Gary
ntop definitively spawns threads at the beginning and not dynamically. If your info is correct there's something very broken on ntop+FreeBSD. Please check again

Luca

On May 15, 2009, at 12:39 AM, Gary Gatten wrote:

<!-- /* Font Definitions */ @font-face {font-family:Tahoma; panose-1:2 11 6 4 3 5 4 4 2 4;} /* Style Definitions */ p.MsoNormal, li.MsoNormal, div.MsoNormal {margin:0in; margin-bottom:.0001pt; font- size:12.0pt; font-family:"Times New Roman";} a:link, span.MsoHyperlink {color:blue; text-decoration:underline;} a:visited, span.MsoHyperlinkFollowed {color:purple; text- decoration:underline;} span.EmailStyle17 {mso-style-type:personal; font-family:Arial; color:windowtext;} span.EmailStyle18 {mso-style- type:personal-reply; color:black;} @page Section1 {size:8.5in 11.0in; margin:1.0in 1.25in 1.0in 1.25in;} div.Section1 {page:Section1;} -->
Definitely odd thread behavior compared to 3.2.1.



3.2.1 will spawn whatever threads it needs and the count is fairly stable as is the priority. 3.3.9 seems to constantly spawn and tear down threads; sometimes there are 14 (was 20, I decreased the DNSAR from 9 to 3) and sometimes only 3 or 4! How can this be when I have 8 interfaces? Nonetheless, when the thread count is high, the priority is higher (4 - 20) and everything is happy – works great. Then for some reason the thread count drops, priority drops (>100) and everything goes to $hit! The cycle repeats over and over.



I’ve tried hacking Make and configure files, running as root, running in fg / non-daemon, updated to gcc44 and compiled with that. Upgraded libtool and several others – nothing seems to be helping much.







From: Gary Gatten
Sent: Thursday, May 14, 2009 11:15 AM
To: '[email protected]'; '[email protected]'
Subject: 3.3.x threading issue on FreeBSD 6.0



Still trying to track down a high load problem since migrating to 3.3.x – tested 3.3.8 and 3.3.9 and maybe 3.3.1 and 3.3.3, but that was so long ago I don’t recall for sure.



Anyway, on 3.2.1 it starts 20 threads for the nTop instance in question (8 netflow interfaces) and it operates perfect using roughly 10%-20% of CPU. 20 threads seem to be constant /stable.



On 3.3.x however, the threads are dynamic from 3 – 20. When the thread count is high it behaves similar to 3.2.1: the netflow queues are serviced promptly and cpu load is 10% - 20%. However, for some reason it kills threads and when there are less than maybe…. 18 threads, bad things happen: cpu maxes out at 100% and netflow queues backup and overflow…



Any help would be great. I’ll even talk my company into making a donation – which I was going to do anyway, but will put on the front burner if I can get some help with this!



TIA!



Gary



"This email is intended to be reviewed by only the intended recipient and may contain information that is privileged and/or confidential. If you are not the intended recipient, you are hereby notified that any review, use, dissemination, disclosure or copying of this email and its attachments, if any, is strictly prohibited. If you have received this email in error, please immediately notify the sender by return email and delete this email from your system."
_______________________________________________
Ntop mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop

_______________________________________________
Ntop-dev mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop-dev

Reply via email to