Looking at nova-compute.log I see errors trying to find LUN 0 during the sysfs stage.
Several machines won't boot because their root disk entries in LVM are seen as PV and booting them doesn't see them in the DM subsystem.
Other machines boot but there attached disks throw LVM errors about duplicate PV and preferring the cinder-volumes VG version.
LVM is showing LVs that have both "bare" entries as well as entries in cinder-volumes and it's complaining about duplicate PVs, not using lvmetad and preferring some entries because they are in the dm subsystem.
I've verified that, so far, I haven't lost any data. The "bare" LV not being used as part of the DM subsystem because it's server won't boot can be mounted on the openstack host and all data on it is accessible.
This host has rebooted cleanly multiple times in the past. This is the first time it's shown any problems.
How can I reset the LVM configuration and convince it that it's not seeing duplicate PV?
2020-09-24 11:15:27.091 13953 INFO os_brick.initiator.connectors.iscsi [req-8d15fb6a-6324-471e-9497-587885eef8f6 396aeda6552f44fdac5f878b90325ee1 54af92f2bb494355b96024076184d1c8
- default default] Trying to connect to iSCSI portal
172.10.0.40:32602020-09-24 11:15:29.721 13953 WARNING nova.compute.manager [req-fd32e16f-c879-402f-a32c-6be45a943c34 48af9a366301467d9fec912fd1c072c6 f9fc7b412a8446d083da1356aa370eb4 - default d
efault] [instance: de7d740c-786a-4aa2-aa09-d447ae7e14b6] Received unexpected event network-vif-unplugged-79aff403-d2e4-4266-bd88-d7bd19d501a9 for instance with vm_state stopped a
nd task_state powering-on.
2020-09-24 11:16:21.361 13953 WARNING os_brick.initiator.connectors.iscsi [req-8d15fb6a-6324-471e-9497-587885eef8f6 396aeda6552f44fdac5f878b90325ee1 54af92f2bb494355b96024076184d
1c8 - default default] LUN 0 on iSCSI portal
172.10.0.40:3260 not found on sysfs after logging in.
2020-09-24 11:16:23.482 13953 INFO os_brick.initiator.connectors.iscsi [req-8d15fb6a-6324-471e-9497-587885eef8f6 396aeda6552f44fdac5f878b90325ee1 54af92f2bb494355b96024076184d1c8
- default default] Trying to connect to iSCSI portal
172.10.0.40:32602020-09-24 11:17:17.741 13953 WARNING os_brick.initiator.connectors.iscsi [req-8d15fb6a-6324-471e-9497-587885eef8f6 396aeda6552f44fdac5f878b90325ee1 54af92f2bb494355b96024076184d
1c8 - default default] LUN 0 on iSCSI portal
172.10.0.40:3260 not found on sysfs after logging in.: VolumeDeviceNotFound: Volume device not found at .
2020-09-24 11:17:21.864 13953 INFO os_brick.initiator.connectors.iscsi [req-8d15fb6a-6324-471e-9497-587885eef8f6 396aeda6552f44fdac5f878b90325ee1 54af92f2bb494355b96024076184d1c8
- default default] Trying to connect to iSCSI portal
172.10.0.40:32602020-09-24 11:18:16.113 13953 WARNING os_brick.initiator.connectors.iscsi [req-8d15fb6a-6324-471e-9497-587885eef8f6 396aeda6552f44fdac5f878b90325ee1 54af92f2bb494355b96024076184d
1c8 - default default] LUN 0 on iSCSI portal
172.10.0.40:3260 not found on sysfs after logging in.: VolumeDeviceNotFound: Volume device not found at .
2020-09-24 11:18:17.252 13953 INFO nova.compute.manager [req-8d15fb6a-6324-471e-9497-587885eef8f6 396aeda6552f44fdac5f878b90325ee1 54af92f2bb494355b96024076184d1c8 - default defa
ult] [instance: de7d740c-786a-4aa2-aa09-d447ae7e14b6] Successfully reverted task state from powering-on on failure for instance.
2020-09-24 11:18:17.279 13953 ERROR oslo_messaging.rpc.server [req-8d15fb6a-6324-471e-9497-587885eef8f6 396aeda6552f44fdac5f878b90325ee1 54af92f2bb494355b96024076184d1c8 - defaul
t default] Exception during message handling: VolumeDeviceNotFound: Volume device not found at .
2020-09-24 11:18:17.279 13953 ERROR oslo_messaging.rpc.server Traceback (most recent call last):
2020-09-24 11:18:17.279 13953 ERROR oslo_messaging.rpc.server File "/usr/lib/python2.7/site-packages/oslo_messaging/rpc/server.py", line 163, in _process_incoming
2020-09-24 11:18:17.279 13953 ERROR oslo_messaging.rpc.server res = self.dispatcher.dispatch(message)
--- lvs output ---
I've annotated 1 machine's disks to illustrate the relationship between the volume-*** cinder-volumes vg entries and the "bare" lv seen as directly accessible from the host.
There are 3 servers that won't boot, they are the one's who's home/vg_home and encrypted_home/encrypted_vg entries are shown.
WARNING: Not using lvmetad because duplicate PVs were found.
WARNING: Use multipath or vgimportclone to resolve duplicate PVs?
WARNING: After duplicates are resolved, run "pvscan --cache" to enable lvmetad.
WARNING: Not using device /dev/sdu for PV yZy8Xk-foKT-ovjV-0EZv-VxEM-GqiP-WH7k53. == backup_lv/encrypted_vg
WARNING: Not using device /dev/sdv for PV tHA9ui-eSIO-MDmI-RM3u-3Bf4-Dznb-Ha3XfP. == varoptgitlab/encrypted_vg
WARNING: Not using device /dev/sdm for PV 5eoyCa-sMO4-b7O4-jIfh-byZE-L5pS-3lOu0D.
WARNING: Not using device /dev/sdp for PV 3BI0nV-TP0k-rgPC-PrjH-FT7z-reMe-ec1spj.
WARNING: Not using device /dev/sdt for PV ILdbcY-VFCm-fnH6-Y3jc-pdWZ-fnl8-PH3TPe. == storage_lv/encrypted_vg
WARNING: Not using device /dev/sdr for PV zowU2N-oaBh-r4cO-cxgX-YYiq-Kf3q-mqlHfK.
WARNING: PV yZy8Xk-foKT-ovjV-0EZv-VxEM-GqiP-WH7k53 prefers device /dev/cinder-volumes/volume-c8da1abf-7143-422c-9ee5-b2724a71c8ff because device is in dm subsystem.
WARNING: PV tHA9ui-eSIO-MDmI-RM3u-3Bf4-Dznb-Ha3XfP prefers device /dev/cinder-volumes/volume-0a12012f-8c2e-41fb-aa0c-a7ae99c62487 because device is in dm subsystem.
WARNING: PV 5eoyCa-sMO4-b7O4-jIfh-byZE-L5pS-3lOu0D prefers device /dev/cinder-volumes/volume-990a057c-46cc-4a81-ba02-28b72c34791d because device is in dm subsystem.
WARNING: PV 3BI0nV-TP0k-rgPC-PrjH-FT7z-reMe-ec1spj prefers device /dev/cinder-volumes/volume-b6a9da6e-1958-46ea-90b4-ac1aebed8c04 because device is in dm subsystem.
WARNING: PV ILdbcY-VFCm-fnH6-Y3jc-pdWZ-fnl8-PH3TPe prefers device /dev/cinder-volumes/volume-302dd53b-7d05-4f6d-9ada-8f2ed6e1d4c6 because device is in dm subsystem.
WARNING: PV zowU2N-oaBh-r4cO-cxgX-YYiq-Kf3q-mqlHfK prefers device /dev/cinder-volumes/volume-df006472-be7a-4957-972a-1db4463f5d67 because device is in dm subsystem.
LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert
home centos_stack3 -wi-ao---- 4.00g
root centos_stack3 -wi-ao---- 50.00g
swap centos_stack3 -wi-ao---- 4.00g
_snapshot-05b1e46b-1ae3-4cd0-9117-3fb53a6d94b0 cinder-volumes swi-a-s--- 20.00g volume-1d0ff5d5-93a3-44e8-8bfa-a9290765c8c6 0.00
lv_filestore cinder-volumes -wi-ao---- 1.00t
...
volume-c8da1abf-7143-422c-9ee5-b2724a71c8ff cinder-volumes -wi-ao---- 100.00g
volume-0a12012f-8c2e-41fb-aa0c-a7ae99c62487 cinder-volumes -wi-ao---- 60.00g
volume-990a057c-46cc-4a81-ba02-28b72c34791d cinder-volumes -wi-ao---- 200.00g
volume-b6a9da6e-1958-46ea-90b4-ac1aebed8c04 cinder-volumes -wi-ao---- 30.00g
volume-302dd53b-7d05-4f6d-9ada-8f2ed6e1d4c6 cinder-volumes -wi-ao---- 60.00g
volume-df006472-be7a-4957-972a-1db4463f5d67 cinder-volumes -wi-ao---- 250.00g
...
volume-f3250e15-bb9c-43d1-989d-8a8f6635a416 cinder-volumes -wi-ao---- 20.00g
volume-fc1d5fcb-fda1-456b-a89d-582b7f94fb04 cinder-volumes -wi-ao---- 300.00g
volume-fc50a717-0857-4da3-93cb-a55292f7ed6d cinder-volumes -wi-ao---- 20.00g
volume-ff94e2d6-449b-495d-82e6-0debd694c1dd cinder-volumes -wi-ao---- 20.00g
data2 data2_vg -wi-a----- <300.00g
data data_vg -wi-a----- 1.79t
backup_lv encrypted_vg -wi------- <100.00g == ...WH7k53
storage_lv encrypted_vg -wi------- <60.00g == ...PH3TPe
varoptgitlab_lv encrypted_vg -wi------- <200.00g
varoptgitlab_lv encrypted_vg -wi------- <30.00g
varoptgitlab_lv encrypted_vg -wi------- <60.00g == ...Ha3XfP
encrypted_home home_vg -wi-a----- <40.00g
encrypted_home home_vg -wi------- <60.00g
pub pub_vg -wi-a----- <40.00g
pub_lv pub_vg -wi------- <250.00g
rpms repo -wi-a----- 499.99g
home vg_home -wi-a----- <40.00g
gtri_pub vg_pub -wi-a----- 20.00g
pub vg_pub -wi-a----- <40.00g
-- Alan Davis
Principal System Administrator
Apogee Research LLC