| Summary: | Add command-line support for enforce-quorum option | ||
|---|---|---|---|
| Product: | [Community] GlusterFS | Reporter: | Jeff Darcy <jdarcy> |
| Component: | cli | Assignee: | Amar Tumballi <amarts> |
| Status: | CLOSED CURRENTRELEASE | QA Contact: | |
| Severity: | low | Docs Contact: | |
| Priority: | medium | ||
| Version: | pre-release | CC: | gluster-bugs, vijay, vraman |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | Type: | --- | |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
|
Description
Jeff Darcy
2011-11-21 17:49:32 UTC
How about making quorum count configurable too? (In reply to comment #1) > How about making quorum count configurable too? What would the options be? Quorum count (Q) is already determined algorithmically, based on the total replica count (R). Let's examine the cases. * Q=1 is equivalent to no quorum enforcement, except that writers get EROFS instead of EIO. Not sure if this is useful. * Q=2, R=2 (i.e. no special case for first brick) seems marginally useful. * Q=2, R=3 is what we already get for any R=3. * Q=3, R=3 also seems marginally useful. * R>3 will remain too rare to worry about. So both of the marginally useful cases require unanimity. Would it make more sense to have a separate quorum-count option, or allow enforce-quorum (quorum-type) to have more than two values? * "none" = no enforcement * "majority" = what's there now * "all" = require presence of all servers CHANGE: http://review.gluster.com/743 (Change-Id: Ia52ddb551e24c27969f7f5fa0f94c1044789731f) merged in master by Vijay Bellur (vijay) |