devil-linux-discuss — General Mailing List for all discussions

Re: [Devil-Linux-discuss] Sagator as MTA?

[Kari M. <kar...@tr...>]

Hi,

Our level of email backscatter has increased 600% in last 6 weeks from =

avg of 6500/day to 35,000/day and is bogging down our mail server.

My first thought was to use my handy dandy DL to solve this. So I'm =

trying to configure a DL box to act as an email filter in front of our =

mail server, so once configured and tested I'll update our MX records ( =

I have 50 domains and 1000 users) to the new DL box.

I've got clamav & spamd configured, but trying to figure out how to get =

sagator to accept the inbound mail on port 25 before relaying the good =

email on to our QMail server.


	Don't do it like that. Sagator was designed to be a Postfix content =
filter. You are propably wasting your time. Instead, have Postfix listening =
port 25, and then forward that mail to Sagator. You get it back to Postfix. =
Then Postfix will use the transport table to forward the good mail to which =
ever IP you want. You can decide domain by domain where Postfix will send =
the good mail.

	The default Sagator conf is rather good. If you want to quarantine mail =
and have lots of spam, adding %d to the quarantine dir to make daily sam =
dir makes sense. Virus feed is so low, that one dir is sufficient.

	You also wnt to check rlimit on sagator.conf. It limit max msg size.

	Sagator source tar ball has a script which adds proper settings to Postfix =
master.cf, if you don't want to do it manually.

	One important note: I once had a spam filter receiving all the feed from =
the net. It failed. 100000+ mails in queues going for filtering. Uh.

	In my experience the Postfix UCE features are excellent. If you add to =
that a list of valid mail addresses, Postfix will not accept mail for =
invalid addresses. These two together reduce the amount of mail to really =
filter with Sagator down to about 1% of the original.

	1 GHz P3 box with 512 MB should alone be ok for 1000 users. I'd have 2-3 =
in parallel for redundancy.


in sagator.conf I tried :

  smtpd(SCANNERS,'64.34.15.2',25),

Where the IP is the IP of the DL box, from an other machine I telnet to =

port 25 and it accepts, but when I do "helo machinename" <cr> I get no =

response and the DL box keeps the connection open until I stop =
sagator.

So I've obviously got something wrong, or do I have to configure postfix =

to accept and hand off to sagator?

Any input would be greatly appreciated.

Mike

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, =
security?
Get stuff done quickly with pre-integrated technology to make your job =
easier.
Download IBM WebSphere Application Server v.1.0.1 based on Apache =
Geronimo
http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&dat=3D12=
1642
_______________________________________________
Devil-linux-discuss mailing list
Dev...@li...
https://lists.sourceforge.net/lists/listinfo/devil-linux-discuss

[Devil-Linux-discuss] Sagator as MTA?

[Mike <Mi...@4a...>]

Hi,

Our level of email backscatter has increased 600% in last 6 weeks from 
avg of 6500/day to 35,000/day and is bogging down our mail server.

My first thought was to use my handy dandy DL to solve this. So I'm 
trying to configure a DL box to act as an email filter in front of our 
mail server, so once configured and tested I'll update our MX records ( 
I have 50 domains and 1000 users) to the new DL box.

I've got clamav & spamd configured, but trying to figure out how to get 
sagator to accept the inbound mail on port 25 before relaying the good 
email on to our QMail server.

in sagator.conf I tried :

  smtpd(SCANNERS,'64.34.15.2',25),

Where the IP is the IP of the DL box, from an other machine I telnet to 
port 25 and it accepts, but when I do "helo machinename" <cr> I get no 
response and the DL box keeps the connection open until I stop sagator.

So I've obviously got something wrong, or do I have to configure postfix 
to accept and hand off to sagator?

Any input would be greatly appreciated.

Mike

Re: [Devil-Linux-discuss] Building the stable version

[Michele P. <10...@ti...>]

At 08.28 08/02/2007 -0600, Heiko Zuerker wrote:
>This is the current 'stable' release, it's just a little more current then
>the official one.

Ok, I hadn't quite got it!

So, can I safely compile some addon binaries on this build system and be 
confident that I can safely add them to the 1.2.12 .iso file downloaded 
from the official mirror?

Widely, my question could expand to "What are the exact prerequisites of an 
"external" binary executable to be able to run on DL?"
For example, can I run a binary from , say, a Debian Sarge install or a 
Centos one?

>Unfortunately we can't separate out the various releases, it would eat up
>too much disk space.

I can understand , that's ok with me.

Thanks again :-) (for everything, from the wonderful distro to the support 
I'm getting here)


Live long and prosper

Re: [Devil-Linux-discuss] Dual VPN Servers

[Frank W. <Fra...@ct...>]

Hi again,

and thanks for the reply...

The Realm stuff is OK, I forgot to show you this one. (I had toyed with a N=
ULL=20
realm too, but realm detection works... when I use fw...@is..., freeradi=
us=20
tries to contact the other server (10.0.0.3)

testradius:~# radtest fw...@ct... blabla 127.0.0.1 1812 testing1=
23
Sending Access-Request of id 75 to 127.0.0.1:1812
        User-Name =3D "fw...@ct..."
        User-Password =3D "blabla"
        NAS-IP-Address =3D testradius
        NAS-Port =3D 1812
rad_recv: Access-Accept packet from host 127.0.0.1:1812, id=3D75, length=3D=
70
        Framed-Protocol =3D PPP
        Service-Type =3D Framed-User
        Framed-IP-Address =3D 172.16.0.111
        Class =3D 0x357004300000013700010a00000201c74b86aecf547b00000000000=
00005

my proxy.conf looks like this:

=2D--8<---
proxy server {
        synchronous =3D no
        retry_delay =3D 5
        retry_count =3D 3
        dead_time =3D 120
        default_fallback =3D yes
        post_proxy_authorize =3D yes
}

realm NULL {
        type            =3D radius
        authhost        =3D 10.0.0.3:1812
        accthost        =3D 10.0.0.3:1813
        secret          =3D testing123
}

realm cte-admin.cte.lu  {
        type            =3D radius
        authhost        =3D 10.0.0.2:1812
        accthost        =3D 10.0.0.2:1813
        secret          =3D testing123
}

realm isp2.com {
        type        =3D radius
        authhost    =3D 10.0.0.3:1812
        accthost    =3D 10.0.0.3:1813
        secret      =3D TheirKey
}

realm LOCAL {
        type            =3D radius
        authhost        =3D LOCAL
        accthost        =3D LOCAL
}

=2D--8<---
This is the suffix part form radiusd.conf:
=2D--8<---
realm suffix {
                format =3D suffix
                delimiter =3D "@"
                ignore_default =3D no
                ignore_null =3D no=20
       }
=2D--8<---



On Thursday 08 February 2007 15:46:57 Serge Leschinsky wrote:
> Hi Frank,
>
> Probably the root cause is in the realm syntax.
>
> Allowed realm formats are:
>
>  username@realm
>  realm/username
>  username%realm
>  realm\username
>
> The realm syntax  is defined via the realm module config in radiusd.conf
>
> When you use radtest, your request is processed by config section with
> realm NULL ( try the execute the following:
> radtest fweis@realm whatever 127.0.0.1 1812  testing123 )
>
>
> So, I believe the problem is in the proxy.conf. May I ask you to show me
> the file? And, may be realm module from radiusd.conf...
>
> PS. Try to run freeradius as "radiusd -x".
>
>

=2D-=20
_______________________________________________
 Centre de Technologie de l'Education
 29 avenue John F. Kennedy
 L-1855 Luxembourg-Kirchberg
 email: Fra...@ct...
 t=E9l.: +352 478-5973
 fax: +352 333797
 _______________________________________________

Re: [Devil-Linux-discuss] [BULK] Re: Dual VPN Servers

[Serge L. <fi...@in...>]

Hi Frank,

Probably the root cause is in the realm syntax.

Allowed realm formats are:

 username@realm
 realm/username
 username%realm
 realm\username

The realm syntax  is defined via the realm module config in radiusd.conf

When you use radtest, your request is processed by config section with realm
NULL ( try the execute the following:
radtest fweis@realm whatever 127.0.0.1 1812  testing123 )


So, I believe the problem is in the proxy.conf. May I ask you to show me the
file? And, may be realm module from radiusd.conf...

PS. Try to run freeradius as "radiusd -x".


--
  Serge


Frank Weis wrote:
> Hi Serge,
> 
> thanks, I begin to understand :-)
> 
> 
> I have successfully tried the setup where pptp accesses MS IAS directly.
> 
> I have then modified the setup such that pptp->local freeradius->proxy to MS 
> IAS based on the realm.
> 
> This is also working as expected, but for some reason the authentication on 
> the IAS side fails. I believe I am very close to the solution, but not really 
> there yet..... I think it could have to be related to the way the password is 
> handled somewhere along the way... Please take a look at the logs below...
> 
> Thanks in advance for any hints,
> 
> Frank.
> 
> The event on the WIndows server says:
> ----8<----
>  
>  ....
>  Authentication-Type = MS-CHAPv2
>  EAP-Type = <undetermined> 
>  Reason-Code = 16
>  Reason = Authentication was not successful because an unknown user name or 
> incorrect password was used. 
> ----8<---
> 
> I am absolutely positive that the credentials are OK because when I change 
> back pptp to access the IAS directly it works.
> 
> Also, when I use radtest on the DL box, it works like a charm:
> ----8<---
> testradius:~# radtest fweis whatever 127.0.0.1 1812 testing123
> Sending Access-Request of id 87 to 127.0.0.1:1812
>         User-Name = "fweis"
>         User-Password = "whatever"
>         NAS-IP-Address = testradius
>         NAS-Port = 1812
> rad_recv: Access-Accept packet from host 127.0.0.1:1812, id=87, length=94
>         Framed-Protocol = PPP
>         Service-Type = Framed-User
>         Framed-IP-Address = 172.16.0.111
>         Class = 0x366504570000013700010a00000201c74acc5a7feb5b0000000000000014
>         MS-MPPE-Encryption-Policy = 0x00000002
>         MS-MPPE-Encryption-Types = 0x0000000e
> ---8<---
> The radius log says this
> Sending Access-Request of id 2 to 10.0.0.2:1812
>         User-Name = "fweis"
>         User-Password = "whatever"
>         NAS-IP-Address = 255.255.255.255
>         NAS-Port = 1812
>         Proxy-State = 0x313134
> --- Walking the entire request list ---
> Waking up in 6 seconds...
> rad_recv: Access-Accept packet from host 10.0.0.2:1812, id=2, length=99
>         Proxy-State = 0x313134
>         Framed-Protocol = PPP
>         Service-Type = Framed-User
>         Framed-IP-Address = 172.16.0.111
>         Class = 0x366704590000013700010a00000201c74acc5a7feb5b0000000000000016
>         MS-MPPE-Encryption-Policy = 0x00000002
>         MS-MPPE-Encryption-Types = 0x0000000e
> ---8<---
> 
> 
> When I use the VPN client, the connection fails:
> 
> ---8<---
> Sending Access-Request of id 3 to 10.0.0.2:1812
>         Service-Type = Framed-User
>         Framed-Protocol = PPP
>         User-Name = "fweis"
>         MS-CHAP-Challenge = 0xb82fba63c4b3bd8a737aebc4b64c269d
>         MS-CHAP2-Response = 0xc500blablabla
>         Calling-Station-Id = "10.0.108.1"
>         NAS-Identifier = "XX_NASPPP"
>         NAS-Port = 0
>         NAS-IP-Address = 127.0.0.1
>         Proxy-State = 0x313433
> --- Walking the entire request list ---
> Waking up in 6 seconds...
> rad_recv: Access-Reject packet from host 10.0.0.2:1812, id=3, length=47
>         Proxy-State = 0x313433
>         MS-CHAP-Error = "\000E=691 R=0 V=3"
> 
> ---8<-----
> 
> 
> 
> On Tuesday 06 February 2007 13:54:52 Serge Leschinsky wrote:
>> Hi Frank,
>>
>> Frank Weis wrote:
>>> Hi,
>>>
>>> I have just found out that 'Portslave' is not checked in the default
>>> config. That package contains radiusclient.
>> Yes. But it's obsolete. It would be better if you will not use portslave at
>> all.
>>
>>> We are effectively speaking about different things: I want to run
>>> pptpd and radiusd on the DL box so I need the radiusclient stuff so pptpd
>>> can talk to radius.
>> Yes I see. You can use radius _plugin_ for pppd (it isn't the same to
>> suchradiusclient) . I sent configs exactly for such case.
>>
>>> And I need the rlm_ modules (like rlm_ldap)  so radius can
>>> talk do my two AD servers.
>> I see. But DL radius server (freeradius) can talk with IAS (MS radius
>> server) without ldap module, I mean proxy mode of freeradius server...
>>
>> Sorry in advance,  if I misunderstood something :-)
>>
>> --
>>  Serge
>>
>> -------------------------------------------------------------------------
>> Using Tomcat but need to do more? Need to support web services, security?
>> Get stuff done quickly with pre-integrated technology to make your job
>> easier. Download IBM WebSphere Application Server v.1.0.1 based on Apache
>> Geronimo
>> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
>> _______________________________________________
>> Devil-linux-discuss mailing list
>> Dev...@li...
>> https://lists.sourceforge.net/lists/listinfo/devil-linux-discuss
> 
> 
> 

Re: [Devil-Linux-discuss] Building the stable version

[Heiko Z. <he...@zu...>]

On Thu, February 8, 2007 04:25, 10...@ti... wrote:
> Hi all.
> I've just built DL from the lfssystem following the
> instructions here http://www.devil-linux.org/documentation/1.2.
> x/ch03s01.html
>
> The build went successfully but the version I got is:
> Devil-Linux 1.2.13-2007-02-08-i586-SMP
>
>
> But the current stable is
> 1.2.12, so how do I get it to build the current version?

This is the current 'stable' release, it's just a little more current then
the official one.
Unfortunately we can't separate out the various releases, it would eat up
too much disk space.
The current development release is 1.3, so you're OK.

> During the
> build process i did as it says:
>
> "To get i.e. the stable 1.2.x release,
> use the following command:
>
> cvs -z3 -d:pserver:anonymous@devil-linux.
> cvs.sourceforge.net:/cvsroot/devil-linux co -r rel-1-2-patches build"
>
>
> Also, is there a way to build earlier versions (1.2.9 for example)?

No not really. The problem is that our FTP server would grow to a couple
hundred gigabytes. Additionally we wouldn't be able to find mirrors....

> Thank you all!

You're welcome

-- 

Regards
  Heiko Zuerker
  http://www.devil-linux.org

Re: [Devil-Linux-discuss] [BULK] Re: Dual VPN Servers

[Frank W. <Fra...@ct...>]

Hi Serge,

thanks, I begin to understand :-)


I have successfully tried the setup where pptp accesses MS IAS directly.

I have then modified the setup such that pptp->local freeradius->proxy to M=
S=20
IAS based on the realm.

This is also working as expected, but for some reason the authentication on=
=20
the IAS side fails. I believe I am very close to the solution, but not real=
ly=20
there yet..... I think it could have to be related to the way the password =
is=20
handled somewhere along the way... Please take a look at the logs below...

Thanks in advance for any hints,

=46rank.

The event on the WIndows server says:
=2D---8<----
=20
 ....
 Authentication-Type =3D MS-CHAPv2
 EAP-Type =3D <undetermined>=20
 Reason-Code =3D 16
 Reason =3D Authentication was not successful because an unknown user name =
or=20
incorrect password was used.=20
=2D---8<---

I am absolutely positive that the credentials are OK because when I change=
=20
back pptp to access the IAS directly it works.

Also, when I use radtest on the DL box, it works like a charm:
=2D---8<---
testradius:~# radtest fweis whatever 127.0.0.1 1812 testing123
Sending Access-Request of id 87 to 127.0.0.1:1812
        User-Name =3D "fweis"
        User-Password =3D "whatever"
        NAS-IP-Address =3D testradius
        NAS-Port =3D 1812
rad_recv: Access-Accept packet from host 127.0.0.1:1812, id=3D87, length=3D=
94
        Framed-Protocol =3D PPP
        Service-Type =3D Framed-User
        Framed-IP-Address =3D 172.16.0.111
        Class =3D 0x366504570000013700010a00000201c74acc5a7feb5b00000000000=
00014
        MS-MPPE-Encryption-Policy =3D 0x00000002
        MS-MPPE-Encryption-Types =3D 0x0000000e
=2D--8<---
The radius log says this
Sending Access-Request of id 2 to 10.0.0.2:1812
        User-Name =3D "fweis"
        User-Password =3D "whatever"
        NAS-IP-Address =3D 255.255.255.255
        NAS-Port =3D 1812
        Proxy-State =3D 0x313134
=2D-- Walking the entire request list ---
Waking up in 6 seconds...
rad_recv: Access-Accept packet from host 10.0.0.2:1812, id=3D2, length=3D99
        Proxy-State =3D 0x313134
        Framed-Protocol =3D PPP
        Service-Type =3D Framed-User
        Framed-IP-Address =3D 172.16.0.111
        Class =3D 0x366704590000013700010a00000201c74acc5a7feb5b00000000000=
00016
        MS-MPPE-Encryption-Policy =3D 0x00000002
        MS-MPPE-Encryption-Types =3D 0x0000000e
=2D--8<---


When I use the VPN client, the connection fails:

=2D--8<---
Sending Access-Request of id 3 to 10.0.0.2:1812
        Service-Type =3D Framed-User
        Framed-Protocol =3D PPP
        User-Name =3D "fweis"
        MS-CHAP-Challenge =3D 0xb82fba63c4b3bd8a737aebc4b64c269d
        MS-CHAP2-Response =3D 0xc500blablabla
        Calling-Station-Id =3D "10.0.108.1"
        NAS-Identifier =3D "XX_NASPPP"
        NAS-Port =3D 0
        NAS-IP-Address =3D 127.0.0.1
        Proxy-State =3D 0x313433
=2D-- Walking the entire request list ---
Waking up in 6 seconds...
rad_recv: Access-Reject packet from host 10.0.0.2:1812, id=3D3, length=3D47
        Proxy-State =3D 0x313433
        MS-CHAP-Error =3D "\000E=3D691 R=3D0 V=3D3"

=2D--8<-----



On Tuesday 06 February 2007 13:54:52 Serge Leschinsky wrote:
> Hi Frank,
>
> Frank Weis wrote:
> > Hi,
> >
> > I have just found out that 'Portslave' is not checked in the default
> > config. That package contains radiusclient.
>
> Yes. But it's obsolete. It would be better if you will not use portslave =
at
> all.
>
> > We are effectively speaking about different things: I want to run
> > pptpd and radiusd on the DL box so I need the radiusclient stuff so ppt=
pd
> > can talk to radius.
>
> Yes I see. You can use radius _plugin_ for pppd (it isn't the same to
> suchradiusclient) . I sent configs exactly for such case.
>
> > And I need the rlm_ modules (like rlm_ldap)  so radius can
> > talk do my two AD servers.
>
> I see. But DL radius server (freeradius) can talk with IAS (MS radius
> server) without ldap module, I mean proxy mode of freeradius server...
>
> Sorry in advance,  if I misunderstood something :-)
>
> --
>  Serge
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job
> easier. Download IBM WebSphere Application Server v.1.0.1 based on Apache
> Geronimo
> http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&dat=
=3D121642
> _______________________________________________
> Devil-linux-discuss mailing list
> Dev...@li...
> https://lists.sourceforge.net/lists/listinfo/devil-linux-discuss



=2D-=20
_______________________________________________
 Centre de Technologie de l'Education
 29 avenue John F. Kennedy
 L-1855 Luxembourg-Kirchberg
 email: Fra...@ct...
 t=E9l.: +352 478-5973
 fax: +352 333797
 _______________________________________________

Re: [Devil-Linux-discuss] Building the stable version

[Oliver N. <dig...@gm...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

10...@ti... wrote:
> Hi all.
> I've just built DL from the lfssystem following the 
> instructions here
> http://www.devil-linux.org/documentation/1.2.
> x/ch03s01.html
> 
> The build went successfully but the version I got is: 
> Devil-Linux 1.2.13-2007-02-08-i586-SMP
> 
> But the current stable is 
> 1.2.12, so how do I get it to build the current version?
> 
> During the 
> build process i did as it says:
> 
> "To get i.e. the stable 1.2.x release, 
> use the following command:
> 
> cvs -z3 -d:pserver:anonymous@devil-linux.
> cvs.sourceforge.net:/cvsroot/devil-linux co -r rel-1-2-patches build"
> 
> Also, is there a way to build earlier versions (1.2.9 for example)?
> 
> Thank you all!
> 
Use the following commands:

> cvs -d:pserver:ano...@de...:/cvsroot/devil-linux login
> cvs -z3 -d:pserver:ano...@de...:/cvsroot/devil-linux co -r rel-1-2-patches build


They work nice for me!

Regards,

Olli



-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)
Comment: Using GnuPG with CentOS - http://enigmail.mozdev.org

iD8DBQFFywvn724ZL5LNhNcRAoEBAJ93wxW7Ryue4OycCVILvYCqa0MMCQCeIbcg
Z/au2qvModSXwsRNjpHQcrA=
=lDPB
-----END PGP SIGNATURE-----

[Devil-Linux-discuss] Building the stable version

[<10...@ti...>]

Hi all.
I've just built DL from the lfssystem following the 
instructions here
http://www.devil-linux.org/documentation/1.2.
x/ch03s01.html

The build went successfully but the version I got is: 
Devil-Linux 1.2.13-2007-02-08-i586-SMP

But the current stable is 
1.2.12, so how do I get it to build the current version?

During the 
build process i did as it says:

"To get i.e. the stable 1.2.x release, 
use the following command:

cvs -z3 -d:pserver:anonymous@devil-linux.
cvs.sourceforge.net:/cvsroot/devil-linux co -r rel-1-2-patches build"

Also, is there a way to build earlier versions (1.2.9 for example)?

Thank you all!

Re: [Devil-Linux-discuss] Postfix and Dovecot SASL startup catch-22

[Matthew H. <mat...@va...>]

I've ran out of RAM on a DL box, then daemons start to die at a alarming
rate. We looks awfully like a crash.

-----Original Message-----
From: dev...@li...
[mailto:dev...@li...] On Behalf Of
Bruce Smith
Sent: 07 February 2007 18:38
To: di...@li...; dev...@li...
Subject: Re: [Devil-Linux-discuss] Postfix and Dovecot SASL startup
catch-22

> > And what happens if your DL box loses power or crashes?  You lose
all
> > the queued email in the ram disk?  Yikes!
>=20
> What???? DL? Crash?  Surely not!

I don't remember any DL crashes, that were software related.

But I've had power outages long enough to deplete my UPS.
I've had CPU fans quit.  I've had power supplies die.
I've had breakers/fuses blow.  I've had mobo capacitors leak.
I've had video cards and networks cards die and bring down a box.
I've seen faulty memory do very strange things to a server.
I once had a steam pipe start spraying water into a server!  :-)

Any of those things (and other hardware problems) will take a toll=20
on your uptime....  :-(

 - BS



------------------------------------------------------------------------
-
Using Tomcat but need to do more? Need to support web services,
security?
Get stuff done quickly with pre-integrated technology to make your job
easier.
Download IBM WebSphere Application Server v.1.0.1 based on Apache
Geronimo
http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&dat=3D=
121642
_______________________________________________
Devil-linux-discuss mailing list
Dev...@li...
https://lists.sourceforge.net/lists/listinfo/devil-linux-discuss

Re: [Devil-Linux-discuss] Problem installing to USB

[Fred F. <ffr...@fr...>]

Dick, thanks for your hint. syslinux was the way to go and setting up a
serial terminal with it was a breeze.

Again thanks.

Fred

Dick Middleton wrote:
> Fred Frigerio wrote:
>> I have tried to get the install to USB to work with no luck.
> 
> FWIW I've never managed to get the grub or lilo installs working.  I 
> always use the syslinux version.
> 
> Dick
> 
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier.
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

Re: [Devil-Linux-discuss] Postfix and Dovecot SASL startup catch-22

[Bruce S. <bw...@ar...>]

> > And what happens if your DL box loses power or crashes?  You lose all
> > the queued email in the ram disk?  Yikes!
> 
> What???? DL? Crash?  Surely not!

I don't remember any DL crashes, that were software related.

But I've had power outages long enough to deplete my UPS.
I've had CPU fans quit.  I've had power supplies die.
I've had breakers/fuses blow.  I've had mobo capacitors leak.
I've had video cards and networks cards die and bring down a box.
I've seen faulty memory do very strange things to a server.
I once had a steam pipe start spraying water into a server!  :-)

Any of those things (and other hardware problems) will take a toll 
on your uptime....  :-(

 - BS

Re: [Devil-Linux-discuss] Postfix and Dovecot SASL startup catch-22

[Dick M. <di...@li...>]

Bruce Smith wrote:

> Is the incoming email stored on the CF?  If not, I hope you have a small
> message size limit and a GOOD spam blocker (or a LOT of memory)!  :-)

Good spam blocker.  Postfix UCE rejection is excellent, and that's without any 
message scanning.

> And what happens if your DL box loses power or crashes?  You lose all
> the queued email in the ram disk?  Yikes!

What???? DL? Crash?  Surely not!

I suppose it's theoretically possible but I think the MTBF is longer than the 
time between updates ;-)

Dick

Re: [Devil-Linux-discuss] Postfix and Dovecot SASL startup catch-22

[Bruce S. <bw...@ar...>]

> > A send-only setup is the only legit reason that I can think of to run
> > Postfix on a diskless system.
> 
> Hmm - I don't know.  Not much mail gets stored in spool files; only those that 
> can't be delivered.  Anyway a send-only postfix still has a queue.

Right, and I control what it's sending.  In my case, they are very small
text-only notification emails from Apache to myself.  Plus they are
infrequent, so the memory outbound queue works fine for my application.
And no big deal if it loses power and the queue is lost.

> My system is not entirely diskless - the mailboxes and mail home dirs are on a 
> hard drive.  Thats if a CF disk counts ;-)
> 
> Putting the spool files on the disk is not such a big deal, I've just never done 
> it.  Doing so would mean my original question is irrelevant.

Is the incoming email stored on the CF?  If not, I hope you have a small
message size limit and a GOOD spam blocker (or a LOT of memory)!  :-)

And what happens if your DL box loses power or crashes?  You lose all
the queued email in the ram disk?  Yikes!

 - BS

Re: [Devil-Linux-discuss] Postfix and Dovecot SASL startup catch-22

[Dick M. <di...@li...>]

Bruce Smith wrote:

> I don't know much about Postfix (I'm a sendmail guy), but I do have
> Postfix configured on a couple of DL boxes without a hard drive.
> HOWEVER, it's only purpose is to SEND email, it cannot receive any email
> (port 25 is blocked), so I don't have the worry about incoming email
> filling up my memory.

> A send-only setup is the only legit reason that I can think of to run
> Postfix on a diskless system.

Hmm - I don't know.  Not much mail gets stored in spool files; only those that 
can't be delivered.  Anyway a send-only postfix still has a queue.

My system is not entirely diskless - the mailboxes and mail home dirs are on a 
hard drive.  Thats if a CF disk counts ;-)

Putting the spool files on the disk is not such a big deal, I've just never done 
it.  Doing so would mean my original question is irrelevant.

Dick

[Devil-Linux-discuss] Directory permissions in etc.tar.bz2

[Dick M. <di...@li...>]

Another thing that's a bit odd:  I've got a subdirectory of /etc/postfix - the 
permissions of this directory are not correct after reboot.  Specifically, the 
execute bits are unset so the directory can't be read.  The permissions are 
correct in the etc.tar.bz2 archive so something must change the permissions 
after reboot.  What does this and can I tell it about my subdir?

Thanks

Dick

Re: [Devil-Linux-discuss] Postfix and Dovecot SASL startup catch-22

[Bruce S. <br...@ar...>]

> > I've been meaning to point this out for a while; I don't reboot very
> > often so it's not a big priority.
> >
> > To use dovecot SASL with postfix you need to start dovecot first.
> > However
> > dovecot needs to create a socket in postfix jail
> > (/var/spool/postfix/private)
> > which, of course, doesn't exist until postfix is started.
> 
> Does that mean your /var/spool/postfix is not on a harddisk?
> I would highly recommend running a mail server only with the spool
> directory mounted to a harddisk.

I don't know much about Postfix (I'm a sendmail guy), but I do have
Postfix configured on a couple of DL boxes without a hard drive.
HOWEVER, it's only purpose is to SEND email, it cannot receive any email
(port 25 is blocked), so I don't have the worry about incoming email
filling up my memory.

> > What I notice is that postfix doesn't complain if
> > /var/spool/postfix/private
> > exists when it is started.
> >
> > I'm thinking that the dovecot init.d script could mkdir -p this directory
> > before starting (perhaps commented out since most users won't need it).
> 
> Anybody have any thoughts on this?

A send-only setup is the only legit reason that I can think of to run
Postfix on a diskless system.  And with this kind of setup, dovecot
doesn't serve any purpose that I can think of.  Does a send-only setup
(without dovecot) need a /var/spool/postfix/private directory?

If so, then I think a mkdir -p would be a good idea.  Otherwise, I don't
know.  I guess it wouldn't hurt to do this in the dovecot init script:

[ ! -d /var/spool/postfix/private ] && mkdir -p /var/spool/postfix/private

 - BS

Re: [Devil-Linux-discuss] Postfix and Dovecot SASL startup catch-22

[Dick M. <di...@li...>]

Heiko Zuerker wrote:
> On Tue, February 6, 2007 15:20, Dick Middleton wrote:
>> I've been meaning to point this out for a while; I don't reboot very
>> often so it's not a big priority.
>>
>> To use dovecot SASL with postfix you need to start dovecot first.
>> However
>> dovecot needs to create a socket in postfix jail
>> (/var/spool/postfix/private)
>> which, of course, doesn't exist until postfix is started.
> 
> Does that mean your /var/spool/postfix is not on a harddisk?

Yes, indeed.  Never even occurred to me - stuck in the mindset of no hard disk ;-)

It would help of course, to keep the spool file over reboots.  As it also avoids 
this problem.

Maybe that's what I'll do.

> I would highly recommend running a mail server only with the spool
> directory mounted to a harddisk.

Point taken.

Dick

Re: [Devil-Linux-discuss] Postfix and Dovecot SASL startup catch-22

[Heiko Z. <he...@zu...>]

On Tue, February 6, 2007 15:20, Dick Middleton wrote:
> I've been meaning to point this out for a while; I don't reboot very
> often so it's not a big priority.
>
> To use dovecot SASL with postfix you need to start dovecot first.
> However
> dovecot needs to create a socket in postfix jail
> (/var/spool/postfix/private)
> which, of course, doesn't exist until postfix is started.

Does that mean your /var/spool/postfix is not on a harddisk?
I would highly recommend running a mail server only with the spool
directory mounted to a harddisk.

> What I notice is that postfix doesn't complain if
> /var/spool/postfix/private
> exists when it is started.
>
> I'm thinking that the dovecot init.d script could mkdir -p this directory
> before starting (perhaps commented out since most users won't need it).

Anybody have any thoughts on this?

-- 

Regards
  Heiko Zuerker
  http://www.devil-linux.org

[Devil-Linux-discuss] Postfix and Dovecot SASL startup catch-22

[Dick M. <di...@li...>]

I've been meaning to point this out for a while; I don't reboot very often so 
it's not a big priority.

To use dovecot SASL with postfix you need to start dovecot first.  However 
dovecot needs to create a socket in postfix jail (/var/spool/postfix/private) 
which, of course, doesn't exist until postfix is started.

What I notice is that postfix doesn't complain if /var/spool/postfix/private 
exists when it is started.

I'm thinking that the dovecot init.d script could mkdir -p this directory before 
starting (perhaps commented out since most users won't need it).

Dick

Re: [Devil-Linux-discuss] Problem installing to USB

[Heiko Z. <he...@zu...>]

On Tue, February 6, 2007 09:01, Fred Frigerio wrote:
>> -----Original Message-----
>> From: dev...@li...
>> [mailto:dev...@li...] On
>> Behalf Of Heiko Zuerker
>> Sent: Tuesday, February 06, 2007 9:58 AM
>> To: dev...@li...
>> Subject: Re: [Devil-Linux-discuss] Problem installing to USB
>>
>>
>>
>> On Tue, February 6, 2007 07:00, Fred Frigerio wrote:
>>
>>>>
>>
>>>> At that point suspecting a GRUB problem I wiped the stick
>>>>
>> and tried to
>>>> rebuild it by hand with the GRUB from my computer (version
>> 0.97). I did
>>
>>>> what I am used to which is a boot partition (sdh1) a root partition
>>>>  (sdh3) and a config partition (sdh2) all ext2. I installed
>>>>
>> grub to sdh1
>>>> and changed the config so that it was (hd0,0)/ instead of
>>>> (hd0,0)/boot/
>>>> and strange enough that worked just fine. No problem
>> booting up, other
>>>> than failing to find the iso to finish booting.
>>>>
>>>> I need help either getting the original scheme working
>>>>
>> (how can I make
>>
>>>> grub see grub.conf?) or my manual scheme working (how can
>> I tell it where
>>
>>>> the ISO is?)
>>>>
>>>
>>> Is there something in the initrd which sets up the / after
>>>
>> the kernel
>>> start? I am wondering if there is any documentation on the
>> boot process
>>> that I can look at to try and get this working.
>>>
>>
>> No there's only the standard linux documentation.
>> Basically what happens is that the boot loader loads the
>> kernel plus the 'initrd' (Initial RamDisk). Then a script called linuxrc
>> is executed which performs the basic setup, locating of the device and
>> mounting it.
>>
>> --
>>
>>
>> Regards
>> Heiko Zuerker
>> http://www.devil-linux.org
>>
>>
>
> OK. That narrows it to figuring out what linuxrc is doing and maybe
> tweaking it. That script is inside the initrd or outside on the stick?

Inside.

-- 

Regards
  Heiko Zuerker
  http://www.devil-linux.org

Re: [Devil-Linux-discuss] Problem installing to USB

[Patrick S. <pat...@we...>]

Hello!

AFAIK, syslinux supports serial console.

See man syslinux for more information.

Regards,

Patrick Schneider

Fred Frigerio schrieb:

> The problem I have is that I need ttyS0 console because this is a
> headless appliance type router with 3 ports (NexGate NSA1035).
> 

Re: [Devil-Linux-discuss] Problem installing to USB

[Fred F. <ffr...@lo...>]

> -----Original Message-----
> From: dev...@li...=20
> [mailto:dev...@li...] On=20
> Behalf Of Heiko Zuerker
> Sent: Tuesday, February 06, 2007 9:58 AM
> To: dev...@li...
> Subject: Re: [Devil-Linux-discuss] Problem installing to USB
>=20
>=20
> On Tue, February 6, 2007 07:00, Fred Frigerio wrote:
> >>
>=20
> >> At that point suspecting a GRUB problem I wiped the stick=20
> and tried to
> >> rebuild it by hand with the GRUB from my computer (version=20
> 0.97). I did
> >> what I am used to which is a boot partition (sdh1) a root partition
> >> (sdh3) and a config partition (sdh2) all ext2. I installed=20
> grub to sdh1
> >>  and changed the config so that it was (hd0,0)/ instead of
> >> (hd0,0)/boot/
> >> and strange enough that worked just fine. No problem=20
> booting up, other
> >> than failing to find the iso to finish booting.
> >>
> >> I need help either getting the original scheme working=20
> (how can I make
> >> grub see grub.conf?) or my manual scheme working (how can=20
> I tell it where
> >> the ISO is?)
> >>
> >
> > Is there something in the initrd which sets up the / after=20
> the kernel
> > start? I am wondering if there is any documentation on the=20
> boot process
> > that I can look at to try and get this working.
> >
>=20
> No there's only the standard linux documentation.
> Basically what happens is that the boot loader loads the=20
> kernel plus the
> 'initrd' (Initial RamDisk). Then a script called linuxrc is=20
> executed which
> performs the basic setup, locating of the device and mounting it.
>=20
> --=20
>=20
> Regards
>   Heiko Zuerker
>   http://www.devil-linux.org
>=20

OK. That narrows it to figuring out what linuxrc is doing and maybe
tweaking it. That script is inside the initrd or outside on the stick?=20

Re: [Devil-Linux-discuss] Problem installing to USB

[Heiko Z. <he...@zu...>]

On Tue, February 6, 2007 07:00, Fred Frigerio wrote:
>>

>> At that point suspecting a GRUB problem I wiped the stick and tried to
>> rebuild it by hand with the GRUB from my computer (version 0.97). I did
>> what I am used to which is a boot partition (sdh1) a root partition
>> (sdh3) and a config partition (sdh2) all ext2. I installed grub to sdh1
>>  and changed the config so that it was (hd0,0)/ instead of
>> (hd0,0)/boot/
>> and strange enough that worked just fine. No problem booting up, other
>> than failing to find the iso to finish booting.
>>
>> I need help either getting the original scheme working (how can I make
>> grub see grub.conf?) or my manual scheme working (how can I tell it where
>> the ISO is?)
>>
>
> Is there something in the initrd which sets up the / after the kernel
> start? I am wondering if there is any documentation on the boot process
> that I can look at to try and get this working.
>

No there's only the standard linux documentation.
Basically what happens is that the boot loader loads the kernel plus the
'initrd' (Initial RamDisk). Then a script called linuxrc is executed which
performs the basic setup, locating of the device and mounting it.

-- 

Regards
  Heiko Zuerker
  http://www.devil-linux.org

Re: [Devil-Linux-discuss] Problem installing to USB

[Matthew H. <mat...@va...>]

Its all pivot root shenanyganz

-----Original Message-----
From: dev...@li...
[mailto:dev...@li...] On Behalf Of
Fred Frigerio
Sent: 06 February 2007 13:00
To: dev...@li...
Subject: Re: [Devil-Linux-discuss] Problem installing to USB

>=20
> At that point suspecting a GRUB problem I wiped the stick and tried to
> rebuild it by hand with the GRUB from my computer (version 0.97). I
did
> what I am used to which is a boot partition (sdh1) a root partition
> (sdh3) and a config partition (sdh2) all ext2. I installed grub to
sdh1
> and changed the config so that it was (hd0,0)/ instead of
(hd0,0)/boot/
> and strange enough that worked just fine. No problem booting up, other
> than failing to find the iso to finish booting.
>=20
> I need help either getting the original scheme working (how can I make
> grub see grub.conf?) or my manual scheme working (how can I tell it
> where the ISO is?)
>=20

Is there something in the initrd which sets up the / after the kernel
start? I am wondering if there is any documentation on the boot process
that I can look at to try and get this working.


------------------------------------------------------------------------
-
Using Tomcat but need to do more? Need to support web services,
security?
Get stuff done quickly with pre-integrated technology to make your job
easier.
Download IBM WebSphere Application Server v.1.0.1 based on Apache
Geronimo
http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&dat=3D=
121642
_______________________________________________
Devil-linux-discuss mailing list
Dev...@li...
https://lists.sourceforge.net/lists/listinfo/devil-linux-discuss