Print | Rate this content

Advisory: (Revised) HPE Fibre Channel Host Bus Adapters - Task Hangs During Fibre Channel Storage Array Port/Node Resets When Using A QLogic-Based FC Adapter Driver Present In HPE Service Pack For ProLiant 2017.07.01 And Later

SUPPORT COMMUNICATION - CUSTOMER ADVISORY

Document ID: a00041102en_us

Version: 2

Advisory: (Revised) HPE Fibre Channel Host Bus Adapters - Task Hangs During Fibre Channel Storage Array Port/Node Resets When Using A QLogic-Based FC Adapter Driver Present In HPE Service Pack For ProLiant 2017.07.01 And Later
NOTICE: The information in this document, including products and software versions, is current as of the Release Date. This document is subject to change without notice.

Release Date: 2018-03-30

Last Updated: 2018-03-30


DESCRIPTION

Document Version
Release Date
Details
2
30 March 2018
Updated driver versions in the Resolution section
1
5 February 2018
Original Document Release

An issue has been found where Linux systems, connected to fibre channel storage arrays via a QLogic-based fibre channel, may experience: hung tasks, an NMI, and possibly go into a kernel panic when an array port or array controller unexpectedly resets multiple times. The issue is only seen when heavy I/O traffic is running to the array, and repeated port resets may be necessary to cause this issue.

The following kernel messages will likely be present in the operating system logs:

kernel: Call Trace:
kernel: [<ffffffff816a94c9>] schedule+0x29/0x70
kernel: [<ffffffff816a6fd9>] schedule_timeout+0x239/0x2c0
kernel: [<ffffffff812f71d3>] ? __blk_run_queue+0x33/0x40
kernel: [<ffffffff812f728a>] ? queue_unplugged+0x2a/0xa0
kernel: [<ffffffff810e939c>] ? ktime_get_ts64+0x4c/0xf0
kernel: [<ffffffff816a8b4d>] io_schedule_timeout+0xad/0x130
kernel: [<ffffffff816a8be8>] io_schedule+0x18/0x20
kernel: [<ffffffff812421d5>] do_blockdev_direct_IO+0x1c45/0x2020
kernel: [<ffffffff8123cb30>] ? I_BDEV+0x10/0x10
kernel: [<ffffffff81242605>] __blockdev_direct_IO+0x55/0x60
kernel: [<ffffffff8123cb30>] ? I_BDEV+0x10/0x10
kernel: [<ffffffff8123d420>] blkdev_direct_IO+0x60/0xa0
kernel: [<ffffffff8123cb30>] ? I_BDEV+0x10/0x10
kernel: [<ffffffff811844fc>] generic_file_aio_read+0x70c/0x790
kernel: [<ffffffff812fab24>] ? blk_finish_plug+0x14/0x40
kernel: [<ffffffff8123d97c>] blkdev_aio_read+0x4c/0x70
kernel: [<ffffffff8120019d>] do_sync_read+0x8d/0xd0
kernel: [<ffffffff81200b9c>] vfs_read+0x9c/0x170
kernel: [<ffffffff81201a5f>] SyS_read+0x7f/0xe0
kernel: [<ffffffff816b4fc9>] system_call_fastpath+0x16/0x1b

This issue has been root caused to a driver architecture change introduced in HPE Service Pack for ProLiant version 2017.07.01. This change would sometimes cause a hang in the driver when devices are added or removed from the FC fabric under high load.

SCOPE

This issue occurs with the following adapters:

  • HPE 81Q Fibre Channel Host Bus Adapter
  • HPE 82Q Fibre Channel Host Bus Adapters
  • HPE SN1000Q Fibre Channel Host Bus Adapters
  • HPE SN1100Q Fibre Channel Host Bus Adapters
  • HPE SN1600Q Fibre Channel Host Bus Adapters
  • HPE QMH2572 Fibre Channel Host Bus Adapter for c-Class BladeSystem
  • HPE QMH2672 Fibre Channel Host Bus Adapter for c-Class BladeSystem
  • HPE Synergy 3830C Fibre Channel Host Bus Adapter

This issue affects all FC drivers that are present in the initial releases of the following HPE Service Pack for ProLiant 2017.07.01, 2017.07.02, 2017.10.1, and 2018.02.01. The driver versions present in those service packs are below:

  • RHEL 6: 8.07.00.50.06.0_k3, 8.07.00.50.06.0_k4, 8.07.00.50.06.0_k6
  • RHEL 7: 8.07.00.50.07.0_k3, 8.07.00.50.07.0_k4, 8.07.00.50.07.0_k6
  • SLES 11: 8.07.00.50.11.3_k3, 8.07.00.50.11.3_k3 (b), 8.07.00.50.11.3_k5
  • SLES 12: 8.07.00.50.12.0_k3, 8.07.00.50.12.2_k4, 8.07.00.50.12.2_k5

RESOLUTION

If performing HPE Storage array firmware upgrades (or other maintenance actions that involve array port resets), HPE recommends to disconnect Linux hosts, using the affected drivers from the array while this maintenance is being performed. If disconnecting the host is not possible, then HPE recommends to attempt to minimize the amount of traffic being generated by these hosts during the maintenance operations, as well as minimize the number of array port resets that occur in quick succession during the maintenance.

A hotfix will be released with updated OS drivers to address this issue. The driver versions that contain the fix are as follows:

  • RHEL 6: 8.07.00.50.06.0_k7
  • RHEL 7: 8.07.00.50.07.0_k7(b)
  • SLES 11: 8.07.00.50.11.3_k6(b) where "#" corresponds to the SLES service pack number
  • SLES 12: 8.07.00.50.12.3_k7(b) where "#" corresponds to the SLES service pack number

Once it is released, HPE recommends upgrading to these versions of drivers as soon as possible.




NOTE: Products sold prior to the November 1, 2015 separation of Hewlett-Packard Company into Hewlett Packard Enterprise Company and HP Inc. may have older product names and model numbers that differ from current models.




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


Hardware Platforms Affected: HPE Synergy 3830C 16Gb Fibre Channel Host Bus Adapter, HPE StoreFabric SN1100Q 16Gb Single Port Fibre Channel Host Bus Adapter, HPE StoreFabric SN1100Q 16Gb Dual Port Fibre Channel Host Bus Adapter, HPE StoreFabric SN1600Q 32Gb Single Port Fibre Channel Host Bus Adapter, HPE StoreFabric SN1600Q 32Gb Dual Port Fibre Channel Host Bus Adapter, HPE 81Q 8Gb 1-port PCIe Fibre Channel Host Bus Adapter, HPE 82Q 8Gb 2-port PCIe Fibre Channel Host Bus Adapter, HPE QMH2572 8Gb Fibre Channel Host Bus Adapter, HPE StoreFabric SN1000Q 16GB 2-port PCIe Fibre Channel Host Bus Adapter, HPE QMH2672 16Gb Fibre Channel Host Bus Adapter
Operating Systems Affected: Red Hat Enterprise Linux (Itanium), SUSE Linux Enterprise Server 11 (x86-64), Red Hat Enterprise Linux 7 (AMD64/EM64T), SUSE Linux Enterprise Server 12 (AMD64/EM64T)
Software Affected: Not Applicable
Support Communication Cross Reference ID: SIK2935
©Copyright 2018 Hewlett Packard Enterprise Development LP
Hewlett Packard Enterprise Development shall not be liable for technical or editorial errors or omissions contained herein. The information provided is provided "as is" without warranty of any kind. To the extent permitted by law, neither HPE nor its affiliates, subcontractors or suppliers will be liable for incidental, special or consequential damages including downtime cost; lost profits; damages relating to the procurement of substitute products or services; or damages for loss of data, or software restoration. The information in this document is subject to change without notice. Hewlett Packard Enterprise Development and the names of Hewlett Packard Enterprise Development products referenced herein are trademarks of Hewlett Packard Enterprise Development in the United States and other countries. Other product and company names mentioned herein may be trademarks of their respective owners.

Provide feedback

Please rate the information on this page to help us improve our content. Thank you!