atop logs, I see
avq ("average queue depth") over 600. This is on an LVM Logical Volume (LV). Both the LV and the physical volume have
nr_requests (maximum queue depth) of 128.
(Note, to see
atop, I reduce my terminal font size (View->Zoom out). This gives
atop enough space to show all the statistics it knows about).
It is not possible to write to the maximum queue depth on the LV. So I don't know if the maximum queue depth on the LV does anything. But I expect there is some limit on the queue depth... and that it would not be over 600!
This log sample is from a time when the system was under stress, and not very responsive. So I am trying to analyse it... but it's not clear whether I can trust this particular data point.
What does it mean? Is
atop showing a wrong or misleading figure? Is the kernel doing something wrong? Could the kernel be working as designed, and there's an explanation I don't know about?
The system is Fedora Workstation 28. I don't remember doing any IO tuning. My
atop version is
2.3.0-10.fc28.x86_64, and the kernel is
LVM | ll_2016-swap | busy 59% | read 24328 | write 175735 | KiB/r 4 | KiB/w 4 | MBr/s 0.2 | MBw/s 1.1 |
| avio 1.76 ms
DSK | sda | busy 93% | read 88967 | write 45808 | KiB/r 81 | KiB/w 152 | MBr/s 11.8 | MBw/s 11.4 | avq 96.50 | avio 4.12 ms
$ lsblk NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT sda 8:0 0 465.8G 0 disk ├─sda1 8:1 0 500M 0 part /boot/efi ~ └─sda7 8:7 0 371.4G 0 part ├─alan_dell_2016-fedora 253:0 0 40G 0 lvm / ├─alan_dell_2016-swap 253:1 0 2G 0 lvm [SWAP] └─alan_dell_2016-home 253:2 0 303G 0 lvm /home $ cat /sys/block/dm-1/dm/name alan_dell_2016-swap $ cat /sys/block/dm-1/queue/nr_requests 128 $ cat /sys/block/sda/queue/nr_requests 128 $ echo 4 | sudo tee /sys/block/dm-1/queue/nr_requests 4 tee: /sys/block/dm-1/queue/nr_requests: Invalid argument $ cat /sys/block/dm-1/queue/nr_requests 128 $ echo 4 | sudo tee /sys/block/sda/queue/nr_requests 4 $ cat /sys/block/sda/queue/nr_requests 4 $ cat /sys/block/dm-1/queue/nr_requests 128