On Tue, May 15, 2018 at 4:21 PM, David Disseldorp <ddiss@xxxxxxx> wrote:
> Hi Ilya,
>
> On Tue, 15 May 2018 15:47:43 +0200, Ilya Dryomov wrote:
>
>> Are you asking from a PoC point of view or with an eye towards
>> upstreaming?
>
> Yes.
>
>> krbd changes look reasonable in principle, but I don't
>> see how any of it is useful without LIO changes. Doesn't the whole
>> thing boil down to where you set the expiration time?
>
> I was under the impression that some other applications (using kRBD
> mapped images) may have a use for ensuring that stale I/O isn't
> processed by the cluster after a user/client specified timeout. If you
> don't see a need for it outside of LIO, then indeed there's no point
> merging the kRBD changes.
Do you have an example? As Maged put it, it's purpose is to deal with
a very specific iSCSI corner case (given the constraint that MCS isn't
an option because it's not supported by VMware).
>
> Given that there haven't been any objections to the OSD class
> functionality, I'll proceed with something for tcmu-runner and withdraw
> the kRBD patch-set.
Where would the expiration timestamp come from in the tcmu-runner case?
Thanks,
Ilya
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
