Add chapters for session 961596767 (Haiku)

chapters/961596767.json

@@ -0,0 +1,14 @@
+[
+  {"timestamp": "3:31", "title": "SFTP File Rename on Delivery Conditional to Existence", "summary": "SFTP does not allow proactive server-side checks before pushing files. Pull from partner step cannot integrate into same flow without disconnecting routes. Solution requires daisy-chaining multiple routes, making implementation complex. Recommendation: negotiate with partner to standardize naming or add timestamp."},
+  {"timestamp": "10:08", "title": "SFTP Connection Packet Size Mismatch Error", "summary": "Packet size violation errors indicate buffer/block size mismatches — adjust SFTP block size in transfer settings. If connection test fails even on edge, likely OpenSSH vs Maverick RFC compliance issue. ST is RFC-strict; some servers use non-compliant features. Test directly from edge via SSH first to diagnose."},
+  {"timestamp": "19:31", "title": "Finding Certificates Attached to Permanent Settings", "summary": "No simple UI method exists. Search server configuration with certificate name to find all references. Admin certificate is always called 'admin'; MDN must be called 'MDN' for disposition notifications. XML export reveals all certificate usages in transfer sites and steps."},
+  {"timestamp": "21:40", "title": "Minimum Certificate Requirements and Replacement", "summary": "Minimum: one admin certificate. With edges: at least two (admin + one for edge-server communication). Replace certificates in-place to auto-update everywhere; keep the name unchanged. Certificate changes on edges must also be updated on edge servers or functionality breaks."},
+  {"timestamp": "25:06", "title": "Transfer Stuck in Canceled State After Upgrade", "summary": "Observed after February patch 32.04. Check event queue for lingering canceled transfers — indicates corrupted tracking table record or event still in queue. Verify previous update did not fail without rollback, then contact support with event queue details for database-level diagnosis."},
+  {"timestamp": "29:46", "title": "Authentication: Password vs SSH Key Security", "summary": "SSH keys always more secure than passwords. ST supports dual authentication (key + password). Set listener to 'optional' for password; set client certificate to 'optional' to allow either key or password fallback. Multiple listeners on different ports enable different auth requirements per entry point."},
+  {"timestamp": "34:30", "title": "Database Partition Strategy for Enterprise Clusters", "summary": "Partitions are date-based in the tracking table. Prebuild 10-14 day partitions to avoid creation delays at midnight when switching dates. Partition creation happens on first write after midnight; if not prebuilt, system pauses for creation. Avoid scheduled maintenance at midnight due to high system load."},
+  {"timestamp": "40:16", "title": "Upgrade Paths and Backward Compatibility Issues", "summary": "Large version jumps encounter cipher/algorithm retirement. Security is tightened; older customers may face connectivity loss (~10%). Max 12-month direct upgrades supported (e.g., Feb 2022 → Jan 2023 → Dec 2023, two hops required). Read release notes for retired ciphers. In-place updates preserve configurations."},
+  {"timestamp": "43:36", "title": "Transfer Site Cipher Defaults After Upgrade", "summary": "Transfer sites with custom cipher lists retain those settings. Sites saved empty inherit server defaults, which change on upgrade — empty-cipher sites may fail to connect after upgrade. Fix by re-saving the site to apply new defaults. Monitor push/pull connectivity issues post-upgrade."},
+  {"timestamp": "47:44", "title": "OS Upgrade from RHEL 7 to RHEL 9", "summary": "Do not attempt in-place upgrade between RHEL 7 and 9. All cluster nodes must run same RHEL version. Build new environment on target RHEL, migrate configurations. Plugins are OS-agnostic; custom components require OS-level dependency review from original builders."},
+  {"timestamp": "50:00", "title": "DR Passive Environment Setup with Shared Database", "summary": "Official procedure: install DR cluster as separate independent cluster against own database following documentation exactly. After complete installation, repoint to shared read-only replica database and replicated file system at final step. Gluster supported for home folders only."},
+  {"timestamp": "52:28", "title": "DR Update Procedure with Read-Only Database", "summary": "Two parameters prevent issues: skip database update (replica updates automatically with production), skip service start (read-only prevents startup). Use command-line installer flags (DST_DB_update, start_services parameters). Do not break replication during update unless documented as required."}
+]