The machine is an old pentium 500Mhz machine (or something similar), with about 128MB of RAM.
 
After the difference in performance that I saw, I wouldn't advise spending any money until you at least try to adjust your max hosts and max checks etc.
 
Checks:
 max_hosts = 5
 max_checks = 10
 
Subnets:
- they are very dense ...
- er .. I mean we have about 500 nodes/devices split across about 10 networks
 
Scanners:
- only run nmap.
 
Server Prefs:
 
begin(SERVER_PREFS)
 max_hosts = 5
 max_checks = 10
 log_whole_attack = no
 cgi_path = /cgi-bin:/scripts
 port_range = default
 optimize_test = yes
 language = english
 checks_read_timeout = 3
 non_simult_ports = 139, 445
 plugins_timeout = 150
 safe_checks = yes
 auto_enable_dependencies = yes
 use_mac_addr = no
 save_knowledge_base = yes
 kb_restore = yes
 > > kb_dont_replay_scanners = no
 kb_dont_replay_info_gathering = no
 kb_dont_replay_attacks = no
 kb_dont_replay_denials = no
 kb_max_age = 2764800
 diff_scan = yes
 plugin_upload = no
 plugin_upload_suffixes = .nasl, .inc
 slice_network_addresses = no
end(SERVER_PREFS)


[EMAIL PROTECTED] wrote:
>>
When I was having trouble with scans finishing/running slowly, I had about 25 hosts and about 25 tests. Scans took days, my RAM was maxed out and the paging to and frmo disk was really high.

Then I dropped my hosts down to about 5 and my tests to about 5 - the results were amazing. Machine no longer maxed out the CPU or the RAM, paging to disk dropped dramatically and, best of all, the tests finished in hours instead of days. I currently scan multiple /24 networks (about 3-4 a night) and the time is between 1-2.5 hours each.
<<

There are a lot of variables you leave out. What is the system
configuration of the scanner? How many total checks do you run?
How dense is the population on those subnets (boy, can that
question have a double meaning!)



When you talk about the the load average sitting at around 30 or more, are you talking about the CPU? If so, that shouldn't be of any concern - the *best* spot your CPU can sit is at 99% - as long as it never hits 100% of course ;)

You certainly have heaps of RAM too ...

Have yuo noticed lots of disk activity? Particularly paging to and from disk?

The reason I ask is that you have a hell of a lot of tests (and presumably threads) running at once: max_hosts = 96  max_checks = 16  gives 1536 threads (I think - anyone ...is it one thread per test??).

When I was having trouble with scans finishing/running slowly, I had about 25 hosts and about 25 tests. Scans took days, my RAM was maxed out and the paging to and frmo disk was really high.

Then I dropped my hosts down to about 5 and my tests to about 5 - the results were amazing. Machine no longer maxed out the CPU or the RAM, paging to disk dropped dramatically and, best of all, the tests finished in hours instead of days. I currently scan multiple /24 networks (about 3-4 a night) and the time is between 1-2.5 hours each.

Hope that helps.

Antony 



Find local movie times and trailers on Yahoo! Movies.
_______________________________________________
Nessus mailing list
[EMAIL PROTECTED]
http://mail.nessus.org/mailman/listinfo/nessus



Find local movie times and trailers on Yahoo! Movies.
_______________________________________________
Nessus mailing list
[EMAIL PROTECTED]
http://mail.nessus.org/mailman/listinfo/nessus

Reply via email to