On Fedora, stack.sh randomly fails on "sudo vgs stack-volumes-default"
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
devstack |
Fix Released
|
Undecided
|
Bogdan Katynski |
Bug Description
When trying to start DevStack (git d1549d624dea9b0
Logs logs:
2015-04-07 15:49:20.727 | + is_service_enabled c-vol
2015-04-07 15:49:20.730 | + return 0
2015-04-07 15:49:20.730 | + [[ -n lvm:lvmdriver-1 ]]
2015-04-07 15:49:20.731 | + local be be_name be_type
2015-04-07 15:49:20.731 | + for be in '${CINDER_
2015-04-07 15:49:20.731 | + be_type=lvm
2015-04-07 15:49:20.731 | + be_name=lvmdriver-1
2015-04-07 15:49:20.731 | + type init_cinder_
2015-04-07 15:49:20.731 | + [[ lvm == \l\v\m ]]
2015-04-07 15:49:20.731 | + init_default_
2015-04-07 15:49:20.731 | + [[ 0 = \0 ]]
2015-04-07 15:49:20.732 | + init_lvm_
2015-04-07 15:49:20.732 | + local vg=stack-
2015-04-07 15:49:20.732 | + local size=10250M
2015-04-07 15:49:20.732 | + is_fedora
2015-04-07 15:49:20.732 | + [[ -z Fedora ]]
2015-04-07 15:49:20.732 | + '[' Fedora = Fedora ']'
2015-04-07 15:49:20.732 | + start_service lvm2-lvmetad
2015-04-07 15:49:20.733 | + is_ubuntu
2015-04-07 15:49:20.733 | + [[ -z rpm ]]
2015-04-07 15:49:20.733 | + '[' rpm = deb ']'
2015-04-07 15:49:20.733 | + sudo /sbin/service lvm2-lvmetad start
2015-04-07 15:49:20.753 | Redirecting to /bin/systemctl start lvm2-lvmetad.
2015-04-07 15:49:20.763 | + '[' tgtadm = tgtadm ']'
2015-04-07 15:49:20.764 | + start_service tgtd
2015-04-07 15:49:20.764 | + is_ubuntu
2015-04-07 15:49:20.764 | + [[ -z rpm ]]
2015-04-07 15:49:20.764 | + '[' rpm = deb ']'
2015-04-07 15:49:20.764 | + sudo /sbin/service tgtd start
2015-04-07 15:49:20.782 | Redirecting to /bin/systemctl start tgtd.service
2015-04-07 15:49:26.143 | + _create_
2015-04-07 15:49:26.144 | + local vg=stack-
2015-04-07 15:49:26.144 | + local size=10250M
2015-04-07 15:49:26.144 | + local backing_
2015-04-07 15:49:26.144 | + sudo vgs stack-volumes-
2015-04-07 15:49:26.162 | Volume group "stack-
2015-04-07 15:49:26.163 | Cannot process volume group stack-volumes-
2015-04-07 15:49:26.165 | + [[ -f /opt/stack/
2015-04-07 15:49:26.165 | ++ sudo losetup -f --show /opt/stack/
2015-04-07 15:49:26.198 | + local vg_dev=/dev/loop0
2015-04-07 15:49:26.198 | + sudo vgs stack-volumes-
2015-04-07 15:49:26.213 | Volume group "stack-
2015-04-07 15:49:26.213 | Cannot process volume group stack-volumes-
2015-04-07 15:49:26.216 | + sudo vgcreate stack-volumes-
2015-04-07 15:49:26.257 | A volume group called stack-volumes-
2015-04-07 15:49:26.260 | + exit_trap
2015-04-07 15:49:26.260 | + local r=5
2015-04-07 15:49:26.261 | ++ jobs -p
2015-04-07 15:49:26.262 | + jobs=
2015-04-07 15:49:26.262 | + [[ -n '' ]]
2015-04-07 15:49:26.262 | + kill_spinner
2015-04-07 15:49:26.262 | + '[' '!' -z '' ']'
2015-04-07 15:49:26.262 | + [[ 5 -ne 0 ]]
2015-04-07 15:49:26.262 | + echo 'Error on exit'
2015-04-07 15:49:26.262 | Error on exit
2015-04-07 15:49:26.263 | + [[ -z /opt/stack/
2015-04-07 15:49:26.263 | + /home/fedora/
2015-04-07 15:49:26.340 | + exit 5
It's strange because stack-volumes-
fedora@myhost$ sudo vgs stack-volumes-
VG #PV #LV #SN Attr VSize VFree
stack-
fedora@myhost$ sudo losetup
NAME SIZELIMIT OFFSET AUTOCLEAR RO BACK-FILE
/dev/loop0 0 0 0 0 /opt/stack/
Its maybe a bug in the code responsible to detect if the volume group is present. Or a race condition? But how? I'm not running stack.sh in parallel?! (hum, I have an environment variable MAKEFLAGS=-j3 but i don't see how it could be related.)
localrc:
# Configuration for DevStack (used on OpenStack Kilo in 2015)
# Read http://
# Passwords
DATABASE_
ADMIN_PASSWORD=
MYSQL_PASSWORD=
RABBIT_
SERVICE_
SERVICE_TOKEN=token
# By default, all git repositories are updated (OFFLINE is False)
# WARNING: local changes are dropped!
#OFFLINE=True
# stack.sh will freshen each repo on each run
RECLONE=True
# Note: it doesn't work to use icehouse, you should checkout the Icehouse
# Note: branch of DevStack too
#GLANCE_
#HORIZON_
#KEYSTONE_
#NEUTRON_
#NOVA_BRANCH=
#SWIFT_
# Write logs into /opt/stack/
LOGFILE=
# Keep 2 days of log max
LOGDAYS=2
# Disable Heat
disable_service heat h-api h-api-cfn h-api-cw h-eng
# Disable swift, tempest
disable_service swift tempest
# Disable horizon
disable_service horizon
# Disable Cinder
disable_service cinder
# Use neutron for manage networks instead of nova ne7twork
Q_PLUGIN=ml2
disable_service n-net
enable_service q-svc q-agt q-dhcp q-l3 q-meta q-lbaas
Note: I disabled SELinux (/etc/selinux/
tags: | added: openstack-version.kilo |
Changed in devstack: | |
assignee: | nobody → Bogdan Katynski (bkatynski) |
Changed in devstack: | |
status: | In Progress → Fix Released |
i saw this once too... and then it disappeared