Slow performance of LUNs or Namespaces on AFF
Applies to
- ONTAP 9
- ONTAP 9.6 - 9.6P7
- ONTAP 9.7 - 9.7P3
- AFF
- Environments serving any SAN protocol (iSCSI / FCP / FCoE / NVMEoF)
Issue
- An All Flash FAS (AFF) with data compaction enabled is running a SAN protocol (iSCSI, FC/FCoE, NVMEoF).
- Hosts will see intermittent access issues that eventually worsen, and cause inaccessibility or a system disruption causing a High Availability Takeover if enabled.
- Aborts and timeouts will be seen in host logs such as the following from ESXi:
EXAMPLES:
Abort: cpu15;20977090NMP: nmp_Throttlelogfordevice;3802; cmd 0x89 90x45a2b689d580, 57664670 to dev 'naa.600a0980383143c642b50365737566d' on path 'vmhba2;C0:t7;l2' failed; h;0x8 d;0x0 p;0x0 Invalid sense data; 0x0 0x0 0x0. act;EvaL
Timeout:2020-01-01t15;00;20.461z cpu26;20997240hbx: 1034; 'NetApp_AFF_Lun-1'; hb at offset 3486100 - setting pulse failed; timeout;
- System event logs shows low memory messages
::> event log show -severity DEBUG -messagename wafl.memory.*
1/01/2020 16:00:00 Cluster-02 ALERT wafl.memory.statusVeryLowMemory: WAFL is running very low on memory, with 795MB remaining.