Support Questions
Find answers, ask questions, and share your expertise
Announcements
Alert: Welcome to the Unified Cloudera Community. Former HCC members be sure to read and learn how to activate your account here.

Network packets were dropping on few datanodes

Network packets were dropping on few datanodes

Hi,

we are using this environment since 1.5 yr and over the period Network packets were dropped to 818733, but observed that dropping rate is very minimal on daily basis.

I want to check how this is going to impact data loading in data nodes?

please suggest me if we have any solution to re-set this value to ZERO.

[root@abc ~]# ifconfig 
eth1        Link encap:Ethernet  HWaddr 00:25:B5:00:00:31 
          inet addr:00.00.00.00  Bcast:00.00.00.00  Mask:255.255.255.0 
          inet6 addr: fe80::225:b5ff:fe00:31/64 Scope:Link 
          UP BROADCAST RUNNING MULTICAST  MTU:9000  Metric:1 
          RX packets:7929368254 errors:0 dropped:818733 overruns:0 frame:0 
          TX packets:1557926025 errors:0 dropped:0 overruns:0 carrier:0 
          collisions:0 txqueuelen:1000 
          RX bytes:14650402362276 (13.3 TiB)  TX bytes:2646572126036 (2.4 TiB) 
lo            Link encap:Local Loopback
          inet addr:127.0.0.1  Mask:255.0.0.0 
          inet6 addr: ::1/128 Scope:Host 
          UP LOOPBACK RUNNING  MTU:65536  Metric:1 
          RX packets:1096399714 errors:0 dropped:0 overruns:0 frame:0 
          TX packets:1096399714 errors:0 dropped:0 overruns:0 carrier:0 
          collisions:0 txqueuelen:0
          RX bytes:1620626363855 (1.4 TiB) 
 TX bytes:1620626363855 (1.4 TiB)
[root@abc ~]#
3 REPLIES 3

Re: Network packets were dropping on few datanodes

Flows that experience packet drops reduce their data rate and bandwidth consumed, with some flows getting very little if any bandwidth, while other flows that experience fewer packet drops receive more than their fair share of the available network bandwidth. This wide range of bandwidth as seen by different flows can lead to highly variable completion times for distributed applications that depend on the all flows or queries to complete. To reduce this effect, ‘’deep buffering’’ is preferable to low-latency in switches. Enabling Jumbo Frames across the cluster improves bandwidth through better checksums and possibly may also provide packet integrity. For more information, please reference our Cluster Planning Guide or contact your Hortonworks field representative.

Re: Network packets were dropping on few datanodes

@Divakar Annapureddy: Was this answer helpful? If so, could you please accept it through the HCC? If not, is there some follow-up I could help you with?

Highlighted

Re: Network packets were dropping on few datanodes

@Divakar Annapureddy.

If you are just looking to reset the counter and see differentials, then below link may help.

http://www.ducea.com/2006/09/08/resetting-ifconfig-counters/

Hope this helps