NAS access slow or hung caused by CRC errors on network port

Applies to

• ONTAP 9
• NFS
• CIFS

Issue

• Observed slow data transfer and high read latency every time application is run
• QOS statistics volume latency command slows little latency
• NFS client report below error and hang:

kernel: nfs: server 10.xx.xx.xx not responding, still trying 

• Packet trace (from client or storage) may show retransmit and out of order packets
• In CIFS the shares experience slowness in reads and writes
• Check for network errors
  • From the filer you can check ports with the below command
    • >>: node run -node node1 -command "ifstat-a" for the affected ports
  • Engage network team to review connected switch ports as well

•  cluster::> system node run -node "nodename" -command ifstat -a
  
  -- interface  e4a  (2 days, 10 hours, 22 minutes, 35 seconds) --
   
  RECEIVE
   Total frames:     5161m | Frames/second:   24558  | Total bytes:      5152g
   Bytes/second:    24516k | Total errors:    35927  | Errors/minute:      10 
   Total discards:      0  | Discards/minute:     0  | Multi/broadcast: 26826k
   Non-primary u/c:     0  | CRC errors:      35927  | Runt frames:         0 
   Fragment:            0  | Long frames:         0  | Jabber:              0 
   Length errors:       0  | No buffer:           0  | Xon:                 0 
   Xoff:                0  | Pause:               0  | Jumbo:             218k
   Noproto:             0  | Error symbol:        0  | Illegal symbol:      0 
   Bus overruns:        0  | Queue drops:         0  | LRO segments:     2060m
   LRO bytes:        4909g | LRO6 segments:       0  | LRO6 bytes:          0 
   Bad UDP cksum:       0  | Bad UDP6 cksum:      0  | Bad TCP cksum:       0 
   Bad TCP6 cksum:      0  | Mcast v6 solicit:    0  | Lagg errors:         0 
   Lacp errors:         0  | Lacp PDU errors:     0 
  

• CRC errors indicate a receive side data issue
• CRCs result in packet loss due to re-transmissions and out-of-order packets received.