Hi All, Hoping someone can help me out with a bit of advice.
I have a custom script that calls postfixs sendmail to send approx 3k emails, and its quick enough that they can all inject into postfixs maildrop queue in under a minute - happy days. Now, in some cases the emails all go into the active queue (seen by running mailq and qshape active) and are pretty much gone within the next minute - alls good. However at other times they can take 30 minutes or more to all go. Now in this instance the maildrop folder holds all the emails, but running qshape maildrop I get alot of "Use of uninitialized value $qt in subtraction (-) at /usr/sbin/qshape line 282." errors, and then the normal qshape table output - but the table total count is less than the number of files in the maildrop folder. Active (10-20), hold(0) and deferred (150) queues are all small/unchanged as well. (reviewing this email before sending I realise I'm not checking incoming so I guess I can try that). I know the server is under some load at this time but I dont believe its massive. I've got bind9 DNS cache running locally, clamd and opendkim - however it doesnt seem likely these are causing the slowness when the emails do send quickly sometimes. postfix reload doesnt fix anything. restarting any of the dependant services doesnt fix anything. The only thing that does seem to work is rebooting the entire server. Upon rebooting the number of emails in the queue seem to be a lot less than prior to reboot (as quickly as I can check that is), and all the emails are in the active queue and go out as quickly as they can. I'm at a bit of a loss as to what the underlying problem is? If any body can offer any suggestions or setting optimisations I can try I would be very grateful as this is driving me mad having to reboot the server every morning Thanks Mark
