Oops. I meant to send this to the list.
Blake Caldwell wrote:
> I used to have the same problem. A while back I increased my packet
> buffer size to 16MB. A recent denial of service attack though casued
> the router to export many more flows than I am used to collecting and
> cflowd crashed. Now I have it set to 32 MB. I hope this is big enough
> ;)
>
> -Blake
>
> "Nunes, Steve" wrote:
>
> >
> >
> > Brett,
> >
> > Some people have been reporting this since the fixes in 2.1-b1. See
> > the archive...
> >
> > http://www.caida.org/archives/cflowd/0154.html
> >
> >
> > Steve Nunes
> > CoreExpress Inc.
> > E-mail: steve.nunes@coreexpress.net
> >
> > -----Original Message-----
> > From: Brett L. Hawn [mailto:brett.hawn@rcn.com]
> > Sent: Saturday, February 10, 2001 9:12 PM
> > To: cflowd@caida.org
> > Subject: cflowd dieing randomly
> >
> > I'm running a semi-production/semi-test version of cflowd on a
> > Netra360 w/
> > 512mb ram, 1gig swap, and tons of drive space. Was working fine when I
> > was
> > only exporting about 3 routers worth of data, kicked it up to 5 and it
> > now
> > its going to hell. After a reboot cflowd will run for about 30 minutes
> > or so
> > then I get spammed with the following in my syslog (and cflowd cores
> > out):
> >
> > Feb 10 22:58:54 flowscan.nms.rcn.net cflowd[184]: [E]
> > semop(0,0xefe34b80,1)
> > failed to release buffer lock: Resource temporarily unavailable
> > {CflowdPacketQueue.cc:438}
> >
> > I've done all the necessary steps like setting my shmem size, etc.
> > Anyone
> > got an idea what might be causing this?
> >
> > --
> > cflowd mailing list
> > cflowd@caida.org
-- cflowd mailing list cflowd@caida.org
This archive was generated by hypermail 2b29 : Mon Feb 12 2001 - 17:26:06 PST