Andras,
Unfortunately your attachment didn't come through the list. (It might
work if you embed it inline? Not sure.) I don't know if anybody's
looked too hard at this before, and without the image I don't know
exactly what metric you're using to say something's 320KB in size. Can
you explain more
Hi Greg,
Much appreciated for the reply, the image is also available at:
https://tracker.ceph.com/attachments/download/5808/Bytes_per_op.png
How the graph is generated: we back the cephfs metadata pool with Azure
ultrassd disks. Azure reports for the disks each minute the average
read/write iops
Hi Greg,
As a follow up, we see items similar to this pop up in the
objecter_requests (when it's not empty). Not sure if reading it right, but
some appear quite large (in the MB range?):
{
"ops": [
{
"tid": 9532804,
"pg": "3.f9c235d7",
"osd": 2,
I see Xiubo started discussing this on
https://tracker.ceph.com/issues/53542 as well.
So the large writes are going to the journal file, and sometimes it's
s single write of a full segment size, which is what I was curious
about.
At this point the next step is seeing what is actually taking up th
Hi Xiubo,
Thanks very much for looking into this, that does sound like what might
be happening in our case.
Is this something that can be improved somehow - would disabling pinning or
some config change help? Or could this be addressed in a future release?
It seems somehow excessive to write so