logs archiveIRC Archive / Freenode / #exim / 2010 / May / 21 / 1
HM2K
howdy
Re: http://wiki.exim.org/FAQ/Performance/Q1002, it says "deliveries per day", how is this usually measured?
phx
by the log? :)
HM2K
essentially, i'm not exactly sure how to best evaluate the log...
eximstats /var/log/exim_mainlog doesn't exactly tell you anything apart from errors
any ideas how to correctly measure the log to show return a "deliveries per day" figure?
or "deliveries per hour"
phx
deliveris, as local or remote or what kind?
you have to exactly define the task
after that check your configuration, check the log syntax, and grep for the proper thing
HM2K
phx remote
in the above website it says: seeing ... " [230,911 deliveries in a day: 4,475MB] "
i'm thinking there's already something that can evaluate performance
phx
i can understand that. you're seeing a statistical data without further explanation
HM2K
*nod*
phx
you'd like to achive this. you'd like to achive an undefined thing.
HM2K
exactly
best approaching the mailing list?
         

phx
you'd like to do something, that you haven't even decided
it's like programming without a specification
HM2K
the specification is there, defined in the url above
phx
best approaching is to FUCING DEFINE your requirements, EXACTLY
HM2K
i'm asking how to return the figure that is defined within that page
phx
and i'm still asking what i just did
HM2K
are you retarded or something?
phx
nope, that'd be you
HM2K
right...
phx
you can't do something without exact specification. but you'd like to. first you have to realize this
HM2K
it's a simple enough question, if you don't know, it's ok.
dude, the spec is as per the page
i don't understand what can be more simple...
phx
i know. and i already suggested you the sufficient resources to get it done
HM2K
you said use the logs and grep, that's about as useful as a chocolate penis
i already know this...
thus why i asked if you're retarded, you seem unable to hold a civil conversation
phx
i don't, and i never will hand the complete solution to anyone. i suggest where i'd begin researching on the topic. and i said more than that
HM2K
at present, you're acting like it's smoke and mirrors
give me some substance to go on
phx
right.
http://www.exim.org/exim-html-current/doc/html/spec_html/ch49.html
here you go
HM2K
if that's the list of commands i'm gonna shoot you
phx
no, it's not
HM2K
ok, that's a start
<= message arrival
sorry
=> normal message delivery
i presume i need to grep for that
phx
< phx> after that check your configuration, check the log syntax, and grep for the proper thing
beg your pardon?
HM2K
my issue is that there are binaries such as eximstats and eximgrep which one would presume would be a it more helpful
         

phx
sorry but i never used those
HM2K
lol, right
phx
if ever i need to do some statistics, i'd use something like continue= and a database
HM2K
i'm not familiar with continue=
phx
should i suggest the docs, or can you find it yourself? :)
but where is our rack-packed screwdriver
HM2K
sometimes a human interpritation to docs is a bit more useful
it helps with conversation too
i'm not really that interested at present to go hunt it down myself
but, since you mentioned it, i'd be willing to hear your take on it
fgrep "2010-05-20 01" /var/log/exim_mainlog | grep -c '<='
seems a bit of a crude way of returning the value; i'd expect there was some pre-existing method to do this; but i guess not
fgrep "2010-05-20 01" /var/log/exim_mainlog | grep -c '=>'
probably want that for delivered
phx
well, you might want to consider [-=]> for grepping
petemc
eximstats does exactly what you want
phx
depends if you mean deliveries as per messages, or per destinations
HM2K
petemc thanks, you're right; turns out i was using it wrong
eximstats -ne -nr -nt /var/log/exim_mainlog
shows me less verbose; before it was useless.
turns out eximstats only shows per hour, but not per day
... figured out i can pipe eximstats, which is nice
fgrep '2010-05-20' /var/log/exim_mainlog | eximstats | grep Delivered
bonhoffer
any life in here?
phx
nope, sorry
bonhoffer
nice, 2010-05-21 11:29:55 socket bind() to port 25 for address 127.0.0.1 failed: Addre
ss already in use: daemon abandoned
any thoughts on what i can do -- ubuntu server
HM2K
netstat -a | grep :smtp
bonhoffer
tcp 0 0 localhost:smtp *:* LISTEN
online it says i can ignore this error, but it is in the paniclog
and i am not receiving mail on test
is this a conflict with sendmail?
HM2K
yes
bonhoffer
how do i remove sendmail and all its cruft
HM2K
try #ubuntu ?
bonhoffer
so i removed sendmail, how do i reload exim4
phx
are you really sure you want to run a server?
bonhoffer
phx: why not?
phx
with such questions...
HM2K
kill -HUP `cat /var/spool/exim/exim-daemon.pid`
bonhoffer
HM2K: thanks
HM2K: so it is dead -- looks like it was not alive, at least not from ps aux | grep exim
how do i give it a try again -- do i need to go through the whole config process
echo "test this" | mail -s testing tim@gmail.com <-- no error, but no message
is there any other way to test?
sorry folks, is this a help channel or just for dev's?
phx
it's a support channel for exim, and not for different linux distributions
so far you had questions related for a given distribution
bonhoffer
actually, i am just curious how i can test to see if exim works? piping data to the mail command?
can i ignore the error: socket bind() to port 25 for address 127.0.0.1 failed: Address already in use: daemon abandoned
phx
first, check if the process is running
if it's running, check whether it's bound to all the specified ports
you can ignore that message as far as you don't need it on that port, obviously
if it's running on all the ports, you can start doing case tests for different ACLs, routers, transports and such, this part is testing your configuration itself
bonhoffer
no, it is not running, but the mail command doesn't return an error -- the binary /usr/sbin/exim doesn't start -- mail is loaded and runs
can i see if mail maps to exim?
exim is running, but mail messages are not sending
phx
what's on 25/tcp? what mta provides the sendmail interface?
petemc
bonhoffer: readlink -f $(which sendmail)
lsof -i :25
bonhoffer
yeah it is /usr/sbin/exim4
endmail- 13871 root 4u IPv4 83158 0t0 TCP localhost:smtp (LISTEN)
that is sendmail . . . result of lsof -i :25
petemc
did you install something outside of the packaging system?
dpkg wont let you install 2 things that provide mail-transport-agent
bonhoffer
petemc: no, but i did install sendmail first, then uninstalled it
just ran dpkg again
followed: http://library.linode.com/email/exim/send-only-mta-ubuntu-9.10-karmic
currently working through http://wiki.exim.org/TestingExim
everything seems to be working, but i am not getting the mails
petemc
what do the logs say?
bonhoffer
petemc: sorry, the panic log?
petemc
mainlog
there shouldnt be anything in paniclog
bonhoffer
looks good to me: http://pastie.org/971037
would it possibly be taking 10 minutes ?
petemc
it was delivered to google
it shouldnt, checked the spam folder?
bonhoffer
ha! -- yes -- they are all there
would have solved all my problems . . .. wow
thanks guys
mrball
Here's one for you. I created a hostlist:
hostlist do_greylist_hosts = !^.*\\d+\.com\$ : ^.*\\d+[x.-]\\d+[x.-]\\d+[x.-] : ^ppp : ^a?dsl : [.-]dynamic[.-] : [.-]a?dsl[.-]
I use it in a block of code to determine if I'm going to greylist an incoming connection. Pretty straightforward.
I missed something though. The above is from a working server. On a second server, I had an older testing version at the end of the line that looked like: \.dynamic\. : \.a?dsl\.
It caused an error that is not clear to me:
>>> check hosts = +do_greylist_hosts
>>> no IP address found for host .dynamic. (during SMTP connection from mail.webtoolset.net [38.101.255.9])
If I change \.dynamic\. to \\.dynamic\\. then I do not get that error and it works as expected. I am just trying to grok why \.dynamic\. results in that weird dns lookup error, when \\.dynamic\\. and [.-]dynamic[.-] do not give me an error.
Simon-
if it's not a regex it's a hostname that will be resolved
\. is not a valid escape sequence
sobersabre
hi.
phx: can you watch svg files ?
my network setup is: http://yfrog.com/0omailerfp
router forwards port 25 and exchange connections to "srv"
srv has internal address, and runs a vm with the "projects" hostname.
that machine runs exim4
I want to be able to send from that internal vm emails to my domain users.
currently it doesn't work.
I think I need to setup the srv server (which is exchange 2007)
I mean I need to configure it differently.
but I don't know how.
basically I want exchange to behave as smart host (If I understand the terminology correctly)
« prev