manila in HA does not deploy
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Manila Charm |
Fix Released
|
High
|
Unassigned | ||
OpenStack Manila-Ganesha Charm |
Fix Released
|
High
|
Chris MacNaughton |
Bug Description
When I deploy manila as a single unit, all is fine. If I use 3 units with vip and hacluster as shown in deployment guide, manila stays blocked
manila/0 blocked idle 4/lxd/4 10.0.2.90 8786/tcp Services not running that should be: manila-scheduler, manila-data; Ports which should be open, but are not: 8776
hacluster-
manila/1 blocked idle 5/lxd/3 10.0.2.100 8786/tcp Services not running that should be: manila-scheduler, manila-data; Ports which should be open, but are not: 8776
hacluster-
manila/2* blocked idle 6/lxd/3 10.0.2.95 8786/tcp Ports which should be open, but are not: 8776
hacluster-
manila-ganesha is fine
Unit Workload Agent Machine Public address Ports Message
manila-ganesha/0* active idle 1/lxd/3 10.0.1.67 Unit is ready
hacluster-
manila-ganesha/1 active idle 2/lxd/4 10.0.1.69 Unit is ready
hacluster-
manila-ganesha/2 active idle 3/lxd/3 10.0.1.68 Unit is ready
hacluster-
If I ssh to units I can start manila-scheduler and manila-data manualy, but I cannot get rid of port error.
manila is listening on port 8786
State Recv-Q Send-Q Local Address:Port Peer Address:Port
LISTEN 0 128 127.0.0.53%lo:53 0.0.0.0:* users:(
LISTEN 0 128 0.0.0.0:22 0.0.0.0:* users:(
LISTEN 0 128 [::]:22 [::]:* users:(
LISTEN 0 128 [::1]:11211 [::]:* users:(
LISTEN 0 128 *:8786 *:* users:(
for a comparison, when running single manila unit, it looks the same
State Recv-Q Send-Q Local Address:Port Peer Address:Port
LISTEN 0 128 127.0.0.53%lo:53 0.0.0.0:* users:(
LISTEN 0 128 0.0.0.0:22 0.0.0.0:* users:(
LISTEN 0 128 [::1]:11211 [::]:* users:(
LISTEN 0 128 *:8786 *:* users:(
LISTEN 0 128 [::]:22 [::]:* users:(
manila:
charm: cs:manila
num_units: 3
options:
openstack
vip: *manila-vip
# default-
# that gets mapped into Manila's configuration file. The names are
# proscribed by the share type drivers in Manila. I suspect
# that we could remove this configuration, but the Manila charm
# currently goes into a blocked state without it set, and is broken
# if set incorrectly. In essence, to use CephFS with Ganesha, it must
# be set to cephfsnfs1
default-
share-
bindings:
"": *oam-space
public: *public-space
admin: *admin-space
internal: *internal-space
shared-db: *internal-space
to:
- lxd:1003
- lxd:1004
- lxd:1005
tags: | added: cdo-qa |
Changed in charm-manila: | |
status: | Incomplete → Confirmed |
Changed in charm-manila: | |
importance: | Undecided → High |
Changed in charm-manila-ganesha: | |
assignee: | nobody → Chris MacNaughton (chris.macnaughton) |
importance: | Undecided → High |
Changed in charm-manila: | |
milestone: | none → 20.05 |
Changed in charm-manila: | |
status: | Fix Committed → Fix Released |
Changed in charm-manila-ganesha: | |
milestone: | none → 21.04 |
status: | Fix Committed → Fix Released |
Hi Marian
Please could you add to the bug report:
- The bundle used for deploying the model.
- Any relevant logs from the affected units.
Many thanks