On Sun, Jan 29, 2023 at 1:12 PM Jiatong Shen <yshxxsjt715@gmail.com> wrote:
Hello Ilya,
Thank you very much for the clarification! Another question, for some historical reasons, there are still some Luminous clients existing. Is it dangerous to sparsify an image which is still being used by a Luminous client?
The atomicity is ensured on the OSD side so with a Nautilus or Octopus cluster and a lone Luminous client it should be fine. Thanks, Ilya
Thank you very much for informing me that N/O are both retired. We are definitely going to investigate on upgrading.
Best,
Jiatong Shen
On Sun, Jan 29, 2023 at 6:55 PM Ilya Dryomov <idryomov@gmail.com> wrote:
On Sun, Jan 29, 2023 at 11:29 AM Jiatong Shen <yshxxsjt715@gmail.com> wrote:
Hello community experts,
I would like to know the status of rbd image sparsify. From the website, it should be added at Nautilus ( https://docs.ceph.com/en/latest/releases/nautilus/ from pr (26226 <https://github.com/ceph/ceph/pull/26226>,) ) but on mimic it is related again https://docs.ceph.com/en/latest/releases/octopus/) Is it still the same commit? or is there any update/bug fixes that is not backported to 14.x?
Hi Jiatong,
"rbd sparsify" for replicated pools [1] was released in Nautilus (14.2.0). Support for EC pools [2] was added a bit later, targeted for Octopus (15.2.0), but it was also made available in an early point release of Nautilus (14.2.2).
Please note that both Nautilus and Octopus releases are EOL and no longer supported though.
I am also curious why it is named re-sparsify in the pr title?
By default, all RBD images are sparse. "rbd sparsify" can bring some of that sparseness back, hence the re- prefix.
[1] https://github.com/ceph/ceph/pull/26226 [2] https://github.com/ceph/ceph/pull/27268
Thanks,
Ilya
--
Best Regards,
Jiatong Shen