My VM was running Reef, actually. But I determined the issue to be write perms on the erasure coded pool. The resolution in my case was editing my client.libvirt user adding to the mon caps, and allowing the erasure coded pool on the osd caps: ceph auth caps client.libvirt mon 'allow r, allow command "osd blacklist", allow command "osd blocklist", allow command "blacklistop", allow command "blocklistop"' osd 'allow class-read object_prefix rbd_children, allow rwx pool=libvirt-pool, allow rwx pool=images, allow rwx pool=libvirt-pool-ec' :) On 2024-09-13 14:58, Anthony D'Atri wrote:
Is your VM running something older than Luminous? Depending on how your attachment is being done, you may need to specify both pools on the client side too.
On Sep 13, 2024, at 1:45 PM, andre@drenet.net wrote:
I have a Ceph instance I'm wherein I'm trying to use erasure coding with RBD (and libvirt). I've followed https://docs.ceph.com/en/latest/rados/operations/erasure-code/#erasure-codin... and enabled overwrites so that RBD can be used. In doing so, I've set the data_pool to the erasure coded pool, with set to the replicated pool.
I have the following in ceph.conf:
rbd_default_data_pool = libvirt-pool-ec
Here's an rbd info on the image I've created (notice the "data_pool" config):
rbd image 'xxxxxxx': size 500 GiB in 128000 objects order 22 (4 MiB objects) snapshot_count: 0 id: 7e37e442592b data_pool: libvirt-pool-ec block_name_prefix: rbd_data.6.7e37e442592b format: 2 features: layering, exclusive-lock, object-map, fast-diff, deep-flatten, data-pool op_features: flags: create_timestamp: Fri Sep 13 16:34:06 2024 access_timestamp: Fri Sep 13 16:34:06 2024 modify_timestamp: Fri Sep 13 16:34:06 2024 The problem, when I attach this rbd image to a VM, I cannot write to it at all, I get an I/O error. But, when I create and rbd image without the "rbd_default_data_pool = libvirt-pool-ec" setting, I can write fine.
Wanted to see if anyone has any ideas, maybe I'm missing something simple. Thanks in advance! _______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io