The hash should be large enough to hold 45 vlans.  Sounds like something's
wrong - probably need to get Luca involved.

-----Burton

-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of Tony
Hernandez
Sent: Thursday, February 27, 2003 7:49 AM
To: [EMAIL PROTECTED]
Subject: RE: [Ntop] updateElementHash(): hash full!


Sorry, around 9,000 hosts as confirmed by my boss. and 45 vlans but very
many subnets. What do you suggest the has size should be in this instance?

-----Original Message-----
From: Tony Hernandez
Sent: Thursday, February 27, 2003 7:38 AM
To: [EMAIL PROTECTED]
Subject: RE: [Ntop] updateElementHash(): hash full!


A lot. I'll get back to you on that number... Last I knew we had over
6,500 hosts.

-----Original Message-----
From: Burton M. Strauss III [mailto:[EMAIL PROTECTED]
Sent: Wednesday, February 26, 2003 10:44 PM
To: [EMAIL PROTECTED]
Subject: RE: [Ntop] updateElementHash(): hash full!


Yeah ... how many dang ASs or VLANs are you monitoring anyway?

-----Burton

-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of Tony
Hernandez
Sent: Wednesday, February 26, 2003 8:39 PM
To: [EMAIL PROTECTED]
Subject: RE: [Ntop] updateElementHash(): hash full!


So, I can try increasing the hash size in globals-defines.h, recompile
and you think it will stop with the messages about that? - unless it
fills up again that is.

-----Original Message-----
From: Burton M. Strauss III [mailto:[EMAIL PROTECTED]
Sent: Wednesday, February 26, 2003 4:37 PM
To: [EMAIL PROTECTED]
Subject: RE: [Ntop] updateElementHash(): hash full!


static void updateElementHashItem(ElementHash **theHash,
                       u_short srcId, u_short dstId,
                       Counter numPkts, Counter numBytes, u_char
dataSent) {
  u_int myIdx = 0, idx = srcId % MAX_ELEMENT_HASH;
  ElementHash *hash, *prev;

  while(1) {
    if((theHash[idx] == NULL) || (theHash[idx]->id == srcId))
      break;

    idx = (idx+1) % MAX_ELEMENT_HASH;
    if(++myIdx == MAX_ELEMENT_HASH) {
      traceEvent(CONST_TRACE_WARNING, "updateElementHash(): hash
full!");
      return;
    }
  }

Basically, it's not finding an empty or matching element in the hash
table...  The hash limit is set in globals-defines.h:

/*
 * Size of the AS and VLAN hashes, created by allocateElementHash() and
dumped by
 * dumpElementHash().
 */
#define MAX_ELEMENT_HASH                    4096


I know of no reason you couldn't make it larger - it's a memory trade
off. If it allocates anything, it allocates that many entries

void allocateElementHash(int deviceId, u_short hashType) {
  int memLen = sizeof(ElementHash*)*MAX_ELEMENT_HASH;

  switch(hashType) {
  case 0: /* AS */
    if(myGlobals.device[deviceId].asHash == NULL) {
      myGlobals.device[deviceId].asHash = (ElementHash**)malloc(memLen);
      memset(myGlobals.device[deviceId].asHash, 0, memLen);
    }
    break;
  case 1: /* VLAN */
    if(myGlobals.device[deviceId].vlanHash == NULL) {
      myGlobals.device[deviceId].vlanHash =
(ElementHash**)malloc(memLen);
      memset(myGlobals.device[deviceId].vlanHash, 0, memLen);
    }
    break;
  }
}

(All code from util.c)


-----Burton

   |\      _,,,---,,_       "If I purr, you must have fed me.
   /,`.-'`'    -.  ;-;;,_    If the network purrs, it must
 <|,4-  ) )-,_..;\ (  `'-'   be ntop!"   -----Pixel
   '---''(_/--'  `-'\_)




-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of Tony
Hernandez
Sent: Wednesday, February 26, 2003 2:38 PM
To: [EMAIL PROTECTED]
Subject: [Ntop] updateElementHash(): hash full!


Hey all, I am running Ntop v2.1.55 ( latest snap wont compile ) I get
these messages in the log:

Feb 26 15:22:06 graback ntop[209]: updateElementHash(): hash full! Feb
26 15:22:30 graback last message repeated 1484 times Feb 26 15:22:46
graback last message repeated 1100 times Feb 26 15:22:46 graback
ntop[209]: WARNING: accessMutex() call with a self-LOCKED mutex [from
209 at pbuf.c:551 updateInterfacePorts]

As you can see, they happen a LOT.

I just got netflows running on it, and sometimes it just bombs without a
trace, not leaving much in the line of *why it crashed in the logs not
really much debug. Seems to run ok when its not listening for netflows
tho. anyone have a clue about this or experiences ?

This is all the logs say when it dumps -

Feb 26 15:27:26 graback /kernel: pid 209 (ntop), uid 65534: exited on
signal 11 Feb 26 15:27:26 graback /kernel: dc0: promiscuous mode
disabled




--
Tony Hernandez (53:61:72:67:65:6E:74)
Computer Programmer Analyst
Division of Housing
University of Florida
"All your base are belong to us!" - Zero Wing
_______________________________________________
Ntop mailing list
[EMAIL PROTECTED]
http://listgateway.unipi.it/mailman/listinfo/ntop

_______________________________________________
Ntop mailing list
[EMAIL PROTECTED]
http://listgateway.unipi.it/mailman/listinfo/ntop
_______________________________________________
Ntop mailing list
[EMAIL PROTECTED]
http://listgateway.unipi.it/mailman/listinfo/ntop

_______________________________________________
Ntop mailing list
[EMAIL PROTECTED]
http://listgateway.unipi.it/mailman/listinfo/ntop
_______________________________________________
Ntop mailing list
[EMAIL PROTECTED]
http://listgateway.unipi.it/mailman/listinfo/ntop
_______________________________________________
Ntop mailing list
[EMAIL PROTECTED]
http://listgateway.unipi.it/mailman/listinfo/ntop

_______________________________________________
Ntop mailing list
[EMAIL PROTECTED]
http://listgateway.unipi.it/mailman/listinfo/ntop

Reply via email to