NAME    [Toc]    [Back]

       niffd - Network Interface Failure Finder daemon

SYNOPSIS    [Toc]    [Back]

       /usr/sbin/niffd  [-d  debug-level]  [-f] [-l logfile] [-a]
       [-n] [-p] [-r] [-t tries] [-w time]

OPTIONS    [Toc]    [Back]

       Sets the debug level as follows:  Logs  NIFF  errors  only
       Logs  NIFF events and errors Logs NIFF caches, events, and
       errors

              By   default,   all   messages   are   written   to
              /var/tmp/niffd.log,  but can be overridden with the
              -l option.  Warnings and errors are still  sent  to
              syslog(3) as well.  Foreground mode.  In this mode,
              niffd does not run as a daemon.  All  messages  are
              written  to standard error (stderr), although warnings
 and errors are  still  sent  to  syslog(3)  as
              well.   Sends  debugging  information  to  logfile,
              instead of the default location /var/tmp/niffd.log.
              Disables  Address Resolution Protocol (ARP) request
              messages.  Disables ICMP ECHO_REQUEST (ping) subnet
              broadcast messages.  Disables all ICMP ECHO_REQUEST
              messages.  Disables  Routing  Information  Protocol
              (RIP)  messages.  Directs niffd to generate network
              traffic tries times per event.  The  default  is  3
              tries.   Directs niffd to wait time seconds between
              traffic generation cycles.  The default is  5  seconds.

DESCRIPTION    [Toc]    [Back]

       The  Network  Interface Failure Finder daemon, niffd, is a
       traffic generator for network interfaces  that  have  been
       classified  inactive  by the kernel Traffic Monitor Thread
       (TMT). The purpose of niffd is to get the interface packet
       counters  to  increment, signifying the interface is still
       alive and well.

       The niffd daemon subscribes to a class  of  events,  using
       the  Event  Management  (EVM)  connection monitoring functions,
 and waits for activity on the event connection.  If
       a  received  event requires traffic generation, niffd continues
 to generate traffic until it receives a  new  event
       that  overrides the old event, or until the maximum number
       of retries has been reached.

       Different events cause different types of network  traffic
       to  be  generated.   Each type can be selectively disabled
       with the -a, -n, -p, or -r  options.   Caution  should  be
       taken  when  disabling  a specific type of network traffic
       because it could result in no traffic being sent  at  all.
       It  is  recommended  that you leave all methods of traffic
       generation  enabled  unless  your   specific   environment
       requires that it be disabled.

       If  the  local machine generates an event, niffd generates
       network traffic based on the alert level of the  event  as
       follows:  Sends an ARP request message to various machines
       on the same subnet Sends either an ICMP ECHO_REQUEST  message
  (ping)  to  various machines on the same subnet or a
       RIP message to the subnet broadcast address Sends an  ICMP
       ECHO_REQUEST message to the subnet broadcast address

       If niffd receives an alarm for an interface that is not on
       the local machine, niffd tries to ping the suspect  interface
  at  all  alert  levels. For example, another machine
       detected a problem and it was forwarded to  niffd  through
       EVM.

       By default, niffd does not start during system boot.  However,
 if you specify the NIFFD and  NIFFC_FLAGS  variables
       in  the  /etc/rc.config  file,  niffd starts during system
       boot. You use rcmgr to specify these  variables.  See  the
       rcmgr(8) reference page for further information.

       During  system  boot,  init  reads  the /sbin/init.d/niffd
       file. First the niffconfig command runs with the arguments
       specified in the NIFFC_FLAGS variable. Then, niffd starts.

EXAMPLES    [Toc]    [Back]

       To enable niffd startup and specify that it monitor the tu
       interface  with  the  default  values,  enter: # rcmgr set
       NIFFD "YES" # rcmgr set NIFFC_FLAGS "-a tu0"

              See the niffconfig(8) reference  page  for  further
              information.   If  you  are running in a TruCluster
              environment, the previous  step  will  have  to  be
              repeated  for all cluster members as actual network
              interface configurations on each member  may  vary.
              See rcmgr(8) for information on how to do this.  To
              start  niffd  with  additional  options,  set   the
              NIFFD_FLAGS  variable accordingly.  For example, to
              start niffd with ICMP ECHO_REQUEST subnet broadcast
              messages  disabled,  enter: # rcmgr set NIFFD_FLAGS
              "-n" To prevent niffd from  starting  automatically
              during  system boot, enter the following command: #
              rcmgr set NIFFD "NO"

FILES    [Toc]    [Back]

       Specifies the command path Default log file when debugging
       is enabled

SEE ALSO    [Toc]    [Back]

       Commands: evmd(8), niffconfig(8), ping(8), rcmgr(8)

       Protocols: arp(7)

       Information: EVM(5), nifftmt(7)



                                                         niffd(8)