From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mike Christie Subject: PATCH 0/4] multipath-tools: Ceph rbd support v2 Date: Mon, 8 Aug 2016 07:01:46 -0500 Message-ID: <1470657710-28081-1-git-send-email-mchristi@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: dm-devel-bounces@redhat.com Errors-To: dm-devel-bounces@redhat.com To: dm-devel@redhat.com, christophe.varoqui@opensvc.com List-Id: dm-devel.ids The following patches made over Christophe's tree today, add Ceph rbd support for handling blacklisted devices. This is not general support for rbd and multipath. There is no boot/root support and it does not support features like multibus. My use is for HA configurations, specifically for exporting rbd images through multiple LIO instances. In this case, we have one rbd instance that has the ceph rbd exclusive lock and it can send WRITE requests. If that host becomes unreachable, then another host will grab the lock, and blacklist the original host to prevent it from sending stale IO (when blacklisted IO will be failed by the OSD). To recover from that type of scenario, this patchset adds a repair() callout to the checker. If the path is in the PATH_DOWN state this callout can be used to fix it up. For my case, I am remapping the device to flush stale IO and cleanup the old lock, and then unblacklisting the path, so it can be used again. Changes since v1: - Drop ID_UID use and implemented sysfs getuid support. - Drop settings that were defaults and follow template. - Fix ceph auth/user.