Description of problem: When a cluster is installed using "osd_auto_discovery" set to true, the purge cluster fails. If "osd_auto_dscovery" is true , the customer would not know which devices have osd's This could be a problem on large clusters. Can we try for a mor eintelligent way of purging clusters? PLAY [confirm whether user really meant to purge the cluster] ***************** GATHERING FACTS *************************************************************** ok: [localhost] TASK: [exit playbook, if user did not mean to purge cluster] ****************** skipping: [localhost] PLAY [stop ceph cluster] ****************************************************** GATHERING FACTS *************************************************************** ok: [magna058] ok: [magna046] ok: [magna052] TASK: [check for a device list] *********************************************** failed: [magna046] => {"failed": true} msg: OSD automatic discovery was detected, purge cluster does not support this scenario. If you want to purge the cluster, manually provide the list of devices in group_vars/osds using the devices variable. failed: [magna052] => {"failed": true} msg: OSD automatic discovery was detected, purge cluster does not support this scenario. If you want to purge the cluster, manually provide the list of devices in group_vars/osds using the devices variable. failed: [magna058] => {"failed": true} msg: OSD automatic discovery was detected, purge cluster does not support this scenario. If you want to purge the cluster, manually provide the list of devices in group_vars/osds using the devices variable. FATAL: all hosts have already failed -- aborting PLAY RECAP ******************************************************************** to retry, use: --limit @/root/purge-cluster.retry localhost : ok=1 changed=0 unreachable=0 failed=0 magna046 : ok=1 changed=0 unreachable=0 failed=1 magna052 : ok=1 changed=0 unreachable=0 failed=1 magna058 : ok=1 changed=0 unreachable=0 failed=1 How reproducible: Always
I don't think we want ceph-ansible to try to auto discover devices to purge, requiring the user to provide the devices seems the safest way to purge. If this is a priority, please reopen.