On 05/29/2018 02:17 PM, Josef Bacik wrote:
> From: Josef Bacik <jba...@fb.com>
>
> A basic documentation to describe the interface, statistics, and
> behavior of io.latency.
>
> Signed-off-by: Josef Bacik <jba...@fb.com>
> ---
> Documentation/blk-iolatency.txt | 80
> +++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 80 insertions(+)
> create mode 100644 Documentation/blk-iolatency.txt
>
> diff --git a/Documentation/blk-iolatency.txt b/Documentation/blk-iolatency.txt
> new file mode 100644
> index 000000000000..9dd86f4f64b6
> --- /dev/null
> +++ b/Documentation/blk-iolatency.txt
> @@ -0,0 +1,80 @@
> +Block IO Latency Controller
> +
> +Overview
> +========
> +
> +This is a cgroup v2 controller for IO workload protection. You provide a
> group
> +with a latency target, and if the average latency exceeds that target the
> +controller will throttle any peers that have a lower latency target than the
> +protected workload.
> +
> +Interface
> +=========
> +
> +- io.latency. This takes a similar format as the other controllers
end above sentence with a period or a colon:
> +
> + "MAJOR:MINOR target=<target time in microseconds"
> +
> +- io.stat. If the controller is enabled you will see extra stats in io.stat
> in
> + addition to the normal ones
end above sentence with a period.
> +
> + - depth=<integer>. This is the current queue depth for the group.
> + - delay=<time in microseconds>. This is the current delay per task that
> + does IO in this group.
> + - use_delay=<integer>. This is how deep into the delay we currently
> + are, the larger this number is the longer it'll take us to get back to
> + queue depth > 1.
> + - total_lat_avg=<time in microseconds>. The running average IO latency
> + for this group. Running average is generally flawed, but will give an
> + admistrator a general idea of the overall latency they can expect for
administrator
> + their workload on the given disk.
> +
> +HOWTO
> +=====
> +
> +The limits are only applied at the peer level in the heirarchy. This means
> that
> +in the diagram below, only groups A, B, and C will influence eachother, and
each other, and
> +groups D and F will influence eachother. Group G will influence nobody.
each other.
> +
> + [root]
> + / | \
> + A B C
> + / \ |
> + D F G
> +
> +
> +So the ideal way to configure this is to set io.latency in groups A, B, and
> C.
> +Generally you do not want to set a value lower than the latency your device
> +supports. Experiment to find the value that works best for your workload,
> start
> +at higher than the expected latency for your device and watch the
> total_lat_avg
> +value in io.stat for your workload group to get an idea of the latency you
> see
> +during normal operation. Use this value as a basis for your real setting,
> +setting at 10-15% higher than the value in io.stat. Experimentation is key
> here
> +because total_lat_avg is a running total, so is the "statistics" portion of
> +"lies, damned lies, and statistics."
> +
> +How Throttling Works
> +====================
> +
> +io.latency is work conserving, so as long as everybody is meeting their
> latency
> +target the controller doesn't do anything. Once a group starts missing it's
its
> +target it begins throttling any peer group that has a higher target than
> itself.
> +This throttling takes 2 forms
2 forms:
> +
> +- Queue depth throttling. This is the number of outstanding IO's a group is
> + allowed to have. We will clamp down relatively quickly, starting at no
> limit
> + and going all the way down to 1 IO at a time.
> +
> +- Artificial delay induction. There are certain types of IO that cannot be
> + throttled without possibly adversely affecting higher priority groups.
> This
> + includes swapping and metadata IO. These types of IO are allowed to occur
> + normally, however they are "charged" to the originating group. If the
> + originating group is being throttled you will see the use_delay and delay
> + fields in io.stat increase. The delay value is how many microseconds that
> are
> + being added to any process that runs in this group. Because this number
> can
> + grow quite large if there is a lot of swapping or metadata IO occuring we
occurring
> + limit the individual delay events to 1 second at a time.
> +
> +Once the victimized group starts meeting it's latency target again it will
> start
its
> +unthrottling any peer groups that were throttled previous. If the victimized
previously.
> +group simply stops doing IO the global counter will unthrottle appropriately.
>
--
~Randy
