> This is a logging (cosmetic) issue: SPF pass got a bonus
> value and the proper text was not set. It should be fixed since
> 1.01
Uhm... on a second thought, that domain has no MX/A record
so how could it pass the SPF test and get that bonus score ??
Mar-23-09 07:20:17 msg-92813-07954 [MissingM
> This is a logging (cosmetic) issue: SPF pass got a
> bonus value and the proper text was not set.
>
> It should be fixed since 1.01
Ok, I see, not a problem but seeing the MXA increasing
and then decreasing the score was a bit confusing :)
I was looking at the ASSP logs and found the following
(note replaced the real recipient with u...@example.com)
<===>
Mar-23-09 07:20:17 msg-92813-07954 213.161.194.44
to: u...@example.com
MessageScore is now 10, after adding 10 (Fore
>> here's a somewhat "longer" list you may use to fill
>> your "whiteorg.txt" file (the use used to whitelist
>> organization names through senderbase lookups)
> Awesome, thanks.
You're most welcome, I just hope someone else will add
more entries.. and so on, while that list is an "optional",
pop
> [RelayAttempt] 192.168.0.60 to:
> externalem...@address.com relay attempt blocked for unknown local sender
> domain
and it's ok, if you've a screwed config, ASSP can't fix it
--
Apps built with the Adobe(R) Flex(R) f
> I am totally confused. I just updated to ClamAV 0.95 RC2 for use with
> 1.5.1.0 final and enabled Scan Whitelisted Senders and set Suspicious
> Virus to UNOFFICIAL|Safebrowsing\. and sent the test email below that
> contains an unsafe url.
on a second thought, did you also enable the "scanlocal"
> Why isn't assp detecting and scoring this? What can I do to help
test?
enable the "clamd" logging and fire up a "tail -f" on the clamd log, do
the
same for the maillog.txt, then use an external IP to send a message
containing an "unsafe" URL to the ASSP box and keep an eye on the
logs, let's se
>> Name:tcool
>> URL: http://ppm.tcool.org/archives/package.xml
> Thanks
Y/W, I discovered it myself while trying to install a module
which wasn't (apparently) available for the windows platform :)
also, since we're at it, I'd suggest to run (from time to time) the
ppm upgrade --install
com
here's a somewhat "longer" list you may use to fill
your "whiteorg.txt" file (the use used to whitelist
organization names through senderbase lookups)
<= LIST BELOW ===>
3Com Corporation
AMERICAN EXPRESS
Apple Computer
Avira GmbH
Bank of America
Capital One Financial
Central In
> Need tagging of subject for easy parse in external programm
sounds quite braindead to me, the same subject may be used
by whatever host and would "pollute" whatever you're using the
collected messages for; if you could detail what you are trying
to achieve, maye you'll get better help
> I'm running 1.08 and have been having problems with hosts timing out
> when trying to make a SSL-->TLS connection. Where are you getting 1.13
> from?
add the following repository:
Name:tcool
URL: http://ppm.tcool.org/archives/package.xml
then run a
ppm s IO::Socket::SSL
HTH :)
> The email address '
> newsletter-accept-1237286985.17931.dfmijekagoaeamlkp...@domain.com'
> is too long. We accepting up to 32 characters for a '*' in a rexec!
> So try the following:
Uhm ... but won't such a limitation generate problems with SRS ?
---
>> I'd switch 1<->2 I mean, FIRST change the password, THEN
>> proceed to the cleanup so that, during the cleanup the account
>> won't be abused by other bots
> But new pass may be also leak while virus continue executing
> on infected computer
Uh... how ? See, you change the email password on
>> One of my users had their PC compromised,
>> and their SMTP password stolen by a trojan.
>> This was then used to log into my server from zombie
>> spam bots and send massive amounts of spam all
>> over the world.
> 1) Remove trojan by antivirus software ( or simply reinstall OS)
>
> 2) ch
> ASSP 1.5.1 should go public this week - so please run the latest
> version.
Running RC 16.00 right now, and also updated the rebuildspamdb
to the latest version online, no particular issues as far as I can tell
and imHo it's ready for prime time, so freeze any feature request
and set it ready fo
> Permissions seem fine to me. The file is already created. Surely the
> problem is the fact there is no such place as "./c:\" on Windows.
Looked at the source, the function loading the parameters from
assp.cfg is already "expanding" the file pathname, but then, for
whatever reason if gets "p
> The problem is that 1) headers are supposed to provide a "chain of
> custody" per the RFC's and 2) RFC 821/2821 para 3.6 requires the
> domain name given with the helo/ehlo must be the primary host name
> (or a domain literal only if the host has no name) and assp fulfills
> neither critera a
> Here's something I haven't seen before. In my ASSP log:
> Mar-13-09 12:59:51 [startup] Info: try loading plugin ASSP_OCR
> Mar-13-09 12:59:51 [startup] ASSP_OCR: can not find convert from
> ImageMagick in PATH!? No images wiil be processed!
> The thing is...it IS in the path. In fact, it is th
> I upgraded my installation to Perl 5.10, and installed ASSP 2.0.0
> 15.06. Running with 30 workers, I started ASSP up. It sits silent and
> deaf for about 3 minutes, as it gradually starts grabbing more and
> more memory. During this time nothing is logged. It finally started
> up, just before 12
> Here is what I have configured. None of these is pointing to discard
> folder - so why is spam still being written out there?
What about pointing the discard folder to /dev/null ?
--
Apps built with the Adobe(R) Flex
> Rebuildspamdb got some extra cleaning features:
>
> discarded folder & viruslog folder are cleaned (records
> older 30 days are removed)
>
> I am unsure what to do with the error folders.
Well... given the fact that the "errors" (spam/notspam) folders
are used to "balance" the bayes database I'
> When I exchange the 15.02 with the 15.03 version, I get an error message,
> that windows cannot start ASSP as a service.
Try running it from cmd prompt
perl assp.pl
and see if it logs any error message
--
Apps buil
> If I am not relaying then why is it working currently?
not having your full config file, and since my crystal ball
is out for repairing I find it difficult to answer such a question
for sure it would be useful to know how your incoming and
outgoing email flows are configured
--
If you think th
> relayHost:=
> relayPort:=
so you didn't set relayHost nor relayPort, I think that it will
be hard to have ASSP relaying your messages
--
Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
powering
> I am still having poll read handles errors occurring on my 2.0.0
> (14.11) (Win2003 Perl 5.8.9) install. The main difference is that
> ASSP is not dying from the errors anymore. However, I now have a new
> problem. When the error occurs, it is when a user first connects,
> before the HELO even o
> I have some ideas in the "virus" handling area.
Uhm... regexp filtering on the result ?
If so, a bit of caution please, the 1.x series is single
threaded and adding such a check may have some
serious impact on it; on the other hand, a better way
to check for ClamD availability would be welcome
>> reports to postmas...@localdomain
> In der Regel sollte das funktionieren, außer wenn die Filter auf
> "early" gestellt sind, dann gibt es noch keinen Empfänger.
Uhm... so if the "early detect" is on the "LOCALDOMAIN" won't be
expanded ?
-
>> that log entry means that the sendind session generated more
>> than 5 errors (e.g. invalid commands and so on);
> Soory to correct but that is not the reason. We are counting 550er
> errors.
yes... and the sender generated more that FIVE errors so ASSP
terminated the session since it exceed
> Mar-7-09 18:02:24 [Worker_1] [MaxErrors] max errors (5) exceeded --
> dropping connection
> I am seeing these lines a lot and getting complains from customers
> which are sending email get a NDR but mail is delivered. Is it
> possible to get some sort of debugging and check what is going wrong
> I want ed to place the logs in /private/var/log/assp but there is no
> place in the GUI to specify local logging paths. My only other option
> is to create a symlink
I don't think there's a need to change the assp code to handle that; you
may just create a link and point the folder elsewhere - a
> Besides that, whitelisting the microsoft.com domain seem to me very
> unwise. But is is your call to open such wide doors.
Oh.. I found a trick for that which allows me to get "genuine" Microsoft
messages and discard all the spoofed ones :)
add Microsoft to the senderbase whitelisted address an
I've been testing it for quite a while and decided it's
a keeper (at least for me); I'm referring to
http://www.wpbl.info/
while I won't suggest using it at "level 1" (blocking)
putting it at "level 2" will help filtering a bunch of
spamming sources; in such a case, just add to your
DNSBL file th
> ==
> perl assp.pl -u will uninstall, but not start assp
> ==
<<>> that's a short "help text" not a dictionary
ok, I think that "perl assp.pl -u will uninstall the service"
will suffice :P
--
Open Source Busi
What follows are a some tricks I learned quite some time ago
and which helped me to fine-tune ASSP performance on win32
platform
* Increase clamd priority: sometimes clamD may slowdown the
mail processing since it will ask for CPU cycles to scan an
incoming message and may start fighting wit
> no the server works ok , the problems starts using 12.09 or 13.00
> which eats all available cpu resources.
> I will try disabling some feature adn I will report ,
> and yes I am using SSL (also in 09.03).
could you please try enabling "DEBUG" and then running a
couple of "tail -f" one against
> yepp, i noticed when checking the various archives. no idea where i
got
> the current file from but i've changed the content of "denysmtp.txt"
to
> the IP-blacklist from http://www.whitehat.ch/bogon/bogon-ips.html and
added the
> bogon-ips from there to denyalways.txt and everything seems to wor
> I take that back, it was posted on Fritz's download page
nope, got those files and neither of those had such a screwed
format nor those entries; again, looks like a cache file or a pb
one which got copied over the "blocking list"... now don't ask
me how it happened
> I have attached the file that is causing the errors.
Heck, that's quite much screwed, the IPs should be
one on each line, and optionally with a CIDR mask,
in your case you have more IPs on the same line
and a lot of intermixed blank lines :(
Also, I can't understand WHERE you got that file
fr
> The only thing I can think of to cause this is "permissions".
> Just do something like
> cd ASSP; chmod 777 *
that assp is on windows... I saw such an issue in some previous
versions, upon startup assp got somewhat "stuck" eating up CPU
but apparently doing nothing, and, after a while it eventu
> same problem here, see also
> http://www.nabble.com/ASSP-1.5.1-rc-Freezing-to22230370.html - even
from the
> command line (with asaservice=0) there are no messages but a simple
> freeze.
been running it fine for several hours now; PLEASE try the following
(just to try nailing the issue) - rename
> i'm still facing the problem that assp.pl > 1.5.1(RC 0.10.05) won't
start
> in my windows xp/perl 5.10 build 1004 enviroment (no messages, no logs
> even when trying to launch from the console)
If assp is installed as a service, if you want to test if from the
console you will
need to manually e
As for subject I downloaded the above version and
gave it a spin; the platform is:
win2k3, Intel Xeon 2.8 Ghz, 2Gb RAM
Active Perl 5.8.8; all needed modules installed (or at least
the ASSP logged an "all ok"); ASSP running as a service
Started up the ASSP, it took some seconds to startup
all t
> Obviously we cought an ugly bug.
> To reach a better footing, we want to go back to 10.5 first. For that
> reason there is now 11.5 on my site which is the same as 10.5 but
> renamed.
> Would everybody please switch to 11.5 and report (hopefully) success?
seems to work ok; I also did some profil
> - ASSP logging stops.
just in case...
install Win32::API::OutputDebugString by running
ppm s Win32::API::OutputDebugString
and then "ppm i #" where "#" is the number of the
newest listed version of the package
download debugview from this page
http://technet.microsoft.com/en-us/sysinternal
> Since it is not recommended to forward spam since it makes ASSP
> think you like it. Would it be possible to augment report-spam
> handling to forward a copy to spamcop just like I manually do it
> today from my mail client.
add the spamcop submission address to the redlist
-
> [Main_Thread] Error: IO-subsystem error - unable to poll handles on OS
[...]
> if something has changed. Either way, 14.02 is still unusable for me.
Yes... :( gave it a spin but it died with the same error you saw, so I
was
forced to rollback; again, I'd really *love* to have a running 2.x but
u
> Sanesecurity\.Junk\.9851=>WHITELIST,0
UH... so you got hit from the 9851 issue ?
--
Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA
-OSBC tackles the biggest issue in open source: Open Sou
tried to install the latest one (from site), started up
just fine... took some time to initialize threads (as
usual) but came up; after about 5 minutes I got the
following:
Feb-25-09 12:07:23 msg-63621-10618 [Worker_2] 202.63.186.169
to: u...@domain.tld recipient delayed:
u...@domain.tld
Feb-25
http://wiki.asrg.sp.am/wiki/Main_Page
http://spamlinks.net/
various documents, links ideas and more
about spam and spam filtering; may be
useful to gather ideas...
--
Open Source Business Conference (OSBC), March 24-25,
> This is a trap address. All trap addresses are being logged as invalid
> instead of [Trap]
switch penalty logging to verbose; ASSP is still correctly handling traps;
to detect that
try changing NoValidRecipient to "550 5.1.1 Recipient address rejected
EMAILADDRESS"
and then look at the rejec
> "some time-consuming tasks"
> which ones?
look at the source code :D
--
Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA
-OSBC tackles the biggest issue in open source: Open Sourcing the En
> what is the point of this if i can have 37 threads with 2.0?
> what does thread 1 do in 1.5.2?
> what does thread 2 do in 1.5.2?
simple, the 2.0 if fully multithreaded, but that means that
it needs a lot more memory; the scope for the additional
threads in 1.5 is just to move some time-consumi
> On the server I was testing it on, it is a quad-core server with 4 GB
> RAM, and had at least 2 GB free when 2.0 would start showing the
> "poll" errors you mention, and then become unresponsive. It was
> utilizing all four CPU's, however not excessively - the server was
> not being strained
> 5.8.9 ? use 5.8.8 or 5.10 - I can not say if 5.8.9 is running well
> but I know that 5.8.8 - 822 and 5.10 - 1004 are running very well.
not in my case (and apparently in other cases) after a while,
ASSP 2.0 dies or, better said, becomes dumb and deaf; in
most cases, before such a thing happens
>> I wonder *where* that "earlytalker" session was rejected; there's
>> no reject message, and looking at the message score the other
>> entries (above and below the earlytalker one) seem unrelated...
>> so... where did our "early talker" connection end up :) ?
> It was delayed.
delayed ?!? But
>> bottom line, if you want to use KS, set it to level 3 or 4
>> but avoid using level 1 or 2 or you'll see a whole lot of
>> false positives
> The rule of thumb with postfix when using multiple RBLs and
> also using spamhaus, is to NOT use individual RBLs that are
> also used by spamhaus.
well
> I wonder *where* that "earlytalker" session was rejected; there's
> no reject message, and looking at the message score the other
> entries (above and below the earlytalker one) seem unrelated...
> so... where did our "early talker" connection end up :) ?
here's another one
Feb-19-09 10:45:51 8
> yep, the message scoring is superfluous.
> The way this is shown in log is caused by the way
> we detect the early talking. I am not willing to change
> the ASSP logic a lot. Look into the code and search
> for earlytalking,
did that, yet it sounds like there's something wrong, at
least judgin
> dnswl is already used with a "trust" value derived from the
> last digit. Karmashere was tested, but not accepted.
about "karmasphere"; if someone is using the DNSBL (not
the whitelist) with 1.5 in weighted mode... be careful, since
it may cause a number of false positives, the reason is
more
> After update to latest version 2 of ASSP, I have a problem to log on
to
> Web GUI. The explanation on the site is not so clear.
> What are the values the assp.cfg should include in order to define a
> admin user.
use the same password as for previous version, but ensure to
use "root" as the user
> yep, the message scoring is superfluous.
well... a dropdown with a "disabled/score/block/monitor" won't hurt
for sure, but in general, I think that an early talker shouldn't be
treated
as a regular SMTP server or client; both will follow RFC and wait for
our server banner before sending out an
>> I also noticed another thing; most early talkers also hit the
>> "forged helo"
> Only because you have the "forced" options in effect and do forged and
> invalid helo checks before delaying. try to set "force" to off.
I may do that; but in my case the delaying "tuples" list would grow up
a who
> that is, apparently the earlytalker message is logged AFTER the
> rejection also, I think that just *dropping* the connection for early
> talkers would be a good idea ;)
just in case, here's another log snippet; i grepped the logs for the
offending IP and here are the lines
Feb-18-09 14:33:57 m
> well... got some of them as for my previous email, I just can't
> understand why the log says...
I also noticed another thing; most early talkers also hit the
"forged helo" or "invalid helo"; now; since those checks are
performed using regexp, so, even if in minimal part they do
require CPU cycl
> This discussion pops up every 6 month or so, there is now
> a PenaltyBox score which can be assigned to an early talker.
Thanks
> Look for a message "earlytalker". You get a price from me ))).
well... got some of them as for my previous email, I just can't
understand why the log says...
Feb-1
> This discussion pops up every 6 month or so, there is now a PenaltyBox
> score which can be assigned to an early talker. Look for a message
> "earlytalker". You get a price from me ))).
Feb-18-09 14:28:06 msg-67286-08218 95.132.124.47
MessageScore is now 150, after adding 150
(forged Helo: 'xxx
FB>> Tried two years ago. Not worth the hassle IMHO.
> Yes. It had no measurable effect.
> But with this delaying, you will produce many open
> connections and much more load on ASSP and your
> MTA.
I see... well, again, I was just reading that page and at first
it didn't seem a bad idea, bu
> But with this delaying, you will produce many
> open connections and much more load on ASSP
> and your MTA.
uhm... on a second thought.. the delay isn't needed
I mean...
* ASSP receives a connection request
* The connection is accepted
* ASSP, before proxying the connection to the local
I was reading this
http://free.acrconsulting.co.uk/email/other.html
and sounds like it shouldn't be so difficult to implement
in ASSP; the idea is that, if a given external host which
connects to ASSP starts sending out commands before
receiving the initial banner, then it should be considered
a
> I assume that there is some unused Cache Memory
> or some lists are growing. As I use MySQL all list
> should be handled by the SQL Server?
hmm... didn't notice that (MySQL); still I suspect that some
lists aren't handled by the backend MySQL and kept in
memory; worst ... I suspect there are m
> Another thing to check:
> Do you have enough free RAM?
> I need minimum of 500MB, going up to 1GB
>
> Maybe GreyHats problem is connected with it.
> If you add more workers, ASSP will need much
> more RAM
Hmm... maybe that's the right track; see, the "problem
machine" isn't "dedicated" to ASSP
> [Worker_x] Error: IO-subsystem error - unable to poll handles on OS
saw exactly the same error with 2.0; in my case ASSP starts up fine
as a service, but after a certain time, it starts logging those "poll"
error messages and then becomes totally deaf; trying to stop the
service at that point d
> SNMP adds an added level of complexity considering MIBS
> to be updated etc. While PRTG looks good being a commercial
> product the vendor could include support for ASSP. Does PRTG
> have scripting ability of any kind?
I think that someone with decent Perl and SNMP knowledge may just
add a sep
>I think we will flag the file "deleted" instead of actually
> deleting it.
Just an idea; prefix the file name with a "_" and on the next
run, delete all the "_" prefixed files :)
--
Open Source Business Conference (OSB
> 12.01 is running fine!
Pity isn't available anymore, otherwise I may give
it a spin and check if it runs ok for me too...
--
Create and Deploy Rich Internet Apps outside the browser with Adobe(R)AIR(TM)
software. With
> In the afternoon I installed 12.06, but now I revert to
> 12.01 as ASSP Crashes every 30-90 min
That's more or less what I saw; in my case, though,
before the hangup ASSP logs a flow of "IO errors"
and then just "hangs"
-
> http://technet.microsoft.com/en-us/sysinternals/bb896647.aspx
> if all's ok you'll then see a "live log" from ASSP
> As a note, the same directions also work for 1.5 :)
Forgot; you'll need to install an additional "ppm";
open up a cmd prompt and issue the command
ppm s Win32::API::OutputDeb
> I'm running on Windows 2008, seeing similar instability, though
> I'm not getting the 'Error: IO-subsystem error - unable to poll
> handles on OS' messages in the log when it dies. I've got TLS
> set to drop and data stored in MySQL 5.1.30. The service runs
> fine for between 2 and 4 days, but
>> assp should send back a "250 Ok" and wait for the client closing
>> the connection before starting report generation
> try to set " store and run at midnight"
<<>> ok; would it be possible to have an option to set
a given generation time instead of midnight ? I'm asking so
since, in some cases
> Just another reason why Firefox rulez! Thanks for the link.
well... with whatever other browser, just add a querystring
(e.g. ?x=1&y=2) to the link and you'll skip the caching
--
Create and Deploy Rich Internet Apps
> Try to install MySQL and switch all files to the Database.
> After doing this, ASSP runs fine, besides the TLS Problem
> - but maybe even this is fixed in the actual version
> - I didn't try.
I can't go to MySQL right now; trying to upgrade from 1.5 to
2.0; have MSSQL on a box, so I may switc
> We developed ASSP2 to do what you are proposing and
> it does it. I am not planning to do the same in a different way
> (but more limited) in V1.
Well... for sure the "blckreports" is a "plus", but the "watchdog"
probably isn't; that would solve any kind of crash or lock issue
and shouldn't b
> not sure, at any rate, I'm retrying now; I switched the
> TLS to "drop"... let's see if this cures the issue...
no way, even switching TLS to off the problem still
resurfaces, after an hour or so, ASSP starts logging
those darn "unable to poll handles" and then becomes
deaf and dumb; it doesn't
> I noticed, that this happend when I enable doTLS.
> Maybe it is a memory issue?
not sure, at any rate, I'm retrying now; I switched the
TLS to "drop"... let's see if this cures the issue...
As an additional note, here's a log snipped from
the 2.0 startup on that box... just in case it may be o
>> nope, just move the RBL to level 1 so that one hits
>> will result in a reject
> That would imply that one trusts the RBL implicitly.
> Whereas, getting (two hits =>2) or (1 hit =>2 and 2 hits =>3)
> does not require implict trust.
Well... maybe I misunderstood, but reading the original mail
I know, maybe it's late, since the final release seems
to be ready, yet, I think that the ideas below may be
worth
One of the main issues when running the 1.5 is the fact
that it's difficult to restart it if it "bombs"; now, there is an
option to automatically restart the assp (either running
as
I've been trying to track this issue, but till now
I was unable to find out what's causing it; running
ASSP 2.0 on Win2k3 (the 1.5 just runs fine on it)
whith ActivePerl 5.8.8; ASSP starts fine, running
as a service and stays up and running for quite
a while, then, suddenly, for no apparent reason
> ;) Now if we could get rid of spam by posting an inter-galactic notice
and
> building a spacelane...
> Okay. The text under RBLmaxhits says hits. When I saw that I was
guessing
> score. I should adjust my scores so that any two are above the
threshold.
nope, just move the RBL to level 1 so that
> in SuspiciousVirus ; ASSP is still blocking the email instead to
> assign the Virus Suspicious Score (25) when a Sanesecurity
> or SecuriteInfo email is found.
Just curious... why would you score such messages instead of
directly blocking them ? Also... consider that the sane signatures
al
I find the "senderbase whitelist" rather useful, and I've built
a list of "whitelisted organizations" which I want to share:
<= COPY BELOW
3Com Corporation
AMERICAN EXPRESS
Apple Computer
Avira GmbH
Bank of America
CHARLES SCHWAB & CO\.
Chrysler Motors Corporation
CISCO SYSTEMS
CloudMark
In case someone doesn't have a local DNS resolver and
needs one, below you'll find a list of "open" DNS resolvers
usable from whatever IP address; checked the list days ago
pruning all the "not working" ones, so they should be pretty
much ok; I know, there are the OpenDNS ones, but sometimes
one m
> Started up V1.5.1 RC 2.05 at about 7:00 PM. It just crashed at 11:21
PM.
> Until that time, it seemed to be processing messages just fine. It did
> seem to clear the DNSBL Cache fairly often, perhaps every pass through
the
> main loop. And there appears to be some contradictions. Note that the
>
> Grayhat,
Hi there :)
> I would like nothing better than to block, but that is a
> political decision as opposed to a technical decision.
Uhm... I see... between hammer and anvil huh :) ?
Anyways, that's up to "you", but not exactly the purpose
for which ASSP was wri
> If you are talking about clamd, this is a knows problem.
> The clamd binary is not working well
Uh ? How comes ? I've ASSP and ClamD running w/o
any particular issue; could you please expand that
"not working well" ?
As a note; if you're referring to slowdowns; those may
happen at times; but co
> to: x...@mathbox.com [monitoring]
(MessageScore
> 100, limit 50)
on a second thought... try setting "DoPenalty" to "monitor" and
"DoPenaltyMessage"
to "block"; and apply the changes; the above message should have been
BLOCKED
due to spamscore over limit, but apparently you have assp configured
> I have watched ASSP crash 3 times now. Memory usage
> was growing from a normal of 87MB to 1100MB.
Had it running from yesterday 19:20 (more or less) till now on
two boxes, both are ok, no eccessive memory usage nor
high CPU usage (aside from the regular one); ok, I'm not
using the SSL port (alt
> OK, it was .14 by the time I got to it, and yes - on first look it
does
> appear to have fixed the issue. However, .14 is really unstable - it
> crashes here every few minutes, making it unusable.
are you using LDAP or VRFY for the recipient validation ?
--
> the verify code corrupted the ldap code.
>
> so I fixed first the old situation. I think, .12 will
> work with ldap and flat and not with verify.
ok, so I'll stay with .08 until VRFY won't be back
--
This SF.net emai
> please try .12
no way; I'm using VRFY (not ldap) still ASSP .12 rejects
valids recipients, rolling back to .08 fixed the issue, so,
I think there must be a bug into the validation code for
both VRFY and LDAP (since I'm using VRFY while the
other poster is using LDAP)
-
> in short, if you query URIBL "too much" (don't ask me how
> much is it) their DNS will start returning a 127.0.0.255 for
> ANY host, and this may cause a whole LOT of false hits
> for innocent machines/uris
for further details, please see
http://www.uribl.com/faq.shtml#q4
---
801 - 900 of 1015 matches
Mail list logo
