Looking at the debug logs in the associated tracker ticket (thanks!), it looks like three update object map ops were sent to the OSD and never completed. This is strange since milliseconds before those three requests were sent, six object map update ops were sent to the same object and completed successfully. Are you still able to reproduce this issue? If so, can you re-run with "--debug-ms=20 --debug-objecter=20" as well? I attempted to import the same cirros image using a Jewel 10.2.7 client and Kraken 11.2.0 OSDs and was unable to reproduce. Thanks, Jason On Tue, May 9, 2017 at 10:18 AM, Florian Haas <florian@xxxxxxxxxxx> wrote: > Hi everyone, > > just a quick heads-up on this issue: http://tracker.ceph.com/issues/19889 > > TL;DR: "rbd import" looks like it breaks when upgrading a cluster from > Jewel to Kraken (and potentially, Luminous — not sure about that) and > before upgrading the client itself. In other words, this might cause > disruption in upgrades with clients using "rbd import"-like > functionality (such as OpenStack Glance). > > Cheers, > Florian > -- > 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 -- 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
