0

We have hadoop cluster , hadoop version - 2.6.X & ambari version - 2.5.0.X

Should avahi-daemon service be disabled or enabled on hadoop cluster machines as ( master / workers / kafka )?

note - we want to disable this service especially on the data-node machines ( workers ), because we think this service could be the reason of OS restart,

for now this service is enable and running on all our machines in the hadoop cluster

reference - https://www.systutorials.com/docs/linux/man/8-avahi-daemon/

service avahi-daemon status
Redirecting to /bin/systemctl status  avahi-daemon.service
● avahi-daemon.service - Avahi mDNS/DNS-SD Stack
   Loaded: loaded (/usr/lib/systemd/system/avahi-daemon.service; enabled; vendor preset: enabled)
   Active: active (running) since Wed 2018-04-25 18:52:11 UTC; 17min ago

Note - I want to add that in /var/log/message file we see a million messages as

Apr 25 00:08:32 worker02 avahi-daemon[953]: Invalid response packet from host 91.61.42.241.
Apr 25 00:08:32 worker02 avahi-daemon[953]: Invalid response packet from host 91.61.60.61.
Apr 25 00:08:32 worker02 avahi-daemon[953]: Invalid response packet from host 91.61.35.160.
Apr 25 00:08:32 worker02 avahi-daemon[953]: Invalid response packet from host 91.61.22.128.
Apr 25 00:08:32 worker02 avahi-daemon[953]: Invalid response packet from host 91.61.56.231.
Phantômaxx
  • 37,901
  • 21
  • 84
  • 115
enodmilvado
  • 443
  • 1
  • 9
  • 20

1 Answers1

1

It should not hurt to disable them. It's not used by any Hadoop or Kafka protocol that I know of.

Its used for DNS, but hopefully you have some other DNS setup if you plan on turning it off.

You should rather try to find what hosts are in that 91.61.x.x range that are trying to communicate with your cluster though.

OneCricketeer
  • 179,855
  • 19
  • 132
  • 245
  • one of the reasons that I want to shout it off because possible of issue as host restart that come suddenly we see in the messages file a lot of these lines and maybe this effected badly the machine , what you think about ? – enodmilvado Apr 26 '18 at 04:53
  • 1
    I mean, sure, excessive logging can flood your disks and they run out of space, causing the machine to crash... That is one possibility, but that isn't the services fault, that is the external hosts trying to communicate with the service – OneCricketeer Apr 26 '18 at 04:56