Our POP server "hangs"

[haim [howard] roman]

We're running UW's ipop3d.  This week, it often happens that users 
cannot retrieve mail.  Sometimes rebooting their station (MS Windows) 
helps, sometimes not.  When rebooting doesn't help, we kill all the 
ipop3d processes; that usually helps.

Does anyone have any ideas?  I searched the c-client mailing list 
archives, but I could not find anything.

Here are  more pieces of information:
(1) Our ipop3d version is v2003.83.
(2) The OS is Red Hat Linux release 9 (Shrike), running on Intel.
(3) Sometimes, we might be having too many POP sessions at once.
   (A) We sometimes see the following message:
           xinetd[2586]: FAIL: pop3s service_limit from=212.235.0.88
       That indicates too many concurrent sessions for that service.
   (B) For POP & IMAP, we changed from 60 (xinetd's default) to 200.
   (C) However, we've experienced this problem of users not getting
       mail even when there are far fewer than 60 processes.
(4) Very often, I see many ipop3d processes, *all* of which are owned
   by root. 

   (A) The ipop3d is supposed to change its UID (or EUID)
       to the user who initiated the session.
   (B) This might be due to a lot of failed logins to POP.  However,
       why are the processes staying around?
(5) Most of our users user mail clients on MS Windows, usually
   Outlook or Outlook Express, sometimes Mozilla or Thunderbird.
(6) I'm not aware of any problem with our IMAP server.  However, there
   are far fewer of them, and most of them probably use webmail.
Thanks, 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Haim (Howard) Roman     |  Computer Center
[EMAIL PROTECTED]         |  Jerusalem College of Technology,  ISRAEL

--
------------------------------------------------------------------
For information about this mailing list, and its archives, see: 
http://www.washington.edu/imap/c-client-list.html
------------------------------------------------------------------