nova quota-update command is executed without checking if the quota exceed
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Compute (nova) |
Fix Released
|
Medium
|
David Geng |
Bug Description
1. Changed number of the cores from 100 to 15 with the following command:
[root@firstbox11-1 ˜]# nova quota-update --cores 15 2bff6058d6454dc
[root@firstbox11-1 ˜]# nova quota-show 2bff6058d6454dc
+------
| Property | Value |
+------
| cores | 15 |
| floating_ips | 40 |
| gigabytes | 1000 |
| injected_
| injected_files | 5 |
| instances | 100 |
| metadata_items | 128 |
| ram | 204800 |
| volumes | 10 |
+------
2. Perform a deploy using a flavor with 8 CPUs like this:
tm_fl: Memory: 2048MB, VCPUS: 8, Root: 8GB, Ephemeral: 0Gb, FlavorID: 69, Swap: 0MB, RXTX Factor: 1.0, public, ExtraSpecs {}
the deploy fails with the following message:
'Quota exceeded for cores: Requested 8, but already used 47 of 15 cores'
This is an issue, because 47 cores are already used, but the nova quota-update does not check and permits to decrease the cores value under
47.....in alternative if it is possible have a number of cores already used greater than of the cores defined on quota....it is an usability issue because minimum a warning should be issued when nova quota-update is issued.
Changed in python-novaclient: | |
status: | New → In Progress |
Changed in nova: | |
importance: | Undecided → Medium |
tags: | added: grizzly-backport-potential |
tags: | removed: grizzly-backport-potential |
Currently we have two ways to update quotas:
1. nova quota-update
2. nova-manage project quota <tenant-id> .....
Looks like the easy way is update nova-manage to check whether the new quota exceeds the "in_use", but it will lead to inconsistent the above two CLI.
Does anyone has some better solution?