[Unbound-users] out of memory caching validator results

[Slingerland, Michael van]

Hi Wouter,

I found the issue:

Even though I set the datasize in login.conf to 2048M.
Ulimit -d still only allows 1024M, That's why unbound crashes.

Manually setting ulimit to 2048M is not possible.

So I need to dig into the kernel parameters to override this.

Thanks for the help :)
Mike 

-----Original Message-----
From: W.C.A. Wijngaards [mailto:wouter at NLnetLabs.nl] 
Sent: Monday, 31 May 2010 10:38
To: Slingerland, Michael van
Subject: Re: Uncaught bounce notification

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

Hi Mike,

You used the wrong email address for this email.  It will not be posted on the mailing list.

(you used unbound-users-bounces, but need to use unbound-users).

The limits look good, for the data-seg-size, but in my ulimit -a output I also see a max memory limit (but this is on linux), so perhaps you need to configure that one too?

If you su - unbound and ulimit -a, what is the output?

Best regards,
   Wouter


On 05/31/2010 10:31 AM, mailman-bounces at NLnetLabs.nl wrote:
> Subject:
> [Unbound-users] out of memory caching validator results
> From:
> "Slingerland, Michael van" <Michael.van.Slingerland at t-mobile.nl>
> Date:
> Mon, 31 May 2010 10:01:21 +0200
> 
> To:
> "unbound-users-bounces at NLnetLabs.nl" 
> <unbound-users-bounces at nlnetlabs.nl>
> 
> 
> Hi Wouter,
> 
> It seems I am hitting some resource limit on openbsd.
> Cause the issue happened again yesterday, at around of 1.5G of total 
> memory usage
> 
> May 30 12:32:19 ypdnscache01 unbound: [26071:0] error: store_msg: 
> malloc failed May 30 12:32:19 ypdnscache01 unbound: [26071:0] error: 
> store_msg: malloc failed May 30 12:32:19 ypdnscache01 unbound: 
> [26071:0] error: out of memory caching validator results May 30 
> 12:32:19 ypdnscache01 unbound: [26071:0] error: store_msg: malloc 
> failed May 30 12:32:19 ypdnscache01 unbound: [26071:0] error: 
> store_msg: malloc failed May 30 12:32:19 ypdnscache01 unbound: 
> [26071:0] error: out of memory caching validator results May 30 
> 12:32:19 ypdnscache01 unbound: [26071:0] error: store_msg: malloc 
> failed May 30 12:32:19 ypdnscache01 unbound: [26071:0] error: out of 
> memory caching validator results May 30 12:32:19 ypdnscache01 unbound: 
> [26071:0] error: store_msg: malloc failed May 30 12:32:19 ypdnscache01 
> unbound: [26071:0] error: error_response_cache: could not store error 
> (nomem) May 30 12:32:19 ypdnscache01 unbound: [26071:0] error: 
> error_response_cache: could not store error (nomem) May 30 12:32:19 
> ypdnscache01 unbound: [26071:0] error: store_msg: malloc failed May 30 
> 12:32:19 ypdnscache01 unbound: [26071:0] error: out of memory caching 
> validator results May 30 12:32:19 ypdnscache01 unbound: [26071:0] 
> error: out of memory caching validator results May 30 12:32:19 
> ypdnscache01 unbound: [26071:0] error: store_msg: malloc failed May 30 
> 12:32:19 ypdnscache01 unbound: [26071:0] error: out of memory caching 
> validator results May 30 12:32:19 ypdnscache01 unbound: [26071:0] 
> error: out of memory caching validator results May 30 12:32:19 
> ypdnscache01 unbound: [26071:0] error: store_msg: malloc failed May 30 
> 12:32:19 ypdnscache01 unbound: [26071:0] error: out of memory caching 
> validator results May 30 12:32:19 ypdnscache01 unbound: [26071:0] 
> error: store_msg: malloc failed May 30 12:32:19 ypdnscache01 unbound: 
> [26071:0] error: out of memory caching validator results May 30 
> 12:32:19 ypdnscache01 unbound: [26071:0] error: out of memory caching 
> validator results May 30 12:32:19 ypdnscache01 unbound: [26071:0] 
> error: store_msg: malloc failed May 30 12:32:19 ypdnscache01 unbound: 
> [26071:0] error: store_msg: malloc failed May 30 12:32:19 ypdnscache01 
> unbound: [26071:0] error: prealloc: out of memory May 30 12:32:19 
> ypdnscache01 unbound: [26071:0] fatal error: alloc_special_obtain: out 
> of memory
> 
> After this I reduced the message cache to 150M and rrset cache to 300M to see if it won't run out of memory again. Up until now it runs stable.
> 
> Now I have to find which resource limit is the culprit..
> My login.conf is set as dscribed below.
> 
> default:\
>         :path=/usr/bin /bin /usr/sbin /sbin /usr/X11R6/bin /usr/local/bin:\
>         :umask=022:\
>         :datasize-max=2048M:\
>         :datasize-cur=2048M:\
>         :maxproc-max=128:\
>         :maxproc-cur=64:\
>         :openfiles-cur=128:\
>         :stacksize-cur=4M:\
>         :localcipher=blowfish,6:\
>         :ypcipher=old:\
>         :tc=auth-defaults:\
>         :tc=auth-ftp-defaults:
> 
> Does anybody has advice which parameters needs to be tuned as well next to "datasize"...
> 
> Thanks,
> Mike

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org/

iEYEARECAAYFAkwDdW8ACgkQkDLqNwOhpPiSHwCdGQzywbd/N2bKej++Jnun889X
Qf4AnAoej4NlOYNp9pK+0NIpiu5BxpyK
=a4wW
-----END PGP SIGNATURE-----