DaanHoogland opened a new pull request, #13951:
URL: https://github.com/apache/cloudstack/pull/13951
# Scope host tag-rule matches to deployment plan in FirstFit/Random
allocators
## Problem
`FirstFitAllocator` and `RandomAllocator` both call
`HostDao#findHostsWithTagRuleThatMatchComputeOferringTags(String)` to pick
up hosts whose
host tag is defined as a *rule* (a boolean/regex expression, e.g. `!ssd`)
rather than a
literal string, so they match the compute offering's host tag.
That DAO method evaluates every rule-tagged host in the **entire
installation** against the
offering tag and returns all matches, with no zone/pod/cluster filtering.
The allocators then
unconditionally `addAll`/`union` those hosts into the candidate list for the
current
deployment plan:
```java
// before — unscoped: can return hosts from any zone/pod/cluster
clusterHosts.addAll(_hostDao.findHostsWithTagRuleThatMatchComputeOferringTags(hostTagOnOffering));
```
As a result, a host in a different datacenter/pod/cluster than the one being
deployed into —
but whose rule tag happens to match the offering's host tag — can be added
as an allocation
candidate. Once storage-pool selection is scoped correctly but the host list
isn't, the two
diverge and valid host+pool pairings that should have worked can be
rejected, or (in the
FirstFitAllocator path) a host outside the deployment's scope can be picked
entirely.
This reproduces most easily in a setup combining host tags and storage tags
with more than
one pod/cluster/zone and a host tag configured as a rule.
## Fix
- Added a scoped overload on `HostDao`:
`findHostsWithTagRuleThatMatchComputeOferringTags(String
computeOfferingTags, Long clusterId, Long podId, Long dcId)`,
implemented in `HostDaoImpl` by reusing the existing rule evaluation and
filtering the
matched hosts down to the given zone/pod/cluster.
- `FirstFitAllocator` (both `allocateTo` overloads) and `RandomAllocator`
now call the scoped
overload with the deployment plan's `dcId`/`podId`/`clusterId` instead of
the unscoped one.
- On `main`, the rule-tag lookup lives in the shared
`BaseAllocator#addHostsBasedOnTagRules`
helper (used by both allocators); that helper and its two call sites were
updated the same
way.
```java
// after — scoped to the deployment plan
clusterHosts.addAll(
_hostDao.findHostsWithTagRuleThatMatchComputeOferringTags(hostTagOnOffering,
clusterId, podId, dcId)
);
```
The old unscoped overload is left in place since it's still the correct
primitive for the
handful of call sites (e.g. `findClustersThatMatchHostTagRule`) that
intentionally look
system-wide.
## Testing
Added/extended unit tests to assert the allocators call the scoped DAO
overload (and never
fall back to the unscoped one) when resolving rule-tagged hosts:
- `FirstFitAllocatorTest` (new on 4.20/4.22; extended on main)
- `RandomAllocatorTest`
- `BaseAllocatorTest` (main only, covers the shared helper directly)
### Types of changes
- [ ] Breaking change (fix or feature that would cause existing
functionality to change)
- [ ] New feature (non-breaking change which adds functionality)
- [ ] Bug fix (non-breaking change which fixes an issue)
- [ ] Enhancement (improves an existing feature and functionality)
- [ ] Cleanup (Code refactoring and cleanup, that may add test cases)
- [ ] Build/CI
- [ ] Test (unit or integration test code)
### Feature/Enhancement Scale or Bug Severity
#### Feature/Enhancement Scale
- [ ] Major
- [ ] Minor
#### Bug Severity
- [ ] BLOCKER
- [ ] Critical
- [ ] Major
- [ ] Minor
- [ ] Trivial
### Screenshots (if appropriate):
### How Has This Been Tested?
<!-- Please describe in detail how you tested your changes. -->
<!-- Include details of your testing environment, and the tests you ran to
-->
#### How did you try to break this feature and the system with this change?
<!-- see how your change affects other areas of the code, etc. -->
<!-- Please read the
[CONTRIBUTING](https://github.com/apache/cloudstack/blob/main/CONTRIBUTING.md)
document -->
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
