/var/lib/nova|glance disk drive is not present

Bug #1317576 reported by Vladimir Kuklin
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Fuel for OpenStack
Fix Released
Critical
Dmitry Ilyin

Bug Description

sometimes ubuntu provisioning leads to broken partitioning and particular partiotions not mounted

vgchange -ay shows following:

  WARNING: Duplicate VG name image: Existing xtYzHT-hjc8-SJ4E-fGVG-AmwD-k8jh-7YhqoX (created here) takes precedence over Jtg0w5-B5Jq-1Nao-0f2A-jKya-8QsD-4o1wAK
  WARNING: Duplicate VG name image: Existing Jtg0w5-B5Jq-1Nao-0f2A-jKya-8QsD-4o1wAK (created here) takes precedence over xtYzHT-hjc8-SJ4E-fGVG-AmwD-k8jh-7YhqoX
  Couldn't find device with uuid 2Rm3lv-jUCv-qdAK-OID3-lPAg-zzYa-j8UwdE.
  Refusing activation of partial LV glance. Use --partial to override.
  0 logical volume(s) in volume group "image" now active
  Refusing activation of partial LV glance. Use --partial to override.
  0 logical volume(s) in volume group "image" now active

Revision history for this message
Vladimir Kuklin (vkuklin) wrote :

root@node-1:~# pvs
  WARNING: Duplicate VG name image: Existing xtYzHT-hjc8-SJ4E-fGVG-AmwD-k8jh-7YhqoX (created here) takes precedence over Jtg0w5-B5Jq-1Nao-0f2A-jKya-8QsD-4o1wAK
  Couldn't find device with uuid 2Rm3lv-jUCv-qdAK-OID3-lPAg-zzYa-j8UwdE.
  Couldn't find device with uuid jh2Mt8-8PQL-Jn4f-9Wcl-jUyD-HZMm-8MLNWu.
  Couldn't find device with uuid dNBNVQ-dViw-1MiN-FoF6-Vc2t-WliX-wU3QQp.
  PV VG Fmt Attr PSize PFree
  /dev/vda5 image lvm2 a- 31.44g 128.00m
  /dev/vdb2 image lvm2 a- 49.44g 0
  /dev/vdc2 image lvm2 a- 49.44g 0
  unknown device image lvm2 a- 31.44g 128.00m
  unknown device image lvm2 a- 49.44g 0
  unknown device image lvm2 a- 49.44g 0

Revision history for this message
Vladimir Kuklin (vkuklin) wrote :

configuration of LVM seems pretty strange

        physical_volumes {

                pv0 {
                        id = "809Ixl-xnm0-qxOr-qqi3-lJB6-edYz-20FF0L"
                        device = "/dev/vda5" # Hint only

                        status = ["ALLOCATABLE"]
                        flags = []
                        dev_size = 65953793 # 31.4492 Gigabytes
                        pe_start = 384
                        pe_count = 1006 # 31.4375 Gigabytes
                }

                pv1 {
                        id = "jh2Mt8-8PQL-Jn4f-9Wcl-jUyD-HZMm-8MLNWu"
                        device = "unknown device" # Hint only

                        status = ["ALLOCATABLE"]
                        flags = ["MISSING"]
                        dev_size = 103702528 # 49.4492 Gigabytes
                        pe_start = 384
                        pe_count = 1582 # 49.4375 Gigabytes
                }

                pv2 {
                        id = "dNBNVQ-dViw-1MiN-FoF6-Vc2t-WliX-wU3QQp"
                        device = "unknown device" # Hint only

                        status = ["ALLOCATABLE"]
                        flags = ["MISSING"]
                        dev_size = 103702528 # 49.4492 Gigabytes
                        pe_start = 384
                        pe_count = 1582 # 49.4375 Gigabytes
                }
        }

        logical_volumes {

                glance {
                        id = "0Pvw0M-cWDG-14ZH-Wg0c-ZEds-78z2-qeX43W"
                        status = ["READ", "WRITE", "VISIBLE"]
                        flags = []
                        segment_count = 3

                        segment1 {
                                start_extent = 0
                                extent_count = 1582 # 49.4375 Gigabytes

                                type = "striped"
                                stripe_count = 1 # linear

                                stripes = [
                                        "pv1", 0
                                ]
                        }
                        segment2 {
                                start_extent = 1582
                                extent_count = 1582 # 49.4375 Gigabytes

                                type = "striped"
                                stripe_count = 1 # linear

                                stripes = [
                                        "pv2", 0
                                ]
                        }
                        segment3 {
                                start_extent = 3164
                                extent_count = 1002 # 31.3125 Gigabytes

                                type = "striped"
                                stripe_count = 1 # linear

                                stripes = [
                                        "pv0", 0
                                ]
                        }
                }
        }
}

Revision history for this message
Vladimir Kuklin (vkuklin) wrote :

it seems like some of devices are duplicated by udev

  PV VG Fmt Attr PSize PFree
  /dev/vda5 image lvm2 a- 31.44g 128.00m
  /dev/vdb2 image lvm2 a- 49.44g 0
  /dev/vdc2 image lvm2 a- 49.44g 0
  unknown device image lvm2 a- 31.44g 128.00m
  unknown device image lvm2 a- 49.44g 0
  unknown device image lvm2 a- 49.44g 0

Revision history for this message
Vladimir Kuklin (vkuklin) wrote :
Mike Scherbakov (mihgen)
Changed in fuel:
assignee: Vladimir Kozhukalov (kozhukalov) → Fuel Library Team (fuel-library)
Revision history for this message
Vladimir Kozhukalov (kozhukalov) wrote :

Last few patches to pmanager.py make additional logging.
https://review.openstack.org/#/c/91531/
https://review.openstack.org/#/c/92621/
https://review.openstack.org/#/c/92849/

Besides that patch which fixes the issue which appears on HP Smart Array tries to clean lvm metadata right after creating new partitions, so let's make this bug medium and try to reproduce it with those patches. If we can reproduce it we'll be able to see what's going on.

Revision history for this message
Dmitry Borodaenko (angdraug) wrote :

Marking Incomplete as we need to reproduce it again as per Vladimir's comment.

Changed in fuel:
status: New → Incomplete
Revision history for this message
Andrey Sledzinskiy (asledzinskiy) wrote :

I get similar problem on {"build_id": "2014-05-14_01-10-31", "mirantis": "yes", "build_number": "203", "ostf_sha": "ef970b442437072bdfa4ea99a7b2971215b2de18", "nailgun_sha": "155acff248aed9a8295d03d58346daa4851d49b4", "production": "docker", "api": "1.0", "fuelmain_sha": "9df792985f8984063979f16dc94b4df24ef40c2d", "astute_sha": "80e60b66e3cb4e3e61b22c61c4acfa127ba1bf7e", "release": "5.0", "fuellib_sha": "89ccab7ee76980e38c4d9a5fbcdf7df87e35d61f"}

Cluster - Ubuntu, Simple, KVM, Flat Nova-Network, Cinder LVM, 1 controller, 1 compute, 1 cinder
At the end of provisioning get next message on nodes - 'The disk drive for var/lib/glance is not ready yet or not present' (screen is attached)
If skip it with 's' deployment would be succsessful

Revision history for this message
Andrey Sledzinskiy (asledzinskiy) wrote :
Revision history for this message
Andrey Sledzinskiy (asledzinskiy) wrote :
Changed in fuel:
status: Incomplete → New
status: New → Confirmed
Revision history for this message
Aleksandra Fedorova (bookwar) wrote :
Revision history for this message
Aleksandra Fedorova (bookwar) wrote :
Changed in fuel:
importance: High → Critical
Mike Scherbakov (mihgen)
Changed in fuel:
assignee: Fuel Library Team (fuel-library) → Dmitry Ilyin (idv1985)
Revision history for this message
Igor Shishkin (teran) wrote :

Got the similar on 205th ISO

Revision history for this message
Vladimir Kozhukalov (kozhukalov) wrote :

The original problem was likely fixed with those patches
 https://review.openstack.org/#/c/91531/
https://review.openstack.org/#/c/92621/

It is easy to see in logs that lv is created successfully. Unfortunately, third patch https://review.openstack.org/#/c/92849/ introduced mkfs.xfs option error. This error was fixed in the following 2 patches.
https://review.openstack.org/#/c/93535/
https://review.openstack.org/93726

Revision history for this message
Aleksandra Fedorova (bookwar) wrote :

Can not reproduce on Fuel CI with iso 207.

https://fuel-jenkins.mirantis.com/job/master_fuellib_review_systest_ubuntu/846/console

Provisioning was successfull (though deployment failed because of something else)

Revision history for this message
Mike Scherbakov (mihgen) wrote :

According to comments of A.Fedorova & V.Kozhukalov, the issue is actually fixed. Closing as Fix Committed. Please reopen and notify the team in #fuel-dev IRC channel if you face this issue ever again.

Changed in fuel:
status: Confirmed → Fix Committed
Igor Shishkin (teran)
Changed in fuel:
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.