LWAYS get it by chunk_ind order?
I would not bet on this ;-) However, while ORDER is usually a very "bad"
command, it should not hurt much in this case, as the resultset is very
small.
Best regards,
Thomas Gelf
--
mail: tho...@gelf.net
web: http://thomas.gelf.net/
-
opinions and to get some feedback!
Best regards,
Thomas Gelf
NB: No, I'm not sorry for spamming and crossposting ;-p
--
mail: tho...@gelf.net
web: http://thomas.gelf.net/
--
Let Crystal Reports handle the re
by defining your very personal variants
in your amavisd.conf.
Best regards,
Thomas Gelf
--
mail: tho...@gelf.net
web: http://thomas.gelf.net/
--
Let Crystal Reports handle the reporting - Free Crystal Reports
Hi Mark!
Mark Martinec wrote:
> Unless you really need to shave off the few milliseconds, I'd suggest
> NOT to move temp directory to a ram disk, which is typically much
> smaller than a real disk partition, so can fill up much faster and
> block mail flow if something unpredictable happens.
Rega
Thomas Gelf wrote:
> Using ADSP will save you a lot of manual work - manually configuring
> DKIM behaviour per domain makes no fun.
NB: ADSP has been published as RFC a few days ago as RFC5617
http://tools.ietf.org/search/r
how_bug.cgi?id=6087
Regards,
Thomas Gelf
--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do b
Mark Martinec wrote:
>> I'm actually kinda leaning towards 52
>> partitions at this point for maximum speed and flexibility
>> while using iso8601 week as the value.
>
> Not to forget that some years can have 53 weeks (1..53)
> according to ISO 8601.
>> 52 partitions -> I've also been evaluating
TABLE quarantine DROP PARTITION p3;
> ALTER TABLE quarantine ADD PARTITION (PARTITION p3 VALUES IN
> (4,8,12,16,20,24,28,32,36,40,44,48,52));
Will work as expected. Cleanup took 4 entire days 'til being finished
on the last slave before using partitions. Now it takes just some
seconds :-
ction when comparing
values.
If running a large DB you should definitively change column types, as
using string functions in your WHERE statements is seldom a good idea
(performance-wise).
Best regards,
Thomas Gelf
Luis Daniel Lucio Quiroz wrote:
> Hi all,
>
> Having SQL schema, in wha
Peer Heinlein wrote:
> Am Donnerstag 09 Juli 2009 schrieb Thomas Gelf:
>
>> As far as I remember all docs I used long time ago to set up my first
>> Amavis boxes clearly state that it is absolutely necessary to avoid
>> REJECTs by the postfix instance Amavis is rei
ecks AFTER pre_queue filters. I do not have any experience with
such setups, so no idea if this assumption is true.
Nonetheless I cannot immagine any special case where this DSNs (in case
of REJECT) from Amavis would make sense - so as you did I'd also opt for
suppressing them.
Best regards,
Michael Scheidell wrote:
> I guess you don't need the 'bigint(20) on id if the id is unique only on
> each partition, and you need to add partition_tag to primary key (since
> a list partition has to use primary keys)
Yes and no: primary key is defined as PRIMARY KEY (id, partition_tag),
however
Xueron Nee wrote:
> Is there any macros can be used in log_templ point to the Helo/Ehlo
> domain passed from upstream?
Not sure how your setup looks like - but usually this is a task that
should be accomplished by your front MX server (Postfix of whatever).
Regards,
Thoma
tion from 4 days to 4 seconds.
Regards,
Thomas Gelf
--
Enter the BlackBerry Developer Challenge
This is your chance to win up to $100,000 in prizes! For a limited time,
vendors submitting new applications to BlackBerry Ap
Michael Scheidell wrote:
> I looked at the readme's and some of the stuff Thomas Gelf wrote about
> mysql partitioning and amavisd partitioning, and tried a couple of them
> without much success.
>
> Did anyone write a 'setup/ quarantine cleanup' cheat sheet
#x27;ve never used it before.
To accomplish what you are suggesting it should IMO suffice to "replace"
penpals_find() (is it possible to do so?) with some custom code reading
from Memcache - plusadding something int before_send(), storing
penpal-information.
Cheers,
Thomas Gelf
-
Mark Martinec wrote:
> Perhaps m...@customer.tld has bypass_virus_checks while o...@mailbox.tld
> does not? Elevated log level would tell. (but see further on)
No, that's what we immediately verified - as it's the most obvious
explaination.
> Having qr'^MAIL$' in @keep_decoded_original_maps seems
se) could also be an option.
That's all so far. I'm pretty sure I've forgotten something - but I'm
confident one of you will for sure find it ;-) Your feedback is more
than welcome, and it doesn't need to be positive - feel free to tell me
why you consider the proposed approa
st seemed reasonable to me. But as run_av does
not even be called for the decoded MIME parts that's probably not the
issue here.
Any suggestions?
Best regards,
Thomas Gelf
--
Crystal Reports - New Free Runtime and
the process to terminate: %s\n",
+ printf STDERR ("Waiting for the process to terminate: %s\n",
join(', ',@killed)) if @killed;
}
$| = 1;
@@ -280,5 +280,5 @@
$cursor->c_close if defined $cursor; # ignoring status
$db->db_close==0 o
led as
follows:
amavisd-nanny -c 1 >/dev/null
That's all, feedback / improvements are welcome.
Best regards,
Thomas Gelf
--
The NEW KODAK i700 Series Scanners deliver under ANY circumstances! Your
production scanning e
Wow, thank you!!
Cheers,
Thomas Gelf
-
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/ser
t; -> ) ENGINE=InnoDB;
> Query OK, 0 rows affected (0.03 sec)
You can't create table quarantine referencing table msgs as long as msgs
doesn't exist. Could this be the root your problem?
Cheers,
Thomas Gelf
-
Ed W schrieb:
> Thomas Gelf wrote:
>> [part 3 of: "Scalable amavisd-new installation"]
>>
>> There is also a self-written component I'm really proud of: the log
>> aggregation system.
>
> Funnily enough I need to now build something similar to th
even if someone is
using aX3C4f6btXgX instead of aX3C4f6btXgX[25] in scenarios where he
should use partition_tag.
Thanks a lot for the great work!!
Thomas Gelf
PS: Thanks also for mentioning myself in the release notes!
PPS: Most work I spent has not been for those little changes but for
--- apps
One could also add
quar_type binary(1),
even if it is it not really needed right now.
Cheers,
Thomas Gelf
-
Check out the new SourceForge.net Marketplace.
It's the best place to buy o
ou want to do
with your quarantine. So just copy & paste default db scheme would
not really help - but such an example could be a hint of how things
COULD be done.
Regards,
Thomas Gelf
Leonardo Rodrigues Magalhães schrieb:
> Please check this message i sent last week, rega
ted MySQL bug - but
it could really be something like that - if both threads are getting
the same autoincremental id - and BOTH INSERT's are therefore failing.
Such behaviour could match your description of amavisd-new's logic.
Don't know whether this i
Thomas Gelf schrieb:
> Thank you for your great work!! I'll switch my mail-traffic away from
> one of my filters within the next hours and test some domain on the
> patched amavisd-new in production. If everything goes well I'll switch
> the routing back and put the patch
server
putting mail in quarantine has been they don't care as it is not
relevant.
With file-based quarantine this would not work unless you are using
a shared filesystem.
Cari saluti,
Thomas Gelf
Paolo Cravero schrieb:
> Anyone? Mark?
>
> When releasing from full MySQL quarantine
tion. If everything goes well I'll switch
the routing back and put the patch on all (both) current live filter
hosts - even if doing so could waste my weekend ;-)
Cheers,
Thomas Gelf
-
This SF.net email is s
Mark Martinec schrieb:
> I can prepare a small change to amavisd which would use
> (partition_tag,mail_id) as a unique key instead of just the (mail_id).
> ...
> Do you think it would help?
That would be great! (and probably also speed up amavisd's lookups
without the need to add an additional ke
Thomas Gelf schrieb:
> Mark Martinec schrieb:
>>> The full patch is on its way...
>> So here it is, adding back the missing 'Ok, id=01234-01-2, from MTA...'
>> to SMTP responses (with some code streamlining):
> Wow, what a service! I'll test it
Thomas Gelf schrieb:
> Thomas Gelf schrieb:
>> db partitioning caused me some headache last days...
>
> It's getting even worse. The following SQL will not work...
>
> > CREATE TABLE IF NOT EXISTS `msgs` (
> > ...
> > PRIMARY KEY (`mail_id`,`
Thomas Gelf schrieb:
> db partitioning caused me some headache last days...
It's getting even worse. The following SQL will not work...
> CREATE TABLE IF NOT EXISTS `msgs` (
> ...
> PRIMARY KEY (`mail_id`,`partition_tag`),
> FOREIGN KEY (`sid`, `partition_tag`)
&g
sting partitions - so IMO the only
option left are LIST partitions.
Kind regards,
Thomas Gelf
NB: All statements above are valid for MySQL only.
-
This SF.net email is sponsored by: Microsoft
Defy all challenges. Micros
ion to my DB structure) and let you know the result.
Thanks a lot!
Thomas Gelf
-
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go
Zdravo Mark,
>> After XFORWARD has been received, the information is in:
>> $msginfo->client_addr
>> $msginfo->client_port
>
> You mean your own IP address? See 'package Amavis::In::Connection',
> such info is stored in $conn->socket_ip and $conn->socket_port.
Slowly I'm understanding small
Thomas Gelf schrieb:
> The problem is, that there is no way to keep an additional unique key
> on the tables. MySQL docs are telling me:
>
> "...every unique key on the table must use every column in the table's
> partitioning expression." (this also includes the t
Thomas Gelf schrieb:
> Can someone give me a little hint where to get this info
> from while being in process_smtp_request()?
"this info" -> [123.45.67.89]:25
-
This SF.net email is sponsored by:
Thomas Gelf schrieb:
> Hi Marc,
>
>> Looking into it... Looks like an undesired side effect
>> of a change in a client-side SMTP pipelining support.
>
> the following patch fixes the problem at least for my
> personal needs
Everything seems to be fine, I'm
MTA([123.45.67.89]:25): ...
Cheers,
Thomas Gelf
--- amavisd 2008-04-23 20:50:05.0 +0200
+++ amavisd.new 2008-05-21 12:19:18.0 +0200
@@ -15668,7 +15668,10 @@
&$check_mail($conn,$msginfo,$lmtp); # do all the contents
checking
$self->{tempdir}->
g into it... Looks like an undesired side effect
> of a change in a client-side SMTP pipelining support.
Thanks! I tried to disable pipelining announcement on the host amavis
has to deliver the mails through (forward_method) - but this does not
seem to have any effect, so no workar
ur?
Best regards,
Thomas Gelf
-
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse012070mrt/
n that I will get the same mail_id more times (once
per week) it should not do any harm if my "3rd-party-apps" are able
to handle it - is that correct? Or could amavisd-new somehow get into
trouble?
Many Thanks,
Thomas Gelf
---
4 VALUES LESS THAN (30),
PARTITION p5 VALUES LESS THAN (36),
PARTITION p6 VALUES LESS THAN (42),
PARTITION p7 VALUES LESS THAN (48),
PARTITION p8 VALUES LESS THAN (54)
);
(and that's why I was confused regarding mail_id). If there is a bett
;myhostname'), $conn->socket_ip),
+c('myproduct_name'),
$conn->socket_port eq '' ? 'unix socket' : "port
".$conn->socket_port);
$s .= "\n with $smtp_proto" if $smtp_proto=~/^(ES|S|L)MTPS?A?\z/i;
# rfc3848
$s .= "\n
so: how can I avoid duplicate mail_id's (and trouble when someone
is going to release such a mail)?
Thanks a lot for clarification!
Regards,
Thomas Gelf
-
This SF.net email is sponsored by: Microsoft
Defy all challenge
mouss schrieb:
> Thomas Gelf wrote:
>> Amavisd-new seems to use localhost_name for it's own EHLO,
>> but is using $myheloname for it's own SMTP greeting banner
>> (and 250 reply to HELO)...
>
> if it helos with myhostname and this happens to be the hostname of
Thomas Gelf schrieb:
...
Attached to this mail there is a little patch using $product_name in
the received header lines. In my old installation I've used a hard-
coded string, but $myproduct_name should do the job. I'll test it
within the next hours to let you know, if it works.
As e
$myheloname = 'localhost';
# $myheloname = '[127.0.0.1]';
$myheloname = '[' . $conn->socket_ip . ']';
IMO c('myhostname') would be the better choice, isn't it?
Cheers,
Thomas Gelf
believe to be useful.
Attached to this mail there is a little patch using $product_name in
the received header lines. In my old installation I've used a hard-
coded string, but $myproduct_name should do the job. I'll test it
within the next hours to let you know, if i
Thomas Gelf schrieb:
> Other questions:
>
> - Is there some better way to achieve a VERY scalable setup I didn't
>think about?
Btw: Anyone successfully using MySQL 5.1 with NDB Backend in production?
at before doing so?
- are you (all of you!!) going to clean up and perfectionate my code
once opensourced - so that I can lay back and let you do my job?
That's all, folks ;o)
Cheers,
Thomas Gelf
-
This SF.net email
this problem (requirement for
new packages / "outdated" distros) in your environments?!
Other questions:
- Is there some better way to achieve a VERY scalable setup I didn't
think about?
Kind regards,
Thomas Gelf
--
l never be THAT large ;-)
So, that's it (more or less). I feel sorry for my broken English, I'm
trying to give my best! You'll find technical details and questions in
the following two mails.
Regards,
Thomas Gelf
-
tick
> to manual partitioning over different tablesets.
Hmmm... and how do you accomplish this?
Cheers,
Thomas Gelf
-
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?
Mark Martinec schrieb:
> I have now a somewhat more advanced variant of my previously posted patch,
> which automatically sets partition_tag to a current ISO8601 week number
> (1..53) if $sql_partition_tag is undef, which will be a default.
> If anyone is interested please drop me a line.
/me drop
Mark Martinec schrieb:
> Your understanding is correct. There is not much need to clear a
> database of process statuses, unless there are hundreds or thousands
> of stale entries left there, but in this case you'd have a more serious
> problem to solve anyway. Apart from amavisd-nanny, an amavisd
Mark Martinec schrieb:
> Well, you already said the mail contents is full of spaces, so
> the above does not apply. It is still good to know nevertheless,
> and the $sa_spawned trick still applies, regardless of a
> reason for SpamAssassin segfaults.
I'm a little bit sceptic about FuzzyOCR & Co -
Jerry Durand schrieb:
> This is probably just a Clamd problem, but I thought I'd run it by
> here, too. I've been busy the past couple of days and haven't kept
> up on list messages.
>
> I've received several e-mails that hang the mail system (Amavisd-new
> reports TROUBLE) and the processor i
Mark Martinec schrieb:
> So if you believe the file(1) is not a cause of delays,
> please show a TIMING report from the amavisd log for
> problematic messages (available at the log level 2 or above).
Log level is 2, but as told before the last line there seems to
be "providing full original messag
Thomas Gelf schrieb:
> thanks for your answer! file is 4.17-5etch2 - which is not 4.21, but
> Debian people should have backported all relevant fixes as of my
> /usr/share/doc/file/changelog.Debian:
NB: CVE-2007-2026 is also fixed, tests from
http://www.amavis.org/security/asa-2007-3.txt
Mark Martinec schrieb:
> My guess is that you are bitten by a Linux regexp library problem,
> triggered by the file(1) utility (it doesn't affect for example FreeBSD
> with the same version of a file utility). It has been discussed before
> on this ML. Also mentioned in: http://www.amavis.org/secur
quot;clamdscan such_a_strange_mail" succeeds.
Amavisd-new version: 2.4.2 from Debian Etch (we'll move away from
Debian's outdated amavis package und use bleeding-edge amavis as soon
as we put this system into full production).
Did anyone encounter similar mails
65 matches
Mail list logo