Large memory VM timeout on Start?

[daemonix]

Hi all,

I have a 1Tb RAM, dual EPYC 7543 system running Linux 5.15.74-1-pve #1 SMP PVE 5.15.74-1
(I cant update at the moment due to a project but its just an internal box with no externally accessible points)

I can start the VM below with up to 300-ish GB RAM. But times out with 400+GB and my target is to have this running with 500GB.
I checked the forum and I even tried changing the timeout from terminal (qm start 141 --timeout 400) but got nowhere.
I also set numa awareness to 1 but nothing changed. Also, I dont care if memory is balluned or not. The server is user with zero over-provisioning, its underutilised.

Any pointers or RTFM?

Thanks

qm config 141
agent: 1
balloon: 64000
boot: order=scsi0;net0
cores: 24
cpu: host
hostpci0: 0000:03:00.0,rombar=0
hostpci1: 0000:04:00.0,rombar=0
hostpci2: 0000:43:00.0,rombar=0
hostpci3: 0000:44:00.0,rombar=0
hostpci4: 0000:83:00.0,rombar=0
hostpci5: 0000:84:00.0,rombar=0
hostpci6: 0000:c3:00.0,rombar=0
hostpci7: 0000:c4:00.0,rombar=0
memory: 300000
meta: creation-qemu=7.0.0,ctime=1664571684
name: u2204Ju
net0: virtio=1E:ED:41:2C:03:04,bridge=vmbr1
numa: 1
ostype: l26
scsi0: Disk8T0:141/vm-141-disk-0.qcow2,size=350G
scsihw: virtio-scsi-pci
smbios1: uuid=06db4a87-3d4f-1515-5757-efd5195c665f
sockets: 2
vmgenid: fc6230ca-713e-48cb-9be2-75c4aad6b992

 

[fiona]

Hi,
how fast does the VM start booting with 300 GiB? How much (free) memory does your host have? Do you use ZFS or something else that might require a lot of memory (for ZFS, see)?

Maybe the 400 seconds are simply not enough, see: https://git.proxmox.com/?p=qemu-server.git;a=commit;h=2e6ea194efb67fa66b32166b2102817ace990426

 

[daemonix]

With the 300GB running I have 500Gb+ free. No ZFS on this system.

Do I really use a high enough timeout to mitigate the problem? I havent timed the 300Gb startup :S (is there a log about it?)

 

[daemonix]

timeout of 2100 (sec) worked!