On Thu, Oct 01, 2026 at 01:27:42PM +0200, Ján Tomko via Devel wrote:
On a Monday in 2026, Yusuke Fujimaki wrote:
Ceph removed auth_supported option in commit 350cc71d, https://github.com/ceph/ceph/commit/350cc71d7d9e9cd6239d24080082de7ac9bf8b70
auth_supported has long been an optional parameter, and has been replaced by auth_*_required. The error occurs because libvirt always uses the auth_supported option when authenticating to rbd storage.Therefore, remove the code that uses this option.
Resolves: https://gitlab.com/libvirt/libvirt/-/work_items/918
Signed-off-by: Yusuke Fujimaki <usk.fujimaki@gmail.com> --- src/libxl/libxl_conf.c | 6 +----- src/libxl/xen_xl.c | 2 -- src/storage/storage_backend_rbd.c | 7 ------- tests/xlconfigdata/test-rbd-multihost-noauth.cfg | 2 +- 4 files changed, 2 insertions(+), 15 deletions(-)
diff --git a/src/storage/storage_backend_rbd.c b/src/storage/storage_backend_rbd.c index 25cad4a28d..3981f81a79 100644 --- a/src/storage/storage_backend_rbd.c +++ b/src/storage/storage_backend_rbd.c @@ -239,10 +239,6 @@ virStorageBackendRBDOpenRADOSConn(virStorageBackendRBDState *ptr,
if (rc < 0) goto cleanup; - - if (virStorageBackendRBDRADOSConfSet(ptr->cluster, - "auth_supported", "cephx") < 0)
This does not look right - if we've been requesting the auth method before, shouldn't we request it via a new option?
IIUC, the suggestion seems to be let the client+server auth-negotiate the auth to make it "do the right thing". If we wanted to still force it though, it appears we could use auth_client_required instead. IIUC, auth_supported has been a no-op for a while, so just removing it is functionally the same as we've been using for a while.
Jano
- goto cleanup; } else { VIR_DEBUG("Not using cephx authorization"); if (rados_create(&ptr->cluster, NULL) < 0) {
With regards, Daniel -- |: https://berrange.com ~~ https://hachyderm.io/@berrange :| |: https://libvirt.org ~~ https://entangle-photo.org :| |: https://pixelfed.art/berrange ~~ https://fstop138.berrange.com :|