Dear Burton,
 
I finaly came back from my trip and i have tried your patch for the FreeBSD
daemon mode, "everything" now work well, thank you very much. Sorry about
the lack of feedback but i was abroad and just able to take my email but not
able to try the suggestion, and to be fair, the high level of the discussion
was far away from my programmer skills. So i watched the battle about
pthreads, kqueue, select....
 
About the memory usage
 
I have deployed Ntop cvs 2.2.99 (last cvs at 5 GMT today),  on a FreeBSD 4.8
STABLE, hardware is Xeon 2.8 Ghz, memory 512, this hardware is acting as a
gateway (one interface on a switched lan), second interface go to a external
link 4 Mb leased line so it's merely snifing the 4mb line 
 
The error message are quite clear, i don't have enough memory, so i'm just
looking for tips or statistics from other ntop user to know what kind of
network they monitor and with how much gigabites :)
 
Looking to the back traffic on ntop, ntop dev, i found few articles about
memory (old version of ntop) more about memory leak,...and also the use of
the switch -C for disabling feature (for a user with 256 Mb)
 
I tried to look at snapshot.ntop.org on the FAQ but this one seems to be
down at the time that i wrote.
 
 
The obvious:
Here is the message in the log file:
 
Feb 11 19:30:47 oxfrouter02 ntop[47072]:   WEB: ntop's web server is now
processing requests
Feb 11 19:30:47 oxfrouter02 ntop[47072]:   THREADMGMT: pcap dispatch thread
running...
Feb 11 19:30:57 oxfrouter02 ntop[47072]:   **ERROR** CHKVER: Unable to
resolve site
Feb 11 20:33:02 oxfrouter02 ntop[47072]:   **FATAL_ERROR** malloc(3836) @
hash.c:936 returned NULL [no more memory?]
Feb 11 20:33:02 oxfrouter02 ntop[47072]:   **WARNING** ntop packet capture
STOPPED
Feb 11 20:33:02 oxfrouter02 ntop[47072]:   NOTE: ntop web server remains up
Feb 11 20:33:02 oxfrouter02 ntop[47072]:   NOTE: Shutdown gracefully and
restart with more memory
Feb 11 20:33:02 oxfrouter02 ntop[47072]:   **ERROR** Sanity check failed (2)
[Low memory?]
Feb 11 20:33:02 oxfrouter02 ntop[47072]:   THREADMGMT: pcap dispatch thread
terminated...
Feb 11 20:36:27 oxfrouter02 ntop[47072]:   THREADMGMT: Idle Scan thread
(137303040) terminated

 
 
So the web report is still alive, and a quick look at the memory in use via
top
 
last pid: 52324;  load averages:  0.00,  0.00,  0.00
up 30+09:09:20  22:04:04
43 processes:  1 running, 42 sleeping
CPU states:  0.0% user,  0.0% nice,  0.0% system,  0.6% interrupt, 99.4%
idle
Mem: 241M Active, 134M Inact, 102M Wired, 22M Cache, 61M Buf, 2348K Free
Swap: 1008M Total, 247M Used, 762M Free, 24% Inuse
 
  PID USERNAME PRI NICE  SIZE    RES STATE  C   TIME   WCPU    CPU COMMAND
52324 root      28   0  1920K  1072K CPU1   1   0:00  1.32%  0.29% top
  480 mysql      2   0 30120K   196K poll   1 111:34  0.00%  0.00% mysqld
47072 ntop       2   0   535M   820K poll   1  26:21  0.00%  0.00% ntop

 
Ntop is using 535 Mb, and i got only 2Mb free, snif !
 
So is it possible for people to tell me some tips/guidelines, i know that
it's not so evident as it depends of the topology, number of hosts,
traffic...peak, average use of the bandwidth. I'm going to upgrade anyway
this box to a bi xeon with 2 gb of memory but would it be enough...
 
Thanks in advance
Laurent Grilli
_______________________________________________
Ntop mailing list
[EMAIL PROTECTED]
http://listgateway.unipi.it/mailman/listinfo/ntop

Reply via email to