On Fri, Jul 21, 2017 at 4:23 AM, Marcus Furlong <furlongm@xxxxxxxxx> wrote: > On 20 July 2017 at 12:49, Matthew Vernon <mv3@xxxxxxxxxxxx> wrote: >> Hi, >> >> On 18/07/17 05:08, Marcus Furlong wrote: >>> >>> On 22 March 2017 at 05:51, Dan van der Ster <dan@xxxxxxxxxxxxxx >>> <mailto:dan@xxxxxxxxxxxxxx>> wrote: >> >> >>> Apologies for reviving an old thread, but I figured out what happened >>> and never documented it, so I thought an update might be useful. >> >> >> [snip detailed debugging] >> >> Thanks for getting to the bottom of this. Have you reported it to the sgdisk >> authors, please? It'd be good to get it fixed :) > > I have reported it, and Rod (CC:d) has pointed out that this is Could you share where you reported it? > probably an XFS filesystem driver bug. Given that this _only_ occurs with > XFS, that is probably correct. I have not reported upstream though. > > As a workaround in sgdisk, I mentioned that these two sgdisk options > together (which ceph-deploy uses): > > --largest-new=1 --set-alignment=2048 > > could create a partition that uses the largest available block of > space, within the constraints of aligning the start _and_ end > boundaries of that partition. If it did that, it would workaround the > filesystem driver bug. That was what I expected would happen, but it > looks like `--set-alignment` only aligns the beginning of the > partition. > > Regards, > Marcus. > > -- > Marcus Furlong > _______________________________________________ > ceph-users mailing list > ceph-users@xxxxxxxxxxxxxx > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com -- Cheers, Brad _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |