15 lines
4.4 KiB
JSON
15 lines
4.4 KiB
JSON
[
|
|
{"timestamp": "0:04", "title": "Zero Downtime Cluster Update Progress", "summary": "Zero downtime update feature is in very early stages with no reference architecture available yet. GA is unlikely this year. Feature will be for enterprise cluster only, never for standard cluster."},
|
|
{"timestamp": "3:14", "title": "Database Failover and Infrastructure Constraints", "summary": "ST uses the database as configuration engine via Coherence layer. Current limitation: cannot tolerate database loss without misbehavior. Cloud edition will eventually enable better failover tolerance but requires different architecture. Secondary concern is event table synchronization lag between primary and replicated databases."},
|
|
{"timestamp": "9:46", "title": "Database Technology Selection Guidance", "summary": "Choose database based on DBA team comfort level. Oracle preferred due to maturity; MSSQL works well on Linux with good admin team; Postgres is newer option. Database choice matters less for typical loads unless moving >1M files/hour. XML export from accounts is database agnostic."},
|
|
{"timestamp": "12:17", "title": "Cloud Migration and Database Switching", "summary": "Owned/self-managed database recommended over RDS. Switching databases requires performance testing before go-live due to parameter tuning differences. ST switched from Oracle Coherence to open Coherence, eliminating the main reason many had chosen Oracle. Reference environments now use Postgres."},
|
|
{"timestamp": "17:01", "title": "UID/GID Account Changes and File System Impacts", "summary": "Non-root install changes only affect access rules; root install can change file system permissions if user approves. Changes only affect specified account's home folder unless accounts have nested folders. Warning: saving from multiple browser tabs can cause one tab to overwrite another's data."},
|
|
{"timestamp": "20:31", "title": "UID/GID Change Validation Popup Behavior", "summary": "Warning popup on account changes was initially showing every time; after user feedback, limited to actual UID/GID changes. May be caught between versions or overzealous validation on non-root installs. Support ticket recommended to diagnose false positives."},
|
|
{"timestamp": "31:23", "title": "SMB Inbound to Outbound Routing Failure Pattern", "summary": "Files pulled from SMB showing failure recorded as 'outbound' instead of 'inbound'. When AR subscription configured with delete-on-success fails, outbound SMB shows zero KB file size. Requires support system log analysis to diagnose; possible error cleanup causing incorrect recording."},
|
|
{"timestamp": "40:23", "title": "SSH Transfer Failed Despite Vendor Successful Disconnect", "summary": "If outbound shows failed but vendor reports successful transfer: SFTP requires precise closing handshake per RFC. If final confirmation packet not received by ST, ST assumes failure despite vendor completing close. Likely cause is customer application not following RFC for proper SFTP closure."},
|
|
{"timestamp": "44:00", "title": "SMB/Folder Monitor Delays and Event Queue Analysis", "summary": "Delays on SSH outbound after SMB/folder monitor inbound indicate potential SSH connection limit exhaustion or server overload. With enterprise cluster, protective mechanisms leave events in database until stabilization rather than crashing."},
|
|
{"timestamp": "47:25", "title": "Enterprise Cluster Tuning for High-Volume Scenarios", "summary": "Overloaded enterprise cluster leaves events in database for later processing vs standard cluster crashing. Requires tuning thread numbers, TM threads, event queue parameters. Use API endpoint GET /events with small limit/offset to check event queue count. Large queues indicate capacity issues."},
|
|
{"timestamp": "52:45", "title": "Event Queue Monitoring and Server Recovery Patterns", "summary": "High event counts (thousands possible with internal retries) indicate processing delays. Server intentionally self-protects by queuing events instead of crashing. Do NOT restart a stuck server — it flushes the queue and prolongs recovery. Check for future-dated files which cause processing delays."},
|
|
{"timestamp": "57:00", "title": "Docker Containerization and Cloud Edition Strategy", "summary": "ST as container SaaS was abandoned: incompatible with database connectivity architecture. Team working on 'SecureTransport cloud edition' as alternative cloud offering with different foundation. Docker images have restricted availability and are still in evaluation."}
|
|
]
|