Message from Log

[stevefxp1]

Hello all,

I am seeing some warning and error messages in my TuneD log. What do these mean and should I be concerned:

1) WARNING tuned.plugins.plugin_cpu: unable to run x86_energy_perf_policy tool, ignoring CPU energy performance bias, is the tool installed?
2) Do you not support NVMe controllers in Ubuntu: INFO tuned.plugins.base: instance disk: assigning devices sda, loop7, loop3, loop2, sdc, loop1, loop4, dm-0, loop9, dm-3, dm-1, loop8, loop6, dm-2, loop0, sdb, loop5
3) ERROR tuned.plugins.plugin_sysctl: Failed to read sysctl parameter 'kernel.sched_min_granularity_ns', the parameter does not exist
4) ERROR tuned.plugins.plugin_sysctl: Failed to read sysctl parameter 'kernel.sched_wakeup_granularity_ns', the parameter does not exist

Thanks,
Steve

[yarda]

1. WARNING tuned.plugins.plugin_cpu: unable to run x86_energy_perf_policy tool, ignoring CPU energy performance bias, is the tool installed?

Is the ' x86_energy_perf_policy' tool installed? It's usually in the kernel-tools package and is built from the kernel sources.

2. Do you not support NVMe controllers in Ubuntu: INFO tuned.plugins.base: instance disk: assigning devices sda, loop7, loop3, loop2, sdc, loop1, loop4, dm-0, loop9, dm-3, dm-1, loop8, loop6, dm-2, loop0, sdb, loop5

NVMe should be supported, could you provide output of:

$ udevadm info /dev/YOUR_NVME_DEVICE

3. ERROR tuned.plugins.plugin_sysctl: Failed to read sysctl parameter 'kernel.sched_min_granularity_ns', the parameter does not exist
4. ERROR tuned.plugins.plugin_sysctl: Failed to read sysctl parameter 'kernel.sched_wakeup_granularity_ns', the parameter does not exist

Which TuneD profile triggers it? Recent upstream TuneD profiles use scheduler plugin, not sysctl plugin and it shouldn't trigger this error with recent kernels, e.g.:

[scheduler]
sched_latency_ns=60000000
sched_migration_cost_ns=500000
sched_min_granularity_ns=15000000
sched_wakeup_granularity_ns=2000000

[stevefxp1]

1) Sorry, which tools need to be installed?? I am on Ubuntu 22.04.4. 2) Attached is the output of the command you asked for. 3) I am running the virtual-guest profile. I have not tried anything but this one. I am running HWE kernel 6.5 on Ubuntu 22.04.4.

On Mon, Jul 1, 2024 at 4:48 PM Jaroslav Škarvada ***@***.***> wrote: 1. WARNING tuned.plugins.plugin_cpu: unable to run x86_energy_perf_policy tool, ignoring CPU energy performance bias, is the tool installed? Is the ' x86_energy_perf_policy' tool installed? It's usually in the kernel-tools package and is built from the kernel sources. 2. Do you not support NVMe controllers in Ubuntu: INFO tuned.plugins.base: instance disk: assigning devices sda, loop7, loop3, loop2, sdc, loop1, loop4, dm-0, loop9, dm-3, dm-1, loop8, loop6, dm-2, loop0, sdb, loop5 NVMe should be supported, could you provide output of: $ udevadm info /dev/YOUR_NVME_DEVICE 3. ERROR tuned.plugins.plugin_sysctl: Failed to read sysctl parameter 'kernel.sched_min_granularity_ns', the parameter does not exist 4. ERROR tuned.plugins.plugin_sysctl: Failed to read sysctl parameter 'kernel.sched_wakeup_granularity_ns', the parameter does not exist Which TuneD profile triggers it? Recent upstream TuneD profiles use scheduler plugin, not sysctl plugin and it shouldn't trigger this error with recent kernels, e.g.: [scheduler] sched_latency_ns=60000000 sched_migration_cost_ns=500000 sched_min_granularity_ns=15000000 sched_wakeup_granularity_ns=2000000 — Reply to this email directly, view it on GitHub <#655 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQ5ST4TBFTVCX2GFXGCAM4LZKG6CXAVCNFSM6AAAAABKGABLPWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMBQHE4TGMBTGY> . You are receiving this because you authored the thread.Message ID: ***@***.***>

P: /devices/pci0000:00/0000:00:18.0/0000:1b:00.0/nvme/nvme0/nvme0n1 N: nvme0n1 L: 0 S: disk/by-id/nvme-VMware_Virtual_NVMe_Disk_VMware_NVME_0002_1 S: disk/by-id/nvme-eui.4a321c9edbcaecbd000c2963a536400b S: disk/by-path/pci-0000:1b:00.0-nvme-1 S: disk/by-id/nvme-VMware_Virtual_NVMe_Disk_VMware_NVME_0002 E: DEVPATH=/devices/pci0000:00/0000:00:18.0/0000:1b:00.0/nvme/nvme0/nvme0n1 E: DEVNAME=/dev/nvme0n1 E: DEVTYPE=disk E: DISKSEQ=23 E: MAJOR=259 E: MINOR=0 E: SUBSYSTEM=block E: USEC_INITIALIZED=2366763 E: MPATH_SBIN_PATH=/sbin E: DM_MULTIPATH_DEVICE_PATH=0 E: ID_SERIAL_SHORT=VMware NVME_0002 E: ID_WWN=eui.4a321c9edbcaecbd000c2963a536400b E: ID_MODEL=VMware Virtual NVMe Disk E: ID_REVISION=1.3 E: ID_NSID=1 E: ID_SERIAL=VMware_Virtual_NVMe_Disk_VMware_NVME_0002_1 E: ID_PATH=pci-0000:1b:00.0-nvme-1 E: ID_PATH_TAG=pci-0000_1b_00_0-nvme-1 E: ID_PART_TABLE_UUID=2a1db2c1 E: ID_PART_TABLE_TYPE=dos E: DEVLINKS=/dev/disk/by-id/nvme-VMware_Virtual_NVMe_Disk_VMware_NVME_0002_1 /dev/disk/by-id/nvme-eui.4a321c9edbcaecbd000c2963a536400b /dev/disk/by-path/pci-0000:1b:00.0-nvme-1 /dev/disk/by-id/nvme-VMware_Virtual_NVMe_Disk_VMware_NVME_0002 E: TAGS=:systemd: E: CURRENT_TAGS=:systemd:

[yarda]

1. Sorry, which tools need to be installed?? I am on Ubuntu 22.04.4.

Probably linux-tools-common as from: https://packages.ubuntu.com/jammy/all/linux-tools-common/filelist

2. Attached is the output of the command you asked for.

LGTM, maybe your TuneD version is too old. NVMe support was added into tuned-2.18.0. If you run Ubuntu 22.04.4 the version in the repo seems to be quite ancient tuned-2.15.0 as from: https://packages.ubuntu.com/jammy/all/tuned
The latest TuneD version is tuned-2.23.0 (https://github.com/redhat-performance/tuned/releases/). This would also explain why you trigger the sysctl errors.

[stevefxp1]

Ok so is there an installer that I can use to get this up to date or are there instructions on how to compile this for Ubuntu 22.04? I do not see anything related to Ubuntu on your Git page.

…

On Wed, Jul 3, 2024 at 3:41 AM Jaroslav Škarvada ***@***.***> wrote: 1. Sorry, which tools need to be installed?? I am on Ubuntu 22.04.4. Probably linux-tools-common as from: https://packages.ubuntu.com/jammy/all/linux-tools-common/filelist 2. Attached is the output of the command you asked for. LGTM, maybe your TuneD version is too old. NVMe support was added into tuned-2.18.0. If you run Ubuntu 22.04.4 the version in the repo seems to be quite ancient tuned-2.15.0 as from: https://packages.ubuntu.com/jammy/all/tuned The latest TuneD version is tuned-2.23.0 ( https://github.com/redhat-performance/tuned/releases/). This would also explain why you trigger the sysctl errors. — Reply to this email directly, view it on GitHub <#655 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQ5ST4UV6FHVKURQJCTTGF3ZKOTK7AVCNFSM6AAAAABKGABLPWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMBVGMYTCMJZGA> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[yarda]

Version bump and rebuilt of the Ubuntu package should probably work as with other Ubuntu packages. Or rebuilt of the TuneD package from the newer Ubuntu version. Instructions how to built/rebuilt packages are probably somewhere in the Ubuntu/Debian doc/wiki. Unfortunately, I don't have Ubuntu, thus I cannot provide detailed instructions.

Installation from sources should also work, you can unpack TuneD archive and run make install. But it isn't recommended to mix upstream tarballs with distro packages.

[yarda]

E.g.: https://help.ubuntu.com/community/UpdatingADeb

But I cannot test the instructions.

[stevefxp1]

Thank you...I will try to capture some install instructions for Ubuntu users.Sent from my T-Mobile 5G Device -------- Original message --------From: Jaroslav Škarvada ***@***.***> Date: 7/3/24 17:16 (GMT-05:00) To: redhat-performance/tuned ***@***.***> Cc: stevefxp1 ***@***.***>, Author ***@***.***> Subject: Re: [redhat-performance/tuned] Message from Log (Issue #655) E.g.: https://help.ubuntu.com/community/UpdatingADeb But I cannot test the instructions. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you authored the thread.Message ID: ***@***.***>