe it was the POP server that was having trouble - not me. Now, if I only new why they were still going to that old user.

9:29 pm

I've just sent out another round of test messages. Let's see what happens...

9:58 pm

Fetchmail ran at 9:50, and one of the five messages in the latest test has arrived. Once again, syslog showed socket errors for the other addresses at one of the servers. I've just got to be patient

Sunday

3:24 pm

I don't know what's wrong with fetchmail. Every time it runs, I see the "socket error" message in syslog, and no mail is downloaded. It only happens on the hotpop.com accounts - others have no error. However, I'm not getting any mail from any account, even though my test messages are still waiting. I tried reordering the poll commands in /etc/fetchmailrc so that the hotpop.com accounts were handled last, but the errors continue. Next, I'll remove the accounts on that server from the file completely to see if the other accounts will still work

3:30 pm

OK. So I'm really getting frustrated by my inability to cause this application to stop. All I want it to do is reread the configuration file. I'm going to reboot the server again, and I'm not even embarrassed about it!

9:30 pm

After removing the hotpop.com accounts from /etc/fetchmailrc, and rebooting the server, the mail from the other accounts came in fine.

I think that it may be time to rethink my plan. With the new disk drives coming today, it doesn't make a lot of sense to setup the current desktop and server, since they will both be replaced. Instead, I'll add the hotpop.com accounts back to the config file at the end, reboot the server yet again, and come back to this later.

Stay tuned.