Search results

it works, and supported , it's up to you : dedup will not effective between datastores/folders, while namespaces are required to keep whole datastore dedup.

@Unimatrix001 Share hw / conf / filesystem of StorageIBM, is ZFS ? and share destination backup hw and filesystem/configuration.

Not really, because "Stop mode" shutdown VM, do snapshot, start backup, then immediately start VM, backup continue as background, so I/O stress can exist. I was impacted sometimes ago with ZFS on slow 2'5 HDD, some services at VM restart did not start correctly, like SQL... disk was replaced but...

https://forum.proxmox.com/threads/amd-bsod-unsupported-processor-since-windows-build-26100-4202-update-kb5060842-its-preview-kb5058499.166828/post-778103

Nobody can really guess your bottleneck. try disabling writecache. try PVE backup to faster destination, like local disk try disable the other backup

You can't view diff directly from PBS as PBS doesn't know data type itself. You can mount or map (ct/vvm) backup with proxmox-backup-client to manually analyse.

If backup destination is too slow, I/O is delayed up to timeout, then VM crash. Be careful to not overlap within guest backup and pve backup. By the way, PVE do VSS snapshot which, by default, break existing SQL chained backups (diff backup), this topic explain VSS options. There is no problem...

Thanks to changed only data written to backup datastore. Even more faster if you snapshot backup (hot backup) where dirty qemu bitmap allow skip reading whole disk source VM. GC job is required to delete older backups. Verify job is required too as existing data / chunk is never write again...

tried here, bosd during booting iso. waiting official MS refreshed iso to test.

Thanks, works here.

typo , of course is systemctl ( the basis ! ) This will down RAM usage to 75% after some minutes.

indeed 16 GB dedicated to ZFS ARC + high memory usage reaches the limit. try tweaking ksm to start at 75% instead default 80%. nano /etc/ksmtuned.conf KSM_THRES_COEF=25 systemctl restart ksmtuned.service I mean Tuesday was Microsoft Patch Tuesday , where Windows OS download and install updates...

Backups within guests ? Windows Update reboot VM ? Patch Tuesday

ZFS is slow and quickly wearout cheap/consumer SSDs because it writes many more data than actual data (overhead). For memory load, current Proxmox VE intaller version limit ZFS ARC Cache to 10% of Physical Memory.

a manual workaround way is move temporarly unwanted snapshots (their manifest folders ) stored on PBS2 datastore elsewhere. Remote Sync then move back manifest folder

Backup to non-Local Proxmox Backup Server each night ?

I woudn't recommend nested virtualization ...

26200 is based on 26100 but 26200.5570 is April Dev Channel update. June Dev Channel build is 26200.5641 which has same problem as 26100 Normal Channel since 26100.4202

is VM on Windows 26100 build ? is VM cpu type "host" vanilla or custom args in VM .conf ?

if PVE acting as gateway for your VMs, try with a dedicated VM/CT as gateway. if dedicated VM/CT act as gateway, try PVE as gateway