Document reasonable radosgw timeout
The documented radosgw timeout of -1 can trigger cinder services to get stuck when experiencing any Ceph/network related issue and does not allow for a graceful recovery. A more reasonable value of 30 (seconds) is allowing for any connectivity releated recovery. Change-Id: I5f04254e77705c13cef165ea24fea3a45f2c20b1 Related-To: #1726936
This commit is contained in:
parent
a7f4d9870a
commit
8d4aed34b8
|
@ -321,7 +321,7 @@ Alternatively, you can also selectively override, like this:
|
|||
rbd_flatten_volume_from_snapshot: 'false'
|
||||
rbd_max_clone_depth: 5
|
||||
rbd_store_chunk_size: 4
|
||||
rados_connect_timeout: -1
|
||||
rados_connect_timeout: 30
|
||||
volume_backend_name: volumes_hdd
|
||||
rbd_user: "{{ cinder_ceph_client }}"
|
||||
rbd_secret_uuid: "{{ cinder_ceph_client_uuid }}"
|
||||
|
|
Loading…
Reference in New Issue