XCP-ng performance loss vs ESXi
Posted: Sat Mar 02, 2024 10:07 am
Hi,
I'm switching over from VMware ESXi 6.7U2 to XCP-ng 8.2.1 and in this process I'm reusing my old hosts. Blue Iris 5.x have been running on one of my HPE DL380 Gen9 servers for years, with 12 virtual cpu's assigned and 12 GB of RAM. When trying to switch over to XCP-ng the CPU usage went up from 40-50% avg. (no clients) to 80-90% avg. (no clients) with identical vm setup. I imported the same Blue Iris settings and I have tried to find the solution, but it seems like ESXi is just more efficient or supports some other form of passthrough to the physical cpu (2 x Intel E5- 2630 v3). Running 9 Axis cameras with 1080p streams and 4 x Sunba PTZ 601-D20X with 1080p streams. I haven't touched the BIOS/setup on my hosts, so the same conditions apply. I have also tried migrating to a Dell PowerEdge R430 running XCP-ng with identical results.
Any ideas?
I'm switching over from VMware ESXi 6.7U2 to XCP-ng 8.2.1 and in this process I'm reusing my old hosts. Blue Iris 5.x have been running on one of my HPE DL380 Gen9 servers for years, with 12 virtual cpu's assigned and 12 GB of RAM. When trying to switch over to XCP-ng the CPU usage went up from 40-50% avg. (no clients) to 80-90% avg. (no clients) with identical vm setup. I imported the same Blue Iris settings and I have tried to find the solution, but it seems like ESXi is just more efficient or supports some other form of passthrough to the physical cpu (2 x Intel E5- 2630 v3). Running 9 Axis cameras with 1080p streams and 4 x Sunba PTZ 601-D20X with 1080p streams. I haven't touched the BIOS/setup on my hosts, so the same conditions apply. I have also tried migrating to a Dell PowerEdge R430 running XCP-ng with identical results.
Any ideas?