Glance had been deleting 4 images more than 7 minutes
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Mirantis OpenStack |
Invalid
|
High
|
MOS Ceph |
Bug Description
Detailed bug description:
It took almost 8 minutes for Glance to delete 4 images. Cluster had background load at that time.
The image IDs were:
46068590-
97117fce-
ce47bec7-
38dca2b2-
Steps to reproduce:
1. Deploy MOS 9.0, 227 build
2. Run rally workload for 4-6 days and make several runs of Rally tests.
3. Try to remove images from rally test runs and workloads
Expected results:
Images has to be deleted pretty quick, but it took enormous amount of time to do that.
Actual result:
It took a lot of time to remove those images.
Reproducibility:
Didn't try
Workaround:
No known workaround.
Impact:
N/A
* 10 baremetal nodes:
- CPU: 12 x 2.10 GHz
- Disks: 2 drives (SSD - 80 GB, HDD - 931.5 GB), 1006.0 GB total
- Memory: 2 x 16.0 GB, 32.0 GB total
- NUMA topology: 1 NUMA node
* Node roles:
- 1 ElasticSearch / Kibana node
- 1 InfluxDB / Grafana node
- 3 controllers (1 was is offline because of disk problems)
- 5 computes
* Details:
- OS: Mitaka on Ubuntu 14.04
- Compute: KVM
- Neutron with VLAN segmentation
- Ceph RBD for volumes (Cinder)
- Ceph RadosGW for objects (Swift API)
- Ceph RBD for ephemeral volumes (Nova)
- Ceph RBD for images (Glance)
* MOS 9.0, build 227
Additional information:
Diagnostic snapshot: http://
Delete was start at 1462348656, finished at 1462349130 (UNIX timestamps)
tags: | added: area-glance |
Changed in mos: | |
importance: | Undecided → High |
status: | New → Confirmed |
Here is results of "top" on both controllers where an issue reproduced: /paste. mirantis. net/show/ 2229/ /paste. mirantis. net/show/ 2230/
https:/
https:/
It looks very suspicious that some processes periodically takes almost 100 CPU on both controllers. This seems the reason why glance-api is so slow.