poor performance on LXC instances

[spacebass]

Hey Proxmox friends,
I've got a cluster with some mixed, older server hardware. For the most part, everything works wonderfully - I've loved this ecosystem for years.

Recently, I set up an LXC container and the performance is pretty slow, even unresponsive at time. I'm trying to understand if that's just because I'm using underpowered hardware or if there's a configuration issue I can address. My biggest issue is that the console and SSH are very very slow and the one user-facing service I run is also quite slow.

Host:
* PVE 5.4.34-2
* PVE Manager 6.2-4
*32 core Xeon E5-2500
* 64gb of ram
* 4TB ZFS pool
* 1 LXC
* 4-5 running VMs (FreeBSD and Linux)
* Average load: 2%
* IO Delay ~ 0.50%

LXC:
* Ubuntu 20.04
* 4 cores
* 8gb ram
* Disk is stored on NFS (wondering if this is my performance issue?) - the NFS share is from a Synology on the same subnet with link aggregation
* Average load when in use: 1%
* Memory usage 22%

Anyone have any tips on troubleshooting and hopefully improving performance?

For additional context, this machine is running a Roon server. Roon is an audiophile platform for storing and playing lossless audio to different 'end points'.

 

[oguz]

hi,

Host:
* PVE 5.4.34-2
* PVE Manager 6.2-4

these versions don't match up. can you post your pveversion -v output? 5.x has been EOL for a while now so you should ideally upgrade to the latest versions [0]

* Disk is stored on NFS (wondering if this is my performance issue?) - the NFS share is from a Synology on the same subnet with link aggregation

it very well could be. you can try moving the disk temporarily to a local storage to see if there's any difference. Container -> Resources -> click volume -> Move Volume -> select a local storage as target


[0]: https://pve.proxmox.com/wiki/Upgrade_from_5.x_to_6.0