Project

General

Profile

Actions

Bug #47502

closed

ceph-volume lvm batch race condition

Added by Marc Gariépy over 3 years ago. Updated over 3 years ago.

Status:
Resolved
Priority:
Normal
Assignee:
-
Target version:
-
% Done:

0%

Source:
Tags:
Backport:
octopus,nautilus
Regression:
No
Severity:
2 - major
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

when usng ceph-volume to create the osd, some ssd are failing to get the correct rotational state.

after some debugging with @Guillaume Abrioux we are pretty sure it's a race condition.

I added some prints in that function and ran the report a few times.
https://github.com/ceph/ceph/blob/master/src/ceph-volume/ceph_volume/util/device.py#L295-L302

 
rotational state of the devices.

   9496 sys_api: 0
    104 sys_api: 1
   9580 disk_api: 0
     20 disk_api: 1

after some more digging i see that udevadm monitor gets triggered at every run of ceph-volume.

issue with ceph-ansible:
https://github.com/ceph/ceph-ansible/issues/5727

# ceph-volume --cluster ceph lvm batch --bluestore --yes --dmcrypt /dev/sdf /dev/sdo /dev/sdw /dev/sdd /dev/sdm /dev/sdu /dev/sdb /dev/sdk /dev/sds /dev/sdi /dev/sdq /dev/sdg /dev/sdx /dev/sde /dev/sdn /dev/sdv /dev/sdc /dev/sdl /dev/sdt /dev/sda /dev/sdj /dev/sdr /dev/sdh /dev/sdp --report

Total OSDs: 24

  Type            Path                                                    LV Size         % of device
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdf                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdo                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdw                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdd                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdm                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdu                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdb                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdk                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sds                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdi                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdq                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdg                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdx                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sde                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdn                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdv                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdc                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdl                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdt                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sda                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdj                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdr                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdh                                                3.49 TB         100.0%
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdp                                                3.49 TB         100.0%

# ceph-volume --cluster ceph lvm batch --bluestore --yes --dmcrypt /dev/sdf /dev/sdo /dev/sdw /dev/sdd /dev/sdm /dev/sdu /dev/sdb /dev/sdk /dev/sds /dev/sdi /dev/sdq /dev/sdg /dev/sdx /dev/sde /dev/sdn /dev/sdv /dev/sdc /dev/sdl /dev/sdt /dev/sda /dev/sdj /dev/sdr /dev/sdh /dev/sdp --report

Total OSDs: 1

Solid State VG:
  Targets:   block.db                  Total size: 80.30 TB                 
  Total LVs: 1                         Size per LV: 80.30 TB                 
  Devices:   /dev/sdo, /dev/sdw, /dev/sdd, /dev/sdm, /dev/sdu, /dev/sdb, /dev/sdk, /dev/sds, /dev/sdi, /dev/sdq, /dev/sdg, /dev/sdx, /dev/sde, /dev/sdn, /dev/sdv, /dev/sdc, /dev/sdl, /dev/sdt, /dev/sda, /dev/sdj, /dev/sdr, /dev/sdh, /dev/sdp

  Type            Path                                                    LV Size         % of device
----------------------------------------------------------------------------------------------------
  [data]          /dev/sdf                                                3.49 TB         100.0%
  [block.db]      vg: vg/lv                                               80.30 TB        100%

Related issues 4 (1 open3 closed)

Related to ceph-volume - Bug #47585: cache disk.get_devices return valueNew

Actions
Has duplicate ceph-volume - Bug #47553: OSDs being deployed bluestore mixed type for SSD disksDuplicate

Actions
Copied to ceph-volume - Backport #47649: octopus: ceph-volume lvm batch race conditionResolvedJan FajerskiActions
Copied to ceph-volume - Backport #47650: nautilus: ceph-volume lvm batch race conditionResolvedJan FajerskiActions
Actions

Also available in: Atom PDF