josh/AIHostingTycoon
1
0
Fork 0
Code Issues 1 Pull Requests Actions Packages Projects Releases Wiki Activity

Fix Fill Capacity exceeding DC slot limit due to double-counted failed racks
CI / build-and-push (push) Successful in 38s
Details

Browse Source 
computeRacksFailed was incremented on production failure and never decremented
when repaired racks came back online, while repair cohorts also tracked the
same racks. This caused usedSlots to inflate past the DC capacity over time.

Fix: derive computeRacksFailed from repair cohorts each tick instead of
maintaining it as a running counter. Include repair cohorts in pipeline slot
accounting so all racks are counted exactly once. Also fixes power limit in
fillDCToCapacity to only count online racks (pipeline racks don't draw power).

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
Josh Wright
2026-04-25 00:00:46 -04:00
parent 9c49a10b31
commit b7d23c8872
4 changed files with 28 additions and 22 deletions
Download Patch File Download Diff File Expand all files Collapse all files
apps/web/src/pages/InfrastructurePage.tsx
+5 -5
View File
@@ -113,7 +113,7 @@ function DeploymentProgressBar({ dc }: { dc: DataCenter }) {
const tierConfig = DC_TIER_CONFIGS[dc.tier];
const maxCompute = maxComputeRacks(tierConfig.rackSlots);
const pipelineRacks = dc.deploymentCohorts.filter(c => c.stage !== 'decommission').reduce((s, c) => s + c.count, 0);
const totalTarget = dc.computeRacksOnline + dc.computeRacksFailed + pipelineRacks;
const totalTarget = dc.computeRacksOnline + pipelineRacks;
const pct = totalTarget > 0 ? (dc.computeRacksOnline / totalTarget) * 100 : 0;
if (totalTarget === 0 && dc.status === 'operational') return null;
@@ -642,8 +642,8 @@ function DataCenterDetailView({ clusterId, campusId, datacenterId }: {
const tierConfig = DC_TIER_CONFIGS[dc.tier];
const maxCompute = maxComputeRacks(tierConfig.rackSlots);
const existingCompute = dc.computeRacksOnline + dc.computeRacksFailed
+ dc.deploymentCohorts.filter(c => c.stage !== 'decommission').reduce((s, c) => s + c.count, 0);
const pipelineCount = dc.deploymentCohorts.filter(c => c.stage !== 'decommission').reduce((s, c) => s + c.count, 0);
const existingCompute = dc.computeRacksOnline + pipelineCount;
const availableSlots = maxCompute - existingCompute;
const sku = dc.rackSkuId ? RACK_SKU_CONFIGS[dc.rackSkuId] : null;
const netSlots = networkSlotsRequired(existingCompute);
@@ -814,7 +814,7 @@ function DataCenterDetailView({ clusterId, campusId, datacenterId }: {
) : (
<>
<p className="text-sm text-surface-400">
Retrofit swaps all {dc.computeRacksOnline + dc.computeRacksFailed} <span className="text-surface-200">{sku!.name}</span> racks to a new SKU.
Retrofit swaps all {dc.computeRacksOnline + pipelineCount} <span className="text-surface-200">{sku!.name}</span> racks to a new SKU.
The DC goes offline during retrofit.
</p>
<div className="space-y-2">
@@ -906,7 +906,7 @@ function DataCenterDetailView({ clusterId, campusId, datacenterId }: {
{confirmRetrofit && (
<ConfirmModal
title="Confirm Retrofit"
message={`This will decommission all ${dc.computeRacksOnline + dc.computeRacksFailed} ${sku?.name} racks and install ${RACK_SKU_CONFIGS[confirmRetrofit].name}. The DC will go offline during this process.`}
message={`This will decommission all ${dc.computeRacksOnline + pipelineCount} ${sku?.name} racks and install ${RACK_SKU_CONFIGS[confirmRetrofit].name}. The DC will go offline during this process.`}
confirmLabel="Start Retrofit"
onConfirm={() => { retrofitDC(datacenterId, confirmRetrofit); setConfirmRetrofit(null); }}
onCancel={() => setConfirmRetrofit(null)}