Around 48hrs ago I noticed that yum downloaded and installed an absolute slew of updates (around 248 I think). Since then I am unable to download any of my e-mails via outlook externally, although I am positive they are actually being delivered as I can read them with the ‘Read User E-Mail’ function in virtualmin.
I have a default Centos 5 install with virtualmin… nothing else fancy or unusual. Any ideas whats gone wrong?
I have the same problem - connection to server timed out error when connecting with external email client and with Usermin.
Yum update upgraded CentOS to 5.1.
Tonight yum updated wbm-virtual-server - I was hoping the update would contain a fix for the problem but it is still occurring.
Maillog shows error:
Dec 5 09:31:26 server dovecot: imap-login: imap-login: error while loading shared libraries: libsepol.so.1: failed to map segment from shared object: Cannot allocate memory<br><br>Post edited by: i00601, at: 2007/12/05 03:37
I had the same problem yesterday. Nobody could receive mail.
I searched and found a solution. There is a setting in the dovecot config needs to be changed.
Find the line that says login_process_size=32. Change to
login_process_size = 64
In my config, that line was remarked out, so it wasn’t being read. I removed the hash and made the change. Restarted dovecot and all was well.
Hope that helps.
sfischer - your hint worked an absolute charm! I made that small adjustment, reset dovecot, and all is running along fine now. Many thanks!
I’m also on CentOS…
What I found was:
outgoing mailserver no longer accepted connections at smtp.domain.net so I changed the clients outgoing mailserver to mail.domain.net
UIDL format had changed from Courier2 to Not Set
I’m guessing the config file was overwritten during the upgrade to 5.1?
Thankfully I have a limited number of clients on this box, so it was not a huge headache.
I changed the UIDL back to C2, restarted Dovecot and Postfix.
I did not need to make the login_process_size change listed above, which seems unnecessary unless you have massive numbers of of mail users with frequent send/receive sessions… I’m guessing dovecot & postfix just needed a restart in this case?
I actually found the login_process_size=32 line commented out after comparing the Dovecot config file to one on a server that has not been updated. Simply uncommenting this line fixed the problem.