Some openstack logs is empty on UI
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Fix Released
|
High
|
Fuel Python (Deprecated) | ||
5.0.x |
Won't Fix
|
High
|
Fuel Python (Deprecated) |
Bug Description
"build_id": "2014-05-
"mirantis": "yes",
"build_number": "212",
"ostf_sha": "353f918197ec53
"nailgun_sha": "ab7f7dfddadfe0
"production": "docker",
"api": "1.0",
"fuelmain_sha": "9de65bfdb7e8bc
"astute_sha": "a3432e6e31ffd6
"release": "5.0",
"fuellib_sha": "b4671dcaa93d45
1. Create new environment (Ubuntu, HA)
2. Choose GRE segmentation
3. Choose Ceph for images
4. Add 3 controllers, compute, cinder, 2 ceph nodes
5. Start deployment. It was successful
6. heat-engine logs for controllers is empty on UI (see screen)
7. But heat-engine logs in /var/log/
Response in firebug:
{"has_more": false, "to": 1639, "entries": []}
Also in other deployment ceilometer-
Changed in fuel: | |
importance: | Undecided → High |
summary: |
- openstack logs is empty on UI + Some openstack logs is empty on UI |
Changed in fuel: | |
status: | New → Confirmed |
no longer affects: | fuel/5.1.x |
The issue is in our regexp matching. The Nailgun tries to parse OpenStack logs with next regexp
https:/ /github. com/stackforge/ fuel-web/ blob/master/ nailgun/ nailgun/ settings. yaml#L82
while some openstack services has another format. For example, heat-engine:
2014- 05-20T13: 25:30.718723+ 00:00 info: INFO: OpenStack Heat Engine started
Since each OpenStack service might have own logging settings, I don't know how to fix in better way.
I see the following ways:
1/ Add new regexp to settings.yaml for the services with a new logging format. Bad solution, IMO.
2/ Change puppet manifests to configure all OpenStack services to log with a certain log format, so we can parse it with one regex.
3/ any suggestions?