Re: [vchkpw] busy try again later error
Rick Macdougall wrote: > > And you have about 6K users ??? Our server with about 30K users almost > never goes over 40 concurrent connections, and we do have quite a few > users checking every minute. > > Are you sure there isn't some issue with authentication or perhaps I/O > that is slowing down the pop connections ? > > Regards, > > Rick > Yes, the processes are a bit overkill, and we maybe have 100 concurrent connections as many of our customers use e-mail as a file delivery service and so utilize the connection for some time when getting messages. I don't believe I remembered to note that the mailboxes are on an NFS mount. The filesystem on that server is JFS mounted with the noatime option. The connection between the mail and file server is on a private switch with the file server being on a gigabit port, and the mailserver on a 100 mbit port. Perhaps it could be some weird locking issues with the cdb files over NFS, but I would think that would have exhibited itself before now when the office staff all connects to get mail. This is a new problem. It started in September and we haven't added many new accounts since then, maybe 100 domains. Nothing mail-related on the mail server, and nothing on the fileserver has changed since before that time. - Michael
Re: [vchkpw] busy try again later error
Michael Johnson wrote: Rick Macdougall wrote: I've been running into an issue on our mail server recently. As traffic has been increasing on it, I've been getting the response that I listed in the subject more often, and now customers are also seeing it (busy try again later!). Of course, this response causes any mail clients to prompt for password, making it rather annoying. I'm using sslserver from the ucspi-ssl package so I can offer ssl-encrypted connections as well. The non-encrypted process is set to 300 connections, and the ssl process is at 200. And you have about 6K users ??? Our server with about 30K users almost never goes over 40 concurrent connections, and we do have quite a few users checking every minute. Are you sure there isn't some issue with authentication or perhaps I/O that is slowing down the pop connections ? Regards, Rick
Re: [vchkpw] busy try again later error
Rick Macdougall wrote: > Michael Johnson wrote: >> I don't have that file, so I'm using the default (7200 then). >> >> How would this have an effect on POP3 connections? >> ___ >> Michael Johnson 909-740-3156 x3163 >> System Administrator [EMAIL PROTECTED] >> PhD Computing http://www.phdcomputing.net/ >> >> >> Rick Macdougall wrote: >>> Michael Johnson wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 I've been running into an issue on our mail server recently. As traffic has been increasing on it, I've been getting the response that I listed in the subject more often, and now customers are also seeing it (busy try again later!). Of course, this response causes any mail clients to prompt for password, making it rather annoying. >>> Is your timeoutsmtpd set in /var/qmail/control to something lower than >>> 7200 ? (I'm using 180, I know at least one person on this list is using >>> 60). >>> > > Heh, sorry. I just glanced through it and assumed it was an smtp error. > I didn't see any mention of pop3 in there. > > You could increase the number of allowed pop3 connections though. The > default is 40. You can increase it by using the -c xx flag with > tcpserver, where xx is your required number of connections. > > Regards, > > Rick > > > Sorry if I wasn't clear on which protocol is having the issue. I'm using sslserver from the ucspi-ssl package so I can offer ssl-encrypted connections as well. The non-encrypted process is set to 300 connections, and the ssl process is at 200. - Michael
Re: [vchkpw] busy try again later error
Michael Johnson wrote: I don't have that file, so I'm using the default (7200 then). How would this have an effect on POP3 connections? ___ Michael Johnson 909-740-3156 x3163 System Administrator [EMAIL PROTECTED] PhD Computing http://www.phdcomputing.net/ Rick Macdougall wrote: Michael Johnson wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 I've been running into an issue on our mail server recently. As traffic has been increasing on it, I've been getting the response that I listed in the subject more often, and now customers are also seeing it (busy try again later!). Of course, this response causes any mail clients to prompt for password, making it rather annoying. Is your timeoutsmtpd set in /var/qmail/control to something lower than 7200 ? (I'm using 180, I know at least one person on this list is using 60). Heh, sorry. I just glanced through it and assumed it was an smtp error. I didn't see any mention of pop3 in there. You could increase the number of allowed pop3 connections though. The default is 40. You can increase it by using the -c xx flag with tcpserver, where xx is your required number of connections. Regards, Rick
Re: [vchkpw] busy try again later error
I don't have that file, so I'm using the default (7200 then). How would this have an effect on POP3 connections? ___ Michael Johnson 909-740-3156 x3163 System Administrator [EMAIL PROTECTED] PhD Computing http://www.phdcomputing.net/ Rick Macdougall wrote: > Michael Johnson wrote: >> -BEGIN PGP SIGNED MESSAGE- >> Hash: SHA1 >> >> I've been running into an issue on our mail server recently. As traffic >> has been increasing on it, I've been getting the response that I listed >> in the subject more often, and now customers are also seeing it (busy >> try again later!). Of course, this response causes any mail clients to >> prompt for password, making it rather annoying. >> > > Is your timeoutsmtpd set in /var/qmail/control to something lower than > 7200 ? (I'm using 180, I know at least one person on this list is using > 60). > > Regards, > > Rick > > >
Re: [vchkpw] busy try again later error
Michael Johnson wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 I've been running into an issue on our mail server recently. As traffic has been increasing on it, I've been getting the response that I listed in the subject more often, and now customers are also seeing it (busy try again later!). Of course, this response causes any mail clients to prompt for password, making it rather annoying. Is your timeoutsmtpd set in /var/qmail/control to something lower than 7200 ? (I'm using 180, I know at least one person on this list is using 60). Regards, Rick
[vchkpw] busy try again later error
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 I've been running into an issue on our mail server recently. As traffic has been increasing on it, I've been getting the response that I listed in the subject more often, and now customers are also seeing it (busy try again later!). Of course, this response causes any mail clients to prompt for password, making it rather annoying. The closest I can find to the problem is an old message on this list back in 2005 from Gaetan Minet (http://www.mail-archive.com/vchkpw@inter7.com/msg22797.html). That particular fix won't help since vpopmail has it already. Does anyone have some suggestions for detecting the cause of this issue? Server Specs: Dual 1.8gHz Xeon Processors (P4 family) 1gB RAM Software: OS: Slackware Linux 11.0 with kernel 2.6.20.3 vpopmail: 5.4.18 Filesystem: ReiserFS with noatime option I am using cdb user maps with pop-before-smtp. I have 2112 domains on the system (counting aliased domains) and 6034 accounts (counting forwards and aliases). I am also using validrcpthosts, but running generation on a cron every 15 minutes instead of via pipe (pipe method was causing slowdowns on qmailadmin). - -- ___ Michael Johnson 909-740-3156 x3163 System Administrator [EMAIL PROTECTED] PhD Computing http://www.phdcomputing.net/ -BEGIN PGP SIGNATURE- iD8DBQFHDR6KehyONlqb1LIRAt9XAKDV0RD/SWrEbMaHZVtVKl83DIDGDwCg81ko ZA+VLuoTWXmdPLyu7yaNcao= =QXgb -END PGP SIGNATURE-
