2 linksets have IP connection excess retransmit alarms

Asset ID:	1-72-2049357.1
Update Date:	2017-06-29
Keywords:

Solution Type Problem Resolution Sure

Solution 2049357.1 : 2 linksets have IP connection excess retransmit alarms

Applies to:

Oracle Communications EAGLE (Hardware) - Version EAGLE 44.0.5 and later
Information in this document applies to any platform.

Symptoms

2 linksets show IP connection excess retransmit alarms

0536 * IP7CONN assoc1 IP Connection Excess Retransmits
0536 * IP7CONN assoc2 IP Connection Excess Retransmits

Cause

- The card 2312 is only having an excess retransmit alarm on one association out of 10, so the issue does not come from the card but from the IP path :

ASSOCIATION STATUS
  ANAME PST SST
LSNXassoc1 IS-NR ESTABLISHED
  LSNXassoc2 IS-NR ESTABLISHED
  LSNXassoc3 IS-NR ESTABLISHED
  LSNXassoc4 IS-NR ESTABLISHED
  LSNXassoc5 IS-NR ESTABLISHED
  LSNXassoc6 IS-NR ESTABLISHED
  LSNXassoc7 IS-NR ESTABLISHED
  lsn2assoc1 IS-ANR EXCESS RETRANS
  LSNXassoc8 IS-NR ESTABLISHED
  LSNXassoc9 IS-NR ESTABLISHED

- Only all associations of each linkset LSN1 and LSN2 are having this Excess retransmits alarm :

****15-08-19 16:00:07****
  9989.0536 * IP7CONN lsn1assoc1 IP Connection Excess Retransmits
  ****15-08-19 16:00:17****
  5010.0536 * IP7CONN lsn1assoc2 IP Connection Excess Retransmits
  ****15-08-19 16:04:37****
  5487.0536 * IP7CONN lsn1assoc3 IP Connection Excess Retransmits

  ****15-08-19 16:00:18****
  5011.0536 * IP7CONN lsn2assoc1 IP Connection Excess Retransmits
  ****15-08-19 16:02:22****
  5275.0536 * IP7CONN lsn2assoc2 IP Connection Excess Retransmits
  ****15-08-19 16:04:04****
  5432.0536 * IP7CONN lsn2assoc3 IP Connection Excess Retransmits

- The statistics of the SCTP layer for one of the associations show important amount of RTT too big and retransmissions, packets duplicated, GAPs :

> pass:loc=2312:cmd="sctp -a lsn2assoc1"

eaglestp 15-08-20 12:01:11 CEST EAGLE5 44.0.5-64.34.26
  Aname           Local        Local   Remote        Remote
Address        Port    Address       Port
  lsn2assoc1    10.217.192.213 2638    10.210.7.178 3638
  10.217.192.149

  Nets Data
  ...
  ip datagrams rcvd = 163021190
  ip datagrams with data chunks rcvd = 110109780
  data chunks rcvd = 163021190
  data chunks read = 163021171
  dup tsns rcvd = 19
  sacks rcvd = 92154618
  gap ack blocks rcvd = 84
  heartbeat requests rcvd = 10783078
  heartbeat acks rcvd = 25668398
  heartbeat requests sent = 25668458
  ip datagrams sent = 237851931
  ip datagrams with data chunks sent = 172645043
  data chunks sent = 180040871
  retransmit data chunks sent = 1642761
  sacks sent = 89604748
  send failed = 0
  retransmit timer count = 1333341
  consecutive retransmit timeouts = 42
  RTT between RMIN and RMAX inclusive = 56688097
  RTT greater than RMAX = 3239492
  fast retransmit count = 47
  recv timer count = 8250175
  heartbeat timer count = 25671452
  none left tosend = 0
  none left rwnd gate = 0
  none left cwnd gate = 0
  UNKNOWN = 1333363

- There is also an excess of RTT for the associations. The RTT should not exceed the Min RTO, if it happens it means there was some retransmissions :

> pass:cmd="assocrtt lsn1assoc1":loc=2312

eaglestp 15-08-27 16:02:20 CEST EAGLE5 44.0.5-64.34.26
  ASSOCRTT: Association round-trip time report (in milliseconds)
  Retransmission Configuration
  Retransmission Mode : RFC
  Minimum RTO : 200
  Maximum RTO : 500

  Normal Traffic Round-Trip Times
  Minimum round-trip time : 5
  Maximum round-trip time : 324
  Weighted Average round-trip time : 53
  Last recorded round-trip time : 19

> pass:cmd="assocrtt lsn2assoc1":loc=1211

eaglestp 15-08-27 16:02:51 CEST EAGLE5 44.0.5-64.34.26
  ASSOCRTT: Association round-trip time report (in milliseconds)
  Retransmission Configuration
  Retransmission Mode : RFC
  Minimum RTO : 200
  Maximum RTO : 500

  Normal Traffic Round-Trip Times
  Minimum round-trip time : 5
  Maximum round-trip time : 235
  Weighted Average round-trip time : 82
  Last recorded round-trip time : 15

- One of the traceroute shows a timed out

> pass:cmd="traceroute 10.210.7.222":loc=1211

eaglestp 15-08-27 16:07:39 CEST EAGLE5 44.0.5-64.34.26
  Traceroute to farend01.fe01 (10.210.7.222),
  10 hops max, 100 byte packets
  1 0ms 0ms 0ms (10.217.192.193)
  2 * * * Request timed out
  3 5ms 5ms 5ms (10.210.7.65)
  4 5ms 5ms 5ms farend01.fe01
  (10.210.7.222)

Solution

The preferred way of eliminating retransmissions is to activate IP quality of service in the network and give SCTP the highest priority.
The quality of the IP network has to be checked.

Attachments

This solution has no attachment