Hello Klaus:

that is very interesting indeed; though I see you are using the Atomic repo.  
From a source perspective I have compiled the following chain of code:

libmicrohttpd-0.9.22
greenbone-security-assistant-3.0.3
openvas-scanner-3.3.1
openvas-libraries-5.0.2
openvas-manager-3.0.3
openvas-administrator-1.2.1
openvas-libraries-5.0.3

as soon as one tries to access the WEBUI GSAD starts to spin out of control; 
the other day when I left it go to a load of 188 :D

something must have got introduced with a recent update that has upset the 
install.

the major difference between our install is that I have tried with a newer 
version of libmicro:

$ ls -l /usr/lib64/libmicrohttpd.so.10
lrwxrwxrwx 1 root root 24 Oct 26 14:01 /usr/lib64/libmicrohttpd.so.10 -> 
libmicrohttpd.so.10.16.1

which is a tad worrying as the one you are running is not even listed on 
http://ftp.heanet.ie/mirrors/gnu/libmicrohttpd/ now so need to see when I can 
get an older version from.

Thanks.

----- Original Message ----- 
From: "Klaus Schulte (NSN - DE/Ulm)" <[email protected]> 
To: "Phil Daws" <[email protected]>, [email protected] 
Sent: Thursday, 1 November, 2012 6:21:37 PM 
Subject: RE: [Openvas-discuss] openvas gnutls bug again? 

Hi Phil, 

gsad ON CentOS 6.3 works here. 

[root@sally gsa]# ldd /usr/sbin/gsad |grep libmicro 
libmicrohttpd.so.10 => /usr/lib64/libmicrohttpd.so.10 
(0x00007fef4c0c3000) 

[root@sally gsa]# ls -l /usr/lib64/libmicrohttpd.so.10 
lrwxrwxrwx. 1 root root 23 Oct 10 15:05 /usr/lib64/libmicrohttpd.so.10 
-> libmicrohttpd.so.10.5.0 

[root@sally ~]# rpm -qa |grep openvas 
openvas-administrator-1.2.1-2.el6.art.x86_64 
openvas-scanner-3.3.1-2.el6.art.x86_64 
openvas-cli-1.1.5-3.el6.art.x86_64 
openvas-1.0-3.el6.art.noarch 
openvas-libraries-5.0.3-4.el6.art.x86_64 
openvas-manager-3.0.3-3.el6.art.x86_64 

[root@sally ~]# rpm -qa |grep green 
greenbone-security-assistant-3.0.3-3.el6.art.x86_64 

[root@sally ~]# rpm -qa |grep tls 
gnutls-2.8.5-4.el6_2.2.x86_64 

[root@sally ~]# cat /etc/centos-release 
CentOS release 6.3 (Final) 

Best regards 
Klaus 

-----Original Message----- 
From: Openvas-discuss 
[mailto:[email protected]] On Behalf Of ext 
Phil Daws 
Sent: Thursday, November 01, 2012 15:49 
To: [email protected] 
Subject: Re: [Openvas-discuss] openvas gnutls bug again? 

am having the same problem on a CentOS 6.3 server with GSAD running at 
198% and driving up the load. 

[root@va1 libmicrohttpd-0.9.22]# rpm -qa | grep gnutls 
gnutls-devel-2.8.5-4.el6_2.2.x86_64 
gnutls-2.8.5-4.el6_2.2.x86_64 

[root@va1 libmicrohttpd-0.9.22]# ldd /usr/local/openvas/sbin/gsad | grep 
libmicro 
libmicrohttpd.so.10 => /usr/lib64/libmicrohttpd.so.10 
(0x00007f318c9ee000) 

have tried a newer version of libmicrohttp built from source and still 
have the same problem. 

Thoughts ? 
----- Original Message ----- 
From: "Reindl Harald" <[email protected]> 
To: [email protected] 
Sent: Thursday, 1 November, 2012 1:55:40 PM 
Subject: Re: [Openvas-discuss] openvas gnutls bug again? 



Am 01.11.2012 14:38, schrieb Michael Meyer: 
>> interesting is that i made a rpm for libmicrohttpd-0.9.15, downgraded 

>> and built any opnvas/gsa package again after downgrade with no 
success 
> 
> There is no need to rebuild the gsad. Just restart it after switching 
> the libmicrohttpd version. 

if you work without package managment maybe 
with RPM packages you have DEPENDENCIES 

so you have to 
* rpm -e --nodeps greenbone-security-assistant 
* update to the new built libmicrohttpd-devel 
* rebuilt GSA against the new library 
* reinstall GSA 

these are the steps on the build/repo-server to 
have clean packages which are deployed vi yum 

> Is the gsa really using the right libmicrohttpd version? 
> (ldd /usr/sbin/gsad | grep libmicro) 

surely beasue ona RPM based system you can only ave one 

> Which gnutls version you are using? Here it is 
> gnutls-2.12.17-1.fc17.i686. 

gnutls-2.12.17-1.fc17.x86_64 
libmicrohttpd-0.9.7-2.fc17.20121016.rh.x86_64 

tried a lot of libmicrohttpd versions in the past 
with two different results "Login failed. OMP service 
is down" or no response at all with high cpu load 

> What exactly means "no success"? Is the CPU at > 95%? 

"Login failed. OMP service is down" on the website 

[root@openvas:/var/log/openvas]$ cat *.log 
lib auth:WARNING:2012-11-01 14h52.55 CET:23258: Authentication 
configuration could not be loaded. 
lib auth:WARNING:2012-11-01 13h52.56 utc:23257: Authentication 
configuration could not be loaded. 
lib serv:WARNING:2012-11-01 13h53.13 utc:23269: openvas_server_new: 
failed to set credentials key file 
lib serv:WARNING:2012-11-01 13h53.13 utc:23269: openvas_server_new: cert 
file: /var/lib/openvas/CA/clientcert.pem 
lib serv:WARNING:2012-11-01 13h53.13 utc:23269: openvas_server_new: key 
file : 
/var/lib/openvas/private/CA/clientkey.pem 
lib serv:WARNING:2012-11-01 13h53.13 utc:23269: Failed to gnutls_bye: 
Error in the push function. 
[Sun Oct 28 00:02:14 2012][6409] openvassd 3.3.1 started 
[Thu Nov 1 13:52:45 2012][6409] received the TERM signal 
[Thu Nov 1 13:52:55 2012][23255] openvassd 3.3.1 started 



_______________________________________________ 
Openvas-discuss mailing list 
[email protected] 
http://lists.wald.intevation.org/cgi-bin/mailman/listinfo/openvas-discus 
s 

_______________________________________________ 
Openvas-discuss mailing list 
[email protected] 
http://lists.wald.intevation.org/cgi-bin/mailman/listinfo/openvas-discus 
s 

_______________________________________________
Openvas-discuss mailing list
[email protected]
http://lists.wald.intevation.org/cgi-bin/mailman/listinfo/openvas-discuss

Reply via email to