Good point there. The part i don’t understand is why he is using a machine with considerable resources to run the manager which isn’t the most resource intensive process of OpenVAS. To used that hardware properly virtualization to have one Manager in one VM and a Scanner in another VM would be a good option.
Fabio > No dia 08/01/2017, às 02:33, Reindl Harald <[email protected]> escreveu: > > to make it more clear for people who think "my single process has to utilize > my hardware": > > you won't get far when you complain why your webserver does not utilize your > CPU and memory as much as you like it because it spends most of the time by > wait for the overloaded database-server which can not serve more concurrency > > in that case you need a database cluster to spread at least your > select-queries over more than one instance to reduce the impact of lockings > > and that may work well even if that instances are on the same host in > multiple virtual machines because even HPC setups exists in the real world > who prove that theory - welcome in 2017! > > https://gcn.com/articles/2014/07/11/vgrid-hpc-virtualization.aspx > > http://www.zdnet.com/article/yes-virtualization-is-faster-sometimes-than-native-hardware/ > > "However, by partitioning each host into two or four virtual machines, they > were able to get significantly better performance" > > Am 08.01.2017 um 03:23 schrieb Reindl Harald: >> Am 08.01.2017 um 03:19 schrieb Fábio Fernandes: >>> In my opinion i think that those resources would be better spent on >>> openvassd process since it does the heavy lifting. >> >> when the bottleneck are shared ressources used by openvassd it can't do >> anyhting about it and so you need to get rid of the concurrecncy by just >> have more of this ressources aka more instances >> >>>> No dia 08/01/2017, às 00:59, Reindl Harald <[email protected]> >>>> escreveu: >>>> >>>> Am 08.01.2017 um 01:30 schrieb TN TN: >>>>> HI Christian, I actually have it installed with postgresql, but when >>>>> you're scanning thousands of internal hosts on a weekly basis it slows >>>>> it down considerably over time.. I'm just curious on why openvasmd is >>>>> not using enough resources on the server. Thanks TN >>>> >>>> because they are not available which means you assume the bottleneck >>>> likely where it isn't - when you scan thousands of hosts how do you >>>> come to the conclusion at a single openvas machine will be enough? >>>> >>>> just install *more* instances as virtual machines when you say you >>>> are not cpu-bound and have enough memory and spread the load - >>>> problem solved >>>> >>>>> On 7 January 2017 at 17:09, Christian Fischer >>>>> <[email protected] >>>>> <mailto:[email protected]>> wrote: >>>>> >>>>> Hi, >>>>> >>>>> On 07.01.2017 19:43, TN TN wrote: >>>>> > Hi, I have a fairly powerful server and it seems like the >>>>> openvasmd >>>>> > process isnt using up alot of the resources on the box. If I run >>>>> > numerous scans (the scanner being off of the main openvasmd >>>>> server), the >>>>> > openvasmd process barely uses any resources, however Greenbone >>>>> slows >>>>> > down and the backend commands take much longer to execute. Is >>>>> there a >>>>> > way to force openvasmd to use more resources (CPU/Memory) so >>>>> that it >>>>> > runs faster? Thanks, TN >>>>> >>>>> i don't think this is possible and it also probably wouldn't help >>>>> here. >>>>> The manager process is using the resources it needs. >>>>> >>>>> I think the bottle neck is more likely the sqlite database >>>>> backend which >>>>> slows down if multiple scanners are sending data to the manager. >>>>> Have a >>>>> look at the postgresql database backend which might help to >>>>> improve the >>>>> performance. > _______________________________________________ > Openvas-discuss mailing list > [email protected] > https://lists.wald.intevation.org/cgi-bin/mailman/listinfo/openvas-discuss _______________________________________________ Openvas-discuss mailing list [email protected] https://lists.wald.intevation.org/cgi-bin/mailman/listinfo/openvas-discuss
