I'm really struggling with the performance of our new on-premise ERP system. We decided to go with a replacement ERP that would remain on-premises rather than switching to cloud solutions. We worked with our managed service provider (MSP) to make necessary storage upgrades, and while the server exceeds the ERP's requirements, users are reporting significant performance issues. Interestingly, the ERP vendor mentioned that our performance is below their expectations. They even provided an AWS instance with lower specs that outperformed our local setup by 60%. We've run network tests showing minimal latency and issues, and monitoring indicates our server has plenty of resources available under load. We even tried direct connections, but those results were worse. Our infrastructure features 1Gbps connections, and there could be a bottleneck at a 1Gbps switch which connects to the server. We're scheduled to consult with HPE next week, but I'm at my wit's end without any clear problem. Any suggestions on where I should focus my efforts next?
5 Answers
It sounds like you might be facing NIC issues on your server or possible bad configurations on your switch port. You could also check if there are MTU settings causing fragmentation. It's worth investigating those areas!
I see you're using a stand-alone server connected to a SAN via 1GbE. Do you have an SNMP network monitoring system in place? If not, I recommend researching tools like LibreNMS or Zabbix. Installing one of these could provide valuable insights into your LAN performance, and you might uncover congestion issues.
Have you thought about using telemetry tools like a logging proxy server or APM tools like Datadog or New Relic? It would be beneficial to measure performance on both your local and cloud setups. Running some A/B tests could help you pinpoint the differences and figure out what’s causing the slowdown.
Which ERP system are you using, and what database is backing it? Make sure to check indexing, whether read committed snapshots are enabled, and if files are pre-allocated correctly. Consider dedicated drives for db, log, and tempdb since that could alleviate some performance issues.
What applications are running on your SAN? It's important to check the IO performance. If you're using virtual machines, have you tried pinning or directly bypassing interfaces? That could help with latency.
Related Questions
Can't Load PhpMyadmin On After Server Update
Redirect www to non-www in Apache Conf
How To Check If Your SSL Cert Is SHA 1
Windows TrackPad Gestures