I have a [bunch of] SME 5.6 with all the updates.
as a precaution I do have a clam+amavis ng on all of them (a good thing - each day they catch a bunch of those pesky worms)
It seems that there is a limit the email can have before it is "frosen" into
the problems directory.
the error in each log has the same description:
Dec 4 17:41:54 sargs amavis[3251]: Unpacking message in /var/spool/amavis-ng/amavis-unpack-3fcf55bb-0cb3
Dec 4 17:41:54 sargs amavis[3251]: AMAVIS: Determined 00000000 to be type message/rfc822
Dec 4 17:41:54 sargs amavis[3251]: Attempting to unpack 00000000 as MIME compliant message
Dec 4 17:41:59 sargs amavis[3251]: AMAVIS::Extract::Mail: Unpacking takes too much space
Dec 4 17:41:59 sargs amavis[3251]: AMAVIS: Error while unpacking 00000000 as message/rfc822
Dec 4 17:41:59 sargs amavis[3251]: AMAVIS: Giving up
Dec 4 17:41:59 sargs amavis[3251]: AMAVIS: Error while unpacking message
Dec 4 17:41:59 sargs amavis[3251]: AMAVIS::MTA::Qmail: Freezing message
/etc/amavis-ng/amavis.conf has:
;; Ignore MIME message extracting errors?
ignore errors = yes
;; What to do if such error occues?
;; freeze - default behaviour, message will be frozen
;; drop+notify - drop message, notify sender
; error action = freeze
where neither freese (the current default action) nor drop and notify will be good.
The system mostly concerned is an 1.7 athlon / 512 MB RAM /10 GB HD @46% full - in theory hardware is not the fault here.
I do not see where can something else been edited for larger "buffers"
P.S. I know that I have to educate my users not to send large emails and use some storage place instead - more asier would be getting some quad processor system through accounting for mail server than make any of them change mind for more than a day.
