[Pkg-torrus-general] Torrus problems

[Nenad Tomasevic]

Hello,

I have strange problem with Torrus on Debian, maybe somebody can give me 
some advice..

I have installed Torrus from packages ( 1.0.8 with Cisco CBQOS plugin 
additionally installed from source ) on Debian 6.0 AMD64 and for testing 
purposes I had only few devices in monitoring and everything worked fine.
After that, I have made a script which pulls hostnames and IP's from 
external database, checks SNMP responses and creates one big XML/subtree 
file with all devices that have responded to snmpget ( I need that for 
integration with other monitoring tools ).

Everything works normal until somebody leave web interface opened in 
browser for a longer period ( few hours ).
After few hours BerkeleyDB crashes and everything stops.

I have been advised to compile Torrus dep's from source with  
https://github.com/oetiker/sdbs scripts and additionally compile Torrus 
from source.

I have done that with latest 2.0.1 version but I didn't test it for a 
longer period so I don't know if it works.

In the same time I have installed current stable Debian, installed 
Torrus from package with all updated dependencies from Sid and then I 
have removed Torrus from packages and I have installed latest version 
from source. That combination surely doesn't work, it took longer but 
database still crashes.

Apache log from installation from source which supports fast-cgi but 
Apache complains on database in every case;

[Tue Oct 04 13:38:18 2011] [error] [client 192.168.250.118] FastCGI: 
server "/opt/torrus/bin/torrus.fcgi" stderr: [04-Oct-2011 13:38:17 ] 
User logged in: bnet, referer: http://torrus.bnet.lan/torrus
[Tue Oct 04 15:47:58 2011] [error] [client 192.168.250.118] FastCGI: 
server "/opt/torrus/bin/torrus.fcgi" stderr: [04-Oct-2011 15:47:58*] 
Cannot open database sub/main/paramprops_1.db:  PANIC: fatal region 
error detected; run recovery, referer: 
http://torrus.bnet.lan/torrus/main?path=/subtree/hostname/Interface_Counters/&view=overview-subleaves-html&OVS=errors
[Tue Oct 04 15:47:58 2011] [error] [client 192.168.250.118] FastCGI: 
server "/opt/torrus/bin/torrus.fcgi" stderr: [04-Oct-2011 15:47:58*] 
Cannot open database sub/main/tokensets_1.db:  PANIC: fatal region error 
detected; run recovery, referer: 
http://torrus.bnet.lan/torrus/main?path=/subtree/hostname/Interface_Counters/&view=overview-subleaves-html&OVS=errors
[Tue Oct 04 15:47:58 2011] [error] [client 192.168.250.118] FastCGI: 
server "/opt/torrus/bin/torrus.fcgi" stderr: [04-Oct-2011 15:47:58*] 
Cannot create BerkeleyDB Environment: DB_RUNRECOVERY: Fatal error, run 
database recovery, referer: 
http://torrus.bnet.lan/torrus/main?path=/subtree/hostname/Interface_Counters/&view=overview-subleaves-html&OVS=errors

Recovery fix things every time.

Did any of you had such problem and did you manage to solve it?


Nenad