<div dir="ltr">Great, I will do that. Thank you for all your help!</div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Wed, Apr 24, 2019 at 6:22 PM Mike Snitzer <<a href="mailto:snitzer@redhat.com">snitzer@redhat.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On Wed, Apr 24 2019 at 5:27pm -0400,<br>
Mike Snitzer <<a href="mailto:snitzer@redhat.com" target="_blank">snitzer@redhat.com</a>> wrote:<br>
<br>
> <br>
> On Wed, Apr 24 2019 at 10:46am -0400,<br>
> Ryan Norwood <<a href="mailto:ryan.p.norwood@gmail.com" target="_blank">ryan.p.norwood@gmail.com</a>> wrote:<br>
> <br>
> > On Wed, Apr 24, 2019 at 9:08 AM Ryan Norwood <[1]<a href="mailto:ryan.p.norwood@gmail.com" target="_blank">ryan.p.norwood@gmail.com</a>><br>
> > wrote:<br>
> > <br>
> > Thank you for your help.<br>
> > You are correct, it appears that the problem occurs when there is a RAID<br>
> > 5 or RAID 50 volume beneath VDO.<br>
> > NAME KNAME RA SIZE ALIGNMENT MIN-IO OPT-IO PHY-SEC LOG-SEC<br>
> > RQ-SIZE SCHED WSAME<br>
> > sdh<br>
> > sdh 128 977.5G 0 512 0 512 512 128<br>
> > deadline 0B<br>
> > +-sed6<br>
> > dm-6 128 977.5G 0 512 0 512 512 128<br>
> > 0B<br>
> > +-md127<br>
> > md127 12288 5.7T 0 1048576 6291456 512 512 128<br>
> > 0B<br>
> > +-vdo_data<br>
> > dm-17 128 5.7T 0 1048576 6291456 512 512 128<br>
> > 0B<br>
> > +-vdo<br>
> > dm-18 128 57.3T 0 4096 4096 4096 4096 128<br>
> > 0B<br>
<br>
<snip><br>
<br>
> > /sys/block/dm-18/queue/max_hw_sectors_kb:4<br>
> > /sys/block/dm-18/queue/max_sectors_kb:4<br>
<br>
These are getting set as a side-effect of MD raid imposing the need for<br>
merge_bvec (in the context of RHEL7.x only, not upstream) otherwise it<br>
goes conservative and forces the IO to be contrained to a single page,<br>
please see:<br>
<br>
drivers/md/dm-table.c:dm_set_device_limits() at the end:<br>
<br>
/*<br>
* Check if merge fn is supported.<br>
* If not we'll force DM to use PAGE_SIZE or<br>
* smaller I/O, just to be safe.<br>
*/<br>
if (dm_queue_merge_is_compulsory(q) && !ti->type->merge)<br>
blk_limits_max_hw_sectors(limits,<br>
(unsigned int) (PAGE_SIZE >> 9));<br>
<br>
With MD raid in the IO stack, dm_queue_merge_is_compulsory() will return<br>
true, so the VDO target not providing ti->type->merge causes this issue.<br>
<br>
Please file a BZ at <a href="http://bugzilla.redhat.com" rel="noreferrer" target="_blank">bugzilla.redhat.com</a> against VDO and I'll continue to<br>
work with the VDO developers to get this fixed for you for RHEL7.5, etc.<br>
<br>
Thanks,<br>
Mike<br>
</blockquote></div>