Unable to redirect nvme disk interrupts to different core

Solution Verified - Updated -

Issue

  • We have a bare metal server with 4 nvme disks. We want to redirect the nvme disk interrupts to only specific set of CPUs, so that the critical application threads running on other CPUs are not interrupted by the disk IRQs.

    We have set isolcpus=19-35 kernel option to isolate the IRQs on specific set of CPUs. But even with this option, we still see nvme interrupts occurring on isolated CPUs:

    $ tuna show_irqs
[...]
143 nvme0q1                18
144 nvme0q2                19
145 nvme0q3                20
146 nvme0q4                21
147 nvme0q5                22
148 nvme0q6                23
149 nvme0q7                24
150 nvme0q8                25
151 nvme0q9                26
152 nvme0q10               27
153 nvme0q11               28
154 nvme0q12               29
155 nvme0q13               30

    Is it possible to completely isolate the CPUs from nvme disk interrupts, so we can dedicate those CPUs only for application processes?

Environment

  • Red Hat Enterprise Linux 8
  • NVMe devices

Subscriber exclusive content

A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.

Current Customers and Partners

Log in for full access

Log In

New to Red Hat?

Learn more about Red Hat subscriptions

Using a Red Hat product through a public cloud?

How to access this content