headers
Martin Hierling | 1 Aug 2006 19:37
Picon

Re: dbmail-smtp 2.1.7 killed by sig9 with specific message

Leif,

thanks for investiation.

>  Based on the link with the logs and your message in question I tried to
> reproduce this issue under valgrind and was unable to with 2.1.7
> dbmail-smtp as well as the retrival of the message after storage matches
> the original. So I would guess that this sig9 is somewhere in your exim

when i feed the message to dbmail-smtp via stdin everything works fine.

> chain or a library diffrence I don't have, what version of glib gmime and
> glibc do you have dbmail compiled against as well as mysql versions.

glib-2.8.6
gmime-2.1.14
glibc-2.3.6
mysql-4.1.20

this are the versions i have compiled dbmail agains. 

at the moment i have about 20 mails in my queue that gets the sig9 while
exim trys to deliver them. I will look at that a little bit closer to
dig what they all have in common. 

regards Martin
--

-- 
----------------------------------------------------------------
  Backups? We doan *NEED* no steenking baX%^~,VbKx NO CARRIER
----------------------------------------------------------------
(Continue reading)

Permalink | Reply | 0 comments |
headers
Simon | 2 Aug 2006 03:11
Picon
Gravatar

Postfix > amamvis > dbmail

Hi there,

We have a setup as follows:

postfix > amamvis > spamas > clamav > dbmail

Does any one have any suggestions on how to go about finding out where
the load is centered over this.. e.g. which parts of the setup are
using the most CPU / Memory over a given period?

I know this is not a dbmail question, but i thought that someone might
have faced a similar problem before.

Thanks

Simon
Permalink | Reply | 5 comments |
headers
Matthew Smith | 2 Aug 2006 09:18
Picon

Re: Postfix > amamvis > dbmail

Quoth Simon at 08/02/06 10:41...

> postfix > amamvis > spamas > clamav > dbmail
> 
> Does any one have any suggestions on how to go about finding out where
> the load is centered over this.. e.g. which parts of the setup are
> using the most CPU / Memory over a given period?
The big question - what OS are you running this on?

M
Permalink | Reply | 3 comments |
headers
Simon | 2 Aug 2006 10:17
Picon
Gravatar

Re: Postfix > amamvis > dbmail

Opps...!

On 8/2/06, Matthew Smith <matt <at> kbc.net.au> wrote:
> Quoth Simon at 08/02/06 10:41...
>
> > postfix > amamvis > spamas > clamav > dbmail
> >
> > Does any one have any suggestions on how to go about finding out where
> > the load is centered over this.. e.g. which parts of the setup are
> > using the most CPU / Memory over a given period?
> The big question - what OS are you running this on?

Debian sarge. (Sorry - Duh!)
Permalink | Reply | 2 comments |
headers
Jake Anderson | 2 Aug 2006 11:29
Picon

Re: Postfix > amamvis > dbmail

try running "top" from the command line and see whats using the cpu?
keep an eye on it during heavy load and it should be pretty apparent

Simon wrote:
> Opps...!
>
> On 8/2/06, Matthew Smith <matt <at> kbc.net.au> wrote:
>> Quoth Simon at 08/02/06 10:41...
>>
>> > postfix > amamvis > spamas > clamav > dbmail
>> >
>> > Does any one have any suggestions on how to go about finding out where
>> > the load is centered over this.. e.g. which parts of the setup are
>> > using the most CPU / Memory over a given period?
>> The big question - what OS are you running this on?
>
> Debian sarge. (Sorry - Duh!)
> _______________________________________________
> Dbmail mailing list
> Dbmail <at> dbmail.org
> https://mailman.fastxs.nl/mailman/listinfo/dbmail
>
Permalink | Reply | 0 comments |
headers
Matthew Smith | 2 Aug 2006 11:35
Picon

Re: Postfix > amamvis > dbmail

Quoth Simon at 08/02/06 17:47...
>> > postfix > amamvis > spamas > clamav > dbmail
>> >
>> > Does any one have any suggestions on how to go about finding out where
>> > the load is centered over this.. e.g. which parts of the setup are
>> > using the most CPU / Memory over a given period?
>> The big question - what OS are you running this on?
> 
> Debian sarge. (Sorry - Duh!)

There's a lot of stuff available for performance monitoring on Un*x
systems; racking my brains to remember what it is I used to use on
AIX...  Linux-wise, it's a case of "Google Is Your Friend".

To find what you are looking for, I would suggest that you Google with
keywords: "debian performance monitor cpu memory"; you could substitute
"linux" for "debian" to look for tools not available as Debian packages.

Example (may obviate the need for a search!) - atop:
http://packages.debian.org/unstable/admin/atop

Hope this helps.

Cheers

M
Permalink | Reply | 0 comments |
headers
Martin Hierling | 2 Aug 2006 17:52
Picon

Mailquota like Unix Filesystem on the Roadmap

Hi,

have not found some thread concerning this so here is my question.

is it planed to implement quotas like it is on unix filesystems:

volume soft and hard limit  (25M, 30M)
messagecount soft and hard limit (1000 Mails, 1100 Mails)

So i could send a 4xx error if soft limit is reached and 5xx if hard
limut is reached. 

It that feature on the roadmap?

regards Martin
--

-- 
----------------------------------------------------------------
  DOS Tip #63: add BUGS=OFF to CONFIG.SYS
----------------------------------------------------------------
Permalink | Reply | 5 comments |
headers
Jesse Norell | 2 Aug 2006 19:16
Favicon
Gravatar

Re: Mailquota like Unix Filesystem on the Roadmap

On Wed, 2006-08-02 at 17:52 +0200, Martin Hierling wrote:
> Hi,
> 
> have not found some thread concerning this so here is my question.
> 
> is it planed to implement quotas like it is on unix filesystems:
> 
> volume soft and hard limit  (25M, 30M)
> messagecount soft and hard limit (1000 Mails, 1100 Mails)

  I don't think so.  It may have been mentioned but I don't remember
anyone ever having put forth any convincing arguments for doing that.  I
can't think of any myself for email right offhand.  Generally any idea
with well thought reasoning, benefits and a little impetus behind it
will be accepted and eventually implemented though, so feel free to
champion the cause if there's a good reason to.  :)

> So i could send a 4xx error if soft limit is reached and 5xx if hard
> limut is reached. 

  That's not how filesystem soft/hard limits work.  To a mail system the
fact that a user is above their soft limit (while within their grace
period) is completely unnoticed; once a user hits their hard limit or
their grace period expires the same thing happens - writes to the
filesystem fail and an error is returned.  The condition is always
temporary (ie. free some disk space and it would then deliver), though
it's probably often/always given a permanent failure code.

  What you propose may cause more harm than good.  If a message was
getting a 4xx error because the user was over their soft quota, the
(Continue reading)

Permalink | Reply | 0 comments |
headers
Simon Lange | 2 Aug 2006 20:06

AW: Mailquota like Unix Filesystem on the Roadmap

I have written a small perl programm which - regulated via cron - does check
the quotas of users. At some point (you would call it soft limit) it sends a
warning message to the primary alias of the desired account until the owner
does remove mails. If the owner does not react and the mailbox runs
completely full (100%) the script will empty the mailbox without further
notice. That way new mails can arrive. Mailboxes are not meant as a static
mailarchive. If you don’t like it its pretty simple to remove this
hardline-process...

If interested, I would post it on my website so u can fetch it and use it.

Another way is to use squirrelmail and using their imap-quota system. Works
nice including warnings... :)

Simon

-----Ursprüngliche Nachricht-----
Von: dbmail-bounces <at> dbmail.org [mailto:dbmail-bounces <at> dbmail.org] Im Auftrag
von Martin Hierling
Gesendet: Mittwoch, 2. August 2006 17:53
An: DBMail mailinglist
Betreff: [Dbmail] Mailquota like Unix Filesystem on the Roadmap

Hi,

have not found some thread concerning this so here is my question.

is it planed to implement quotas like it is on unix filesystems:

volume soft and hard limit  (25M, 30M)
(Continue reading)

Permalink | Reply | 3 comments |
headers
Jorge Bastos | 2 Aug 2006 21:03
Picon

Re: Mailquota like Unix Filesystem on the Roadmap

That script is interesting.
There only one thing i do not agree, is the full delete of the user mailbox 
when 100% is reached.

----- Original Message ----- 
From: "Simon Lange" <sl <at> polynaturedesign.com>
To: "'DBMail mailinglist'" <dbmail <at> dbmail.org>
Sent: Wednesday, August 02, 2006 7:06 PM
Subject: AW: [Dbmail] Mailquota like Unix Filesystem on the Roadmap

>I have written a small perl programm which - regulated via cron - does 
>check
> the quotas of users. At some point (you would call it soft limit) it sends 
> a
> warning message to the primary alias of the desired account until the 
> owner
> does remove mails. If the owner does not react and the mailbox runs
> completely full (100%) the script will empty the mailbox without further
> notice. That way new mails can arrive. Mailboxes are not meant as a static
> mailarchive. If you don't like it its pretty simple to remove this
> hardline-process...
>
> If interested, I would post it on my website so u can fetch it and use it.
>
> Another way is to use squirrelmail and using their imap-quota system. 
> Works
> nice including warnings... :)
>
> Simon
>
(Continue reading)

Permalink | Reply | 0 comments |

Gmane