View Issue Details

IDProjectCategoryView StatusLast Update
0012992CentOS-7-OTHERpublic2017-03-22 07:45
Reporterhearnsj 
PrioritynormalSeverityminorReproducibilityalways
Status newResolutionopen 
Product Version7.3.1611 
Target VersionFixed in Version 
Summary0012992: irqbalance update - now affinity_hint and banned cpus confict
DescriptionI updated some systems in an HPC cluster on Sunday with:
irqbalance-1.0.7-6.el7.x86_64

Systems are now at CentOS Linux release 7.3.1611 (Core)
Kernel 3.10.0-514.10.2.el7.x86_64
When irqbalance is tarted the system repeatedly logs:
 /usr/sbin/irqbalance: irq NN affinity_hint and banned cpus confict

Where NN is a different number

Is this a known issue?
Additional InformationThe /etc/sysconfig/irqbalance file contains:

# irqbalance is a daemon process that distributes interrupts across
# CPUS on SMP systems. The default is to rebalance once every 10
# seconds. This is the environment file that is specified to systemd via the
# EnvironmentFile key in the service unit file (or via whatever method the init
# system you're using has.
#
# ONESHOT=yes
# after starting, wait for a minute, then look at the interrupt
# load and balance it once; after balancing exit and do not change
# it again.
#IRQBALANCE_ONESHOT=

#
# IRQBALANCE_BANNED_CPUS
# 64 bit bitmask which allows you to indicate which cpu's should
# be skipped when reblancing irqs. Cpu numbers which have their
# corresponding bits set to one in this mask will not have any
# irq's assigned to them on rebalance
#
#IRQBALANCE_BANNED_CPUS=

#
# IRQBALANCE_ARGS
# append any args here to the irqbalance daemon as documented in the man page
#
#IRQBALANCE_ARGS=


IRQBALANCE_ARGS=--hintpolicy=exact
TagsNo tags attached.
abrt_hash
URL

Activities

pkenyon

pkenyon

2017-03-21 15:45

reporter   ~0028904

This is probably a duplicate of https://bugs.centos.org/view.php?id=12794.

Are you using Intel Omni-Path? I have this problem as well, on an OPA cluster.
hearnsj

hearnsj

2017-03-22 07:45

reporter   ~0028915

Pkenyon - thankyou. Yes I am using Omnipath - and I thought I did say that.
Of course I omitted that, sorry.
I think the issue is the recommendation with Omnipath to set hintpolicy=exact

I see it is not yet fixed in CentOS - I will downgrade the RPM maybe
and wait for the fix!

If the CentOS maintainer sees this I will be very happy to test on our system.

How can I change the Category of this bug to 'irqbalance' please?

Issue History

Date Modified Username Field Change
2017-03-20 11:41 hearnsj New Issue
2017-03-21 15:45 pkenyon Note Added: 0028904
2017-03-22 07:45 hearnsj Note Added: 0028915