| Summary: | Setting dynamic quotas concurrently can lead to cumin reporting > 100% total quota | ||
|---|---|---|---|
| Product: | Red Hat Enterprise MRG | Reporter: | Trevor McKay <tmckay> |
| Component: | cumin | Assignee: | grid-maint-list <grid-maint-list> |
| Status: | CLOSED WONTFIX | QA Contact: | MRG Quality Engineering <mrgqe-bugs> |
| Severity: | low | Docs Contact: | |
| Priority: | low | ||
| Version: | 2.0 | CC: | eallen, eerlands, ltoscano, matt, mkudlej |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2016-05-26 20:14:27 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
|
Description
Trevor McKay
2011-04-14 19:50:33 UTC
There are no safeguards on the backend to prevent accounting group quotas from exceeding 100%. Implications are as yet unknown. Restricting such behavior on the backend may not be desirable. (In reply to comment #1) > There are no safeguards on the backend to prevent accounting group quotas from > exceeding 100%. Implications are as yet unknown. Restricting such behavior on > the backend may not be desirable. FYI: The negotiator does handle this: when the negotiator encounters groups whose dynamic quotas add up to > 1, it will re-normalize them to 1. So, if you define some dynamic quotas [0.5, 0.5, 0.5], they will be normalized to [0.3333, 0.3333, 0.3333], and that is what the negotiator will use. This is recursive. The "add up to <= 1" requirement is checked for all child groups of each group, etc. This is probably a side effect of how cumin gets the values -- they are retrieved through a collection of individual calls rather than a single atomic call that gets all the quota values at once. In that case, normalization would not be the right thing to do -- the numbers are just wrong. Based on Erik's comment above, perhaps the thing to do is retry some number of tiems if the total is greater than 100 and/or flag the numbers with a color or some indicator to the user that they should refresh. MRG-Grid is in maintenance and only customer escalations will be considered. This issue can be reopened if a customer escalation associated with it occurs. |