Cinder doesn't allow to create several volumes in parallel: ERROR: Request Entity Too Large (HTTP 413)
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Mirantis OpenStack |
Invalid
|
High
|
MOS Cinder | ||
6.1.x |
Invalid
|
High
|
MOS Cinder | ||
7.0.x |
Invalid
|
High
|
MOS Cinder |
Bug Description
Steps To Reproduce:
1. Deploy OpenStack cluster with Cinder
2. Login to OpenStack controller node
3. Run the following script:
#______
for i in `seq 1 10`
do
cinder create --display-name test$i 2 &
done
#______
4. Get list of volumes
Expected Results:
All 10 volumes should be created successfully, without any errors. User should see all new 10 volumes in the list.
Observed Results:
Some volumes failed to create, and we get "ERROR: Request Entity Too Large (HTTP 413) (Request-ID: req-0f12b6d5-
Example:
http://
Note:
Looks like it is Cinder API performance issue, cinder failed because of "many" API requests. We will get more errors if we will increase the count of iterations in script.
This bug was reproduced on MOS 6.1 RC3 (#525) ISO. Deployment configuration: CentOS, HA with 3 controllers and 1 compute, with Ceph backend for Cinder, 1 ceph-osd node (on compute node). Snapshot is attached.
Looks like this issue not reproduced on backends with Cinder LVM, only with Cinder+Ceph deployments.
Changed in mos: | |
status: | New → Confirmed |
importance: | Undecided → High |
assignee: | nobody → MOS Cinder (mos-cinder) |
tags: | added: cinder |
summary: |
- Cinder doesn't allow to create several volumes in parallel + Cinder doesn't allow to create several volumes in parallel: ERROR: + Request Entity Too Large (HTTP 413) |
description: | updated |
description: | updated |
Diagnostic snapshot is attached ^^
We can see in the log the following errors: PIException: Bad or unexpected response from the storage volume backend API: error connecting to ceph cluster
VolumeBackendA