Skip to content

Instantly share code, notes, and snippets.

September 29

Customer Bundle ID Hostname Path
RBC 207967 rbcgithub-fg-rbc-com-primary /data/multi-disk/db
RBC 207868 rbcgithub-fg-rbc-com-replica-babffafcfc /data/multi-disk/db
SAP SE 207830 prod-tools-frontend-sofia-84a78 /data/multi-disk/db
BMW AG 207821 cc-github-bmwgroup-net-replica-fcebfafffe /data/multi-disk/db
BMW AG 207821 cc-github-bmwgroup-net-replica-fcebfafffe /data/multi-disk/git
SAP SE 207647 prod-tools-frontend-athens-7e663 /data/multi-disk/db

MySQL multi-disk during GHES install and upgrade

Summary

Customers should be able to place GHES MySQL data on multi-disk storage at the two moments when the product can make the safest storage decision:

  1. First install: the first-run wizard can guide the admin to attach and confirm an eligible MySQL multi-disk device before initial config-apply initializes MySQL.
  2. Upgrade: existing standalone customers and HA can opt in to moving MySQL to newly attached multi-disk storage during that node's planned upgrade window.

The product outcome is simple: customers get a clearer, safer way to separate database storage from the primary data disk without needing to discover and run a manual storage procedure after the instance is already in service.

Validate Muti-disk setup for BMW

Setup Multi-disk

Create stand-alone ghebooted instance

.gheboot 3.18.10 
#!/bin/bash
#/ Usage: ghe-storage-verify-mysql-migration <backup-path> <destination-path>
#/ Verifies that two MySQL datadir trees are identical by comparing md5sums
#/ of every file beneath each path.
#/
#/ Example (after ghe-storage-migrate-mysql):
#/ ghe-storage-verify-mysql-migration /data/user/mysql-backup/github_enterprise /data/multi-disk/db/mysql/github_enterprise
set -e
func (a *UpgradeGatewayAPIV1) runCluster() http.HandlerFunc {
	return func(w http.ResponseWriter, r *http.Request) {
		clientNodes, _ := connector.AgentClients(r.Context())
		
		// filter out cluster delegate and non-delegate nodes
		var clusterDelegateNode *catalog.Agent
		var nonDelegateNodes []*catalog.Agent
		for _, node := range clientNodes {

🚀 Improving the GHES Manage Developer Experience — What We Did and What's Changed

You've told us this was hard!!
Building new APIs in GHES Manage has too often felt like a challenging expedition: confusing paths, hidden "gotchas," and the occasional moment where you're just… banging your head against the desk wondering why something that should be straightforward is taking so long.

GHES Manage was created three years ago as a modern alternative to the legacy Enterprise Management Console, but adoption hasn't met our expectations. We've heard consistent feedback that implementing new API endpoints can be challenging and unintuitive.

So we've been focused on improving the GHES Manage API.

Fix-It

The following snippet helps to copy script from enterprise2 to a ghebooted instance:

#!/bin/bash
#/ Usage: ./copy.sh <ghe-boot-host-name>
TARGET_HOST="${1}"
TARGET_USER="admin"
TARGET_PORT="122"

set -e

What Operators do

  • Operator manage and configure instance with UI
  • Operator builds automation with API
  • Operator does fire fighting with ssh + bash script

What can the purpose of an API/CLI

  • Part of the UI architecture and support a robust UI solution
  • Help operators build an reliable and efficient automation

Availability Review Retrospect

We initiated a systematic availability review process following our July 2024 offsite (see Revival of the GHES Availability Review Process). The first availability issue was then created on August 16th, marking almost a year since our previous review.

Our journey began by exploring what availability truly means for GHES. We recognized that an escalation's value extends beyond mere resolution - we aimed to foster deeper discussions, prevent recurrence through measured repair items, and share knowledge via comprehensive runbooks.

Over the past 6 months, we've made significant strides in our availability review processes:

  • Created 29 availability issues
  • Generated 29 repair items, with 22 successfully resolved
  • Conducted 9 availability review meetings