[gentoo-hardened] resource overstep RLIMIT_CORE

[gentoo-hardened] resource overstep RLIMIT_CORE

[Jan Klod]

Hi list,

could someone explain the meaning and possible causes (if possible) of this 
message: "denied resource overstep by requesting 4096 for RLIMIT_CORE against 
limit 0?

Re: [gentoo-hardened] resource overstep RLIMIT_CORE

[Manuel Leithner]

On Tue, 20 Jan 2009 11:40:33 +0100, Jan Klod wrote:
> Hi list,
> 
> could someone explain the meaning and possible causes (if possible) of this 
> message: "denied resource overstep by requesting 4096 for RLIMIT_CORE against 
> limit 0?
> 
> 

One of your processes crashed and tried to dump its core, but since the RLIMIT_CORE is set to 0 (i. e. no core dumps are to be created), it can't do so.

Re: [gentoo-hardened] resource overstep RLIMIT_CORE

[Jan Klod]

On Tuesday 20 January 2009 10:49:24 Manuel Leithner wrote:
> On Tue, 20 Jan 2009 11:40:33 +0100, Jan Klod wrote:
> > Hi list,
> >
> > could someone explain the meaning and possible causes (if possible) of
> > this message: "denied resource overstep by requesting 4096 for
> > RLIMIT_CORE against limit 0?
>
> One of your processes crashed and tried to dump its core, but since the
> RLIMIT_CORE is set to 0 (i. e. no core dumps are to be created), it can't
> do so.

Ah, so it is just a secondary problem. Its ssmtp crashing. First it makes a 
segfault, then receives signal 11 by grsec, then overstep. Anyone here using 
it with grsec? 

Re: [gentoo-hardened] resource overstep RLIMIT_CORE

[Ned Ludd]

On Tue, 2009-01-20 at 11:17 +0000, Jan Klod wrote:
> On Tuesday 20 January 2009 10:49:24 Manuel Leithner wrote:
> > On Tue, 20 Jan 2009 11:40:33 +0100, Jan Klod wrote:
> > > Hi list,
> > >
> > > could someone explain the meaning and possible causes (if possible) of
> > > this message: "denied resource overstep by requesting 4096 for
> > > RLIMIT_CORE against limit 0?
> >
> > One of your processes crashed and tried to dump its core, but since the
> > RLIMIT_CORE is set to 0 (i. e. no core dumps are to be created), it can't
> > do so.
> 
> Ah, so it is just a secondary problem. Its ssmtp crashing. First it makes a 
> segfault, then receives signal 11 by grsec, then overstep. 

A segfault is signal 11. It's not two events but one. And grsec is not
sending the signal in anyway but rather just logging that it happened.

> Anyone here using 
> it with grsec? 

Lots of people.

Re: [gentoo-hardened] resource overstep RLIMIT_CORE

[Jan Klod]

On Tuesday 20 January 2009 15:45:25 Ned Ludd wrote:
> On Tue, 2009-01-20 at 11:17 +0000, Jan Klod wrote:
> > On Tuesday 20 January 2009 10:49:24 Manuel Leithner wrote:
> > > On Tue, 20 Jan 2009 11:40:33 +0100, Jan Klod wrote:
> > > > Hi list,
> > > >
> > > > could someone explain the meaning and possible causes (if possible)
> > > > of this message: "denied resource overstep by requesting 4096 for
> > > > RLIMIT_CORE against limit 0?
> > >
> > > One of your processes crashed and tried to dump its core, but since the
> > > RLIMIT_CORE is set to 0 (i. e. no core dumps are to be created), it
> > > can't do so.
> >
> > Ah, so it is just a secondary problem. Its ssmtp crashing. First it makes
> > a segfault, then receives signal 11 by grsec, then overstep.
>
> A segfault is signal 11. It's not two events but one. And grsec is not
> sending the signal in anyway but rather just logging that it happened.
>
> > Anyone here using
> > it with grsec?
>
> Lots of people.

In my case it is 2.62-r{2,5} on hardened-sources-2.6.25 with most of 
grsecurity options enabled and I can't get it working... Maybe someone of 
those lots has some suggestion? How to find out the problem or what is known 
to work?

Re: [gentoo-hardened] resource overstep RLIMIT_CORE

[Magnus Granberg]

On Tuesday 20 January 2009 20.25.22 Jan Klod wrote:
> On Tuesday 20 January 2009 15:45:25 Ned Ludd wrote:
> > On Tue, 2009-01-20 at 11:17 +0000, Jan Klod wrote:
> > > On Tuesday 20 January 2009 10:49:24 Manuel Leithner wrote:
> > > > On Tue, 20 Jan 2009 11:40:33 +0100, Jan Klod wrote:
> > > > > Hi list,
> > > > >
> > > > > could someone explain the meaning and possible causes (if possible)
> > > > > of this message: "denied resource overstep by requesting 4096 for
> > > > > RLIMIT_CORE against limit 0?
> > > >
> > > > One of your processes crashed and tried to dump its core, but since
> > > > the RLIMIT_CORE is set to 0 (i. e. no core dumps are to be created),
> > > > it can't do so.
> > >
> > > Ah, so it is just a secondary problem. Its ssmtp crashing. First it
> > > makes a segfault, then receives signal 11 by grsec, then overstep.
> >
> > A segfault is signal 11. It's not two events but one. And grsec is not
> > sending the signal in anyway but rather just logging that it happened.
> >
> > > Anyone here using
> > > it with grsec?
> >
> > Lots of people.
>
> In my case it is 2.62-r{2,5} on hardened-sources-2.6.25 with most of
> grsecurity options enabled and I can't get it working... Maybe someone of
> those lots has some suggestion? How to find out the problem or what is
> known to work?

http://forums.gentoo.org/viewtopic-t-728429.html
http://bugs.gentoo.org/show_bug.cgi?id=253733
http://bugs.gentoo.org/show_bug.cgi?id=254843
We have some RLIMIT bug in the hardened-sources
Gordon Malm know more about it.
/Zorry

Re: [gentoo-hardened] resource overstep RLIMIT_CORE

[Jan Klod]

On Tuesday 20 January 2009 15:45:25 Ned Ludd wrote:
> On Tue, 2009-01-20 at 11:17 +0000, Jan Klod wrote:
> > On Tuesday 20 January 2009 10:49:24 Manuel Leithner wrote:
> > > On Tue, 20 Jan 2009 11:40:33 +0100, Jan Klod wrote:
> > > > Hi list,
> > > >
> > > > could someone explain the meaning and possible causes (if possible)
> > > > of this message: "denied resource overstep by requesting 4096 for
> > > > RLIMIT_CORE against limit 0?
> > >
> > > One of your processes crashed and tried to dump its core, but since the
> > > RLIMIT_CORE is set to 0 (i. e. no core dumps are to be created), it
> > > can't do so.
> >
> > Ah, so it is just a secondary problem. Its ssmtp crashing. First it makes
> > a segfault, then receives signal 11 by grsec, then overstep.
>
> A segfault is signal 11. It's not two events but one. And grsec is not
> sending the signal in anyway but rather just logging that it happened.
>
> > Anyone here using
> > it with grsec?
>
> Lots of people.

Ok, my advice is: don't use some of characters \#,\$,\,, maybe \( in password 
for ssmtp...