Operational Defect Database
BugZero found this defect 1362 days ago.
Data sources
All data on this page is proprietary to BugZero® or gathered from public sources
Hewlett Packard Enterprise | a00104372en_us
Advisory: (Revision) HPE 16GB/32GB FC QLogic Host Bus Adapters, HPE 3PAR/HPE Primera - FC Adapters In An ESXi Environment May Encounter Failure Symptoms Under Heavy Load When Connected To Storage Arrays Implementing Different DIF Technologies
Last update date:
1/31/2024
Affected products:
HPE Primera Controller Nodes
HPE QMH2672 16Gb Fibre Channel Host Bus Adapter for BladeSystem c-Class
HPE SN1000Q 16Gb PCIe Host Bus Adapter
HPE SN1100Q 16Gb Host Bus Adapter
HPE SN1600Q 32Gb Host Bus Adapter
HPE SN1610Q 32Gb Fibre Channel Host Bus Adapter
HPE Synergy 3830C 16Gb Fibre Channel Host Bus Adapter
HPE 3PAR StoreServ 20000 Storage
HPE 3PAR StoreServ 8000 Storage
Affected releases:
No affected releases provided.
Fixed releases:
No fixed releases provided.
Description:
Info
Document Version Release Date Details 2 27 August 2020 Updated document content to help clear any ambiguity as to whether this is the adapter on the host or the array that is affected 1 3 August 2020 Original Document Release HPE has identified a compatibility issue in the ESXi host server drivers related to DIF functionality. This issue is caused by the ESXi driver sending Persistent Checksum info to an array LUN which does not support Persistent Checksum, in a mixed storage environment. Customers who have ESXi host server(s) connected to multiple target LUNs from multiple arrays may encounter loss of connectivity to the target under heavy load if the arrays implement different Data Integrity Field (DIF) checksum technologies. For example, customers who have a host connected to an HPE 3PAR/Primera Storage array using the Persistent Checksum feature may encounter this issue under heavy load if they also have another HPE 3PAR/Primera array connected to that host which does not support, has not enabled, or is configured not to use the feature. Similarly, customers who have a host connected to an HPE 3PAR/Primera array with Persistent Checksum enabled and another array with host-based T10-DIF or no DIF technologies enabled under heavy load may encounter this issue. This lost connectivity may be exhibited as Data underrun errors, lost data-store connections, DIF errors, or increased latency recorded in the VMkernel logs. NOTE : This lost connectivity may also result in a host system hang. Following is sample output of what may be seen in the VMkernel logs: 2020-02-27T09:08:20.666Z cpu37:2103083)qlnativefc:vmhba2(d8:0.1): iocb(s) 0x430bd32eff00 Returned STATUS. 2020-02-27T09:08:20.666Z cpu37:2103083)qlnativefc: vmhba2(d8:0.1): DIF ERROR in cmd: 0x28 Type=0x0 lba=0xad00 actRefTag=0x0, expRefTag=0xad00, actAppTag=0x0,expAppTag=0x0, actGuard=0x0, expGuard=0xf2a2. 2020-02-19T03:28:41.546Z cpu2:65831)qlnativefc: vmhba1(12:0.0): (29:14): Mid-layer underflow detected (8 of 512 bytes).
Scope
This advisory applies to HPE 3PAR/Primera storage arrays and HPE Fibre Channel Host Bus Adapters with ESXi environments running the following 16/32GB FC device drivers: 3.1.29.0 (ESXi 6.7) 3.1.31.0 (ESXi 6.7) 2.1.94.0 (ESXi 6.5) 2.1.96.0 (ESXi 6.5) The following operating systems are affected: VMware vSphere 6.5 VMware vSphere 6.7
Resolution
HPE recommends that you update the ESXi host server to the latest driver revisions for the 16Gb and 32Gb adapters: ESXi6.5 - 2.1.101.0 ESXi6.7 - 3.1.36.0 Workaround If the customer is unwilling to update the host server driver at the current time, then they may avoid this issue by either: Disable Persistent Checksum on the host, use the following ESXi CLI command: esxcli system module parameters set -p ql2xt10difvendor=0 -m qlnativefc NOTE : After running the above command, the ESXi hypervisor needs to be rebooted. OR Disable support for Persistent Checksum on the exported Virtual Volumes (VV) on the supported HPE 3PAR/Primera storage arrays with the following HPE 3PAR CLI command: setvv -pol no_host_dif <VV name> NOTE : This command on the HPE 3PAR takes effect immediately, and it is recommended that any host which has this LUN exported be shut down prior to making the change at the virtual volume level. RECEIVE PROACTIVE UPDATES : Receive support alerts (such as Customer Advisories), as well as updates on drivers, software, firmware, and customer replaceable components, proactively in your e-mail through HPE Subscriber's Choice. Sign up for Subscriber's Choice at the following URL: Proactive Updates Subscription Form.
