On 12/09/2015 01:32 AM, Dave Chinner wrote:
On Tue, Dec 08, 2015 at 03:56:52PM +0200, Avi Kivity wrote:
On 12/08/2015 08:03 AM, Dave Chinner wrote:
On Wed, Dec 02, 2015 at 10:34:14AM +0200, Avi Kivity wrote:
On 12/02/2015 02:13 AM, Brian Foster wrote:
Metadata is modified in-core and handed off to the logging
infrastructure via a transaction. The log is flushed to disk some time
later and metadata writeback occurs asynchronously via the xfsaild
thread.
Unless, I expect, if the log is full. Since we're hammering on the
disk quite heavily, the log would be fighting with user I/O and
possibly losing.
Does XFS throttle user I/O in order to get the log buffers recycled faster?
No. XFS tags the metadata IO with REQ_META that the IO schedulers
can tell the difference between metadata and data IO, and schedule
them appropriately. Further. log buffers are also tagged with
REQ_SYNC to indicate they are latency sensitive IOs, whcih the IO
schedulers again treat differently to minimise latency in the face
of bulk async IO which is not latency sensitive.
IOWs, IO prioritisation and dispatch scheduling is the job of the IO
scheduler, not the filesystem. The filesystem just tells the
scheduler how to treat the different types of IO...
Is there any way for us to keep track of it, and reduce disk
pressure when it gets full?
Only if you want to make more problems for yourself - second
guessing what the filesystem is going to do will only lead you to
dancing the Charlie Foxtrot on a regular basis. :/
So far the best approach I found that doesn't conflict with this is
to limit io_submit iodepth to the natural disk iodepth (or a small
multiple thereof). This seems to keep XFS in its comfort zone, and
is good for latency anyway.
That's pretty much what I just explained in my previous reply. ;)
The only issue is that the only way to obtain this parameter is to
measure it.
Yup, exactly what I've been saying ;)
However, You can get a pretty good guess on max concurrency from the
device characteristics in sysfs:
/sys/block/<dev>/queue/nr_requests
That's just a fixed number. AFAICT, it isn't derived from the actual device.
"measure it" is better than nothing, but when you want to distribute
software that works out of the box and does not need extensive tuning,
it leaves something to be desired.
I'm thinking about detecting the limit dynamically (below the limit,
throughput is roughly proportional to concurrency; above the limit,
throughput is fixed while latency is proportional to concurrency). The
problem is that the measurement is very noisy, the more so because we
are two layers above the hardware, and driving it from cores that try
very hard not to communicate.
The right place to do this is the block layer.
gives you the maximum IO scheduler request queue depth, and
/sys/block/<dev>/device/queue_depth
gives you the hardware command queue depth.
That's more useful, but it really describes the bus/link/protocol rather
than the device itself.
I don't have this queue_depth attribute for my nvme0n1 device (4.1.7).
E.g. a random iscsi device I have attached to a test VM:
$ cat /sys/block/sdc/device/queue_depth
32
$ cat /sys/block/sdc/queue/nr_requests
127
Which means 32 physical IOs can be in flight concurrently, and the
IO scheduler will queue up to roughly another 100 discrete IOs
before it starts blocking incoming IO requests (127 is the typical
io scheduler queue depth default). That means maximum non-blocking
concurrency is going to be around 100-130 IOs in flight at once.
I wrote a small tool to do this [1], but it's a hassle for users.
[1] https://github.com/avikivity/diskplorer
I note that the NVMe device you tested in the description hits
maximum performance with concurrency at around 110-120 read IOs in
flight. :)
We increased nr_requests for the test so it wouldn't block. So it's the
actual device characteristics, not an artifact of the software stack.
If you consider a RAID of these, you can easily need a few hundred
concurrent ops.
IIRC nvme's maximum iodepth is 64k.
_______________________________________________
xfs mailing list
xfs@xxxxxxxxxxx
http://oss.sgi.com/mailman/listinfo/xfs
