@pierrebrunet said: Another thing to check is the exact time between your XO and your host. I installed the host with NTP from "static" DHCP (OK according to https://docs.xcp-ng.org/installation/requirements/) not made any manual changes/configs After I did a RTC adjustment in BIOS 3 days ago, the time now is ~1s ahead This is the output from timedatectl [22:04 x2 ~]# timedatectl status Local time: tor 2026-07-16 22:06:45 CEST Universal time: tor 2026-07-16 20:06:45 UTC RTC time: tor 2026-07-16 20:06:42 Time zone: Europe/Stockholm (CEST, +0200) NTP enabled: yes NTP synchronized: no RTC in local TZ: no DST active: yes Last DST change: DST began at sön 2026-03-29 01:59:59 CET sön 2026-03-29 03:00:00 CEST Next DST change: DST ends (the clock jumps one hour backwards) at sön 2026-10-25 02:59:59 CEST sön 2026-10-25 02:00:00 CET NTP synchronized: no I thought that the time would be adjusted automatically at reboot The first NTP sync should allow a bigger step than when it's already synced. @pierrebrunet said: do you make these backup regularly? They run once a day @ 03:04 before the normal backup that starts @ 03:10 and they have all run OK since July 8, the failed one reported 8 days ago ( post ~85 ) Maybe you should resync your clock NTP is enabled but not synced, I don't know the correct way to sync it In xsconsole/ I tried Use DHCP NTP Servers It changed the RTC clock to the same as UTC time but time was still 1s ahead and NTP not synced. I rebooted the host and I got this: [23:36 x2 ~]# timedatectl status Local time: tor 2026-07-16 23:37:17 CEST Universal time: tor 2026-07-16 21:37:17 UTC RTC time: tor 2026-07-16 21:37:16 Time zone: Europe/Stockholm (CEST, +0200) NTP enabled: yes NTP synchronized: no RTC in local TZ: no DST active: yes Last DST change: DST began at sön 2026-03-29 01:59:59 CET sön 2026-03-29 03:00:00 CEST Next DST change: DST ends (the clock jumps one hour backwards) at sön 2026-10-25 02:59:59 CEST sön 2026-10-25 02:00:00 CET NTP still not synced, Local time 2s ahead, RTC 1s ahead, I don't know, If not someone can show me the proper way to manually resync, maybe I should try to reinstall the host

@ravenet Sure, gave that a whirl and no change, though I did notice some nvtop weirdness as it would show some load, but most of the wait time there was actually no load on the GPUs instead of the constant load matching the model being loaded. For some historical context ARI support was initially disabled in the bios when I started this thread. That was on the list of things I enabled when I started seeing some success with ollama (something in the changes since has broken ollama now too, but there was at least some forward progress after enabling). dmesg output overall looked the same, but I did see this output on the console (and in dmesg) that seemed interesting. Not 100% sure at this point if this was in the previous dmesg outputs or not, but may be worth sharing. EDIT: looks like this may actually be new... I looked back through the past dm dmesg outputs and I did not see this output. [ 108.547683] amdgpu 0000:00:09.0: MES(0) failed to respond to msg=REMOVE_QUEUE [ 108.547729] amdgpu 0000:00:09.0: failed to remove hardware queue from MES, doorbell=0x1202 [ 108.547746] amdgpu 0000:00:09.0: MES might be in unrecoverable state, issue a GPU reset [ 108.547774] amdgpu 0000:00:09.0: Failed to evict queue 2 [ 108.547789] amdgpu 0000:00:09.0: Failed to evict process queues [ 108.547803] amdgpu: Failed to quiesce KFD [ 108.547870] amdgpu 0000:00:09.0: GPU reset begin!. Source: 3 [ 109.656850] amdgpu 0000:00:09.0: Failed to remove queue 0 [ 109.657324] amdgpu 0000:00:09.0: Dumping IP State [ 109.756265] amdgpu 0000:00:09.0: Dumping IP State Completed [ 112.047695] amdgpu 0000:00:09.0: MODE1 reset [ 112.047797] amdgpu 0000:00:09.0: GPU mode1 reset [ 112.054505] amdgpu 0000:00:09.0: GPU smu mode1 reset [ 113.075393] amdgpu 0000:00:09.0: GPU reset succeeded, trying to resume [ 113.090354] amdgpu 0000:00:09.0: [drm] PCIE GART of 512M enabled (table at 0x00000087D6B00000). [ 113.092905] amdgpu 0000:00:09.0: [drm] AMDGPU device coredump file has been created [ 113.092913] amdgpu 0000:00:09.0: [drm] Check your /sys/class/drm/card1/device/devcoredump/data [ 113.092917] amdgpu 0000:00:09.0: VRAM is lost due to GPU reset! [ 113.092921] amdgpu 0000:00:09.0: PSP is resuming... [ 114.997591] amdgpu 0000:00:09.0: GECC is disabled, set amdgpu_ras_enable=1 to enable GECC in next boot cycle if needed [ 115.091125] amdgpu 0000:00:09.0: RAP: optional rap ta ucode is not available [ 115.091130] amdgpu 0000:00:09.0: SECUREDISPLAY: optional securedisplay ta ucode is not available [ 115.091134] amdgpu 0000:00:09.0: SMU is resuming... [ 115.091375] amdgpu 0000:00:09.0: smu driver if version = 0x0000002e, smu fw if version = 0x00000033, smu fw program = 0, smu fw version = 0x00684c00 (104.76.0) [ 115.432371] amdgpu 0000:00:09.0: SMU is resumed successfully! [ 115.445556] amdgpu 0000:00:09.0: program CP_MES_CNTL : 0x4000000 [ 115.445731] amdgpu 0000:00:09.0: program CP_MES_CNTL : 0xc000000 [ 115.750240] amdgpu 0000:00:09.0: [drm] DMUB hardware initialized: version=0x0A000800 As an update for other things I have tried, in order to eliminate hardware issues, or bios settings I tried installing proxmox and spinning up a VM there with both GPUs passed through and it worked just fine... With that feedback I did a fresh install of XCP-NG 8.3 and spun up a fresh VM using the same steps as I used on proxmox and still no dice. This leads me to believe the issue is somewhere in the XCP-NG passthrough stack with my specific hardware...

Also a question for @Team-XAPI-Network

@pierrebrunet Hello! I've checked B2 logs and found some errors like you had foreseen. I had a look and this kind of 500 INTERNAL ERROR several times on other jobs that seem ok. How many retries does XO attempt to send the part when this happens? I'll do another run that includes only a few large VMs. BTW, is there a way to interrupt the job withou restarting XO ? REST.PUT.PART Weekly/xo-vm-backups/60b4d2a7-3f91-f0f9-b0a0-f834df3ff9ee/20260705T070618Z.xva "PUT /zzzzzkly/xo-vm-backups/60b4d2a7-3f91-f0f9-b0a0-f834df3ff9ee/20260705T070618Z.xva?partNumber=3954&uploadId=4_ze092693d1b65962e9fe6051b_f25969add7e42e384_d20260713_m140016_c003_v0312040_t0013_u01783951216751&x-id=UploadPart HTTP/1.1" 500 INTERNAL_ERROR 145 - 3762 3762 "-" "aws-sdk-js/3.1065.0 ua/2.1 os/linux#5.15.0-185-generic lang/js md/nodejs#24.18.0 api/s3#3.1065.0 m/N,a,c,E,e xen-orchestra-fs-4.9.2" - - - TLS_AES_256_GCM_SHA384 AuthHeader s3.eu-central-003.backblazeb2.com TLSv1.3 - -

@christopher-petzel Thanks!

@olivierlambert said: I'm not sure it's a great idea, because even 64 hosts is huge in terms of VMs and fallout if you have a problem on your pool DB. Even if you drastically improve the current mechanism, the impact of a problem pool wide is far bigger with 64 hosts than 24 for example. It's more than purely tech, it's also a tech design/choice. Exactly, the failure domain just grows exponentially when going beyond a certain collective pool size. I was doing some more research and I found a few possible choices that Tesco may have moved too, but don't want to detract from this forum here.

So if i set VMs to Best-effort or disabled for HA i get not enough memory. @pkgw Do you have Twinstore configured and/or HA enabled? If no I will continue my issues at the twinstore post. If i set to restart i get this.... host.setMaintenanceMode { "id": "35233210-4e37-4703-9bf6-9e8a9c24df9f", "maintenance": true } { "code": "HA_OPERATION_WOULD_BREAK_FAILOVER_PLAN", "params": [], "call": { "duration": 5, "method": "host.disable", "params": [ "* session id *", "OpaqueRef:d9bd976f-445c-8244-1b56-879f446efc12" ] }, "message": "HA_OPERATION_WOULD_BREAK_FAILOVER_PLAN()", "name": "XapiError", "stack": "XapiError: HA_OPERATION_WOULD_BREAK_FAILOVER_PLAN() at XapiError.wrap (file:///opt/xen-orchestra/packages/xen-api/_XapiError.mjs:16:12) at file:///opt/xen-orchestra/packages/xen-api/transports/json-rpc.mjs:38:21 at runNextTicks (node:internal/process/task_queues:65:5) at processImmediate (node:internal/timers:472:9)" }

@MajorP93 said: What you are describing must be a storage live migration then (XenMotion). That should work across pools, yes. I stand corrected because I did not factor in this variant in my initial answer. Yes, that's it - sorry - I should have been more explicit, even if 'lumbers along' wasn't a hint //EDIT: also there is no such thing as non-pool hosts. Even a single host setup has it's own pool. Pedantic, but true I technically meant 'not hosts in the source pool' I guess.

@simonp , thank you very much and all of your team ! ! ! Keep up the outstanding work. Regards, Marc

@kagbasi-wgsdac said: @poddingue Bug report filed as requested — https://github.com/xcp-ng/xcp/issues/825 — and tagging @Team-Storage per your suggestion. Full evidence bundle is attached to the issue (versions, sweep output, vhd-util vs xe comparison, SMlog). Summary of what I found: One correction to the mechanism, and I think it matters. The recap describes is-a-snapshot being flipped to true. On my system that isn't what's happening — is-a-snapshot is false on every affected VDI. The field being wrongly written is snapshot-of, which is getting populated on base disks that aren't snapshots at all. XO's disappearing-disks symptom is consistent with either (it filters on a non-empty snapshot-of), but if the storage team is hunting for a bad is-a-snapshot write, that may be the wrong field. Every affected VDI here looks like: is-a-snapshot: false <-- correct snapshot-of: <populated with an unrelated VDI's UUID> <-- wrong A VDI that is a snapshot of itself. The clearest single artifact: uuid: 806f7f42-083f-4a40-b3f1-0700d00bab5a name-label: WinSrv2022SHB_Disk1_Data is-a-snapshot: false snapshot-of: 806f7f42-083f-4a40-b3f1-0700d00bab5a <-- itself snapshot-time: 20260709T11:19:15Z sm-config: vhd-parent: c86e3247-... <-- bears no relation to the snapshot-of value No valid code path produces snapshot-of = self. Whatever writes this field isn't validating the target. It's still actively corrupting new VDIs — this is not just legacy damage. That self-referential VDI was created 2026-07-09, a week after my patch + reboot. Sweeps 9 days apart went from ~180 → 191 affected VDIs on one SR, and a fourth anchor UUID appeared that didn't exist in the first sweep. Newly created VHDs keep landing in the affected set. So "stop it happening again" is the urgent half of the two-part fix, at least in my case. The bogus targets cluster onto a tiny anchor set, and the anchors point at each other: Count Anchor 97 937c3945 (→ a893fdb4) 50 a893fdb4 (→ ea150883) 37 ea150883 7 806f7f42 (→ itself, new since Jul 9) That looks less like corrupted lineage and more like the field being filled from an incorrect/uninitialised source. On-disk VHDs are completely healthy. vhd-util check says valid, parent locators are consistent, GC reports no work. The two VDIs the DB calls parent/child are, on disk, siblings under a common parent. The corruption is purely in the XAPI database — which is good news for recoverability. The VDI_IN_USE is not a real lock. current-operations is empty, xe task-list is empty, no tapdisk holds it. VM.start fails because it's walking a snapshot relationship that doesn't exist on disk. Reproduces from xe on the pool master with XO entirely out of the path — which is why I filed against xcp-ng/xcp rather than the XO tracker. Versions: XCP-ng 8.3.0, xapi 26.1.11 (xapi-core-26.1.11-1.2), sm-3.2.12-17.9, sm-fairlock-3.2.12-17.9, blktap 3.55.5-9.1, build 20260618. I have not attempted to bulk-clear the fields — on-disk data is intact and I'd rather not do a mass write against the XAPI DB on a live SR without guidance. Backing store snapshotted as a safety net. Happy to run whatever diagnostics would help. And +1 to the hand-grenade feeling — the affected set growing on its own is the part that worries me. Has this issue been validated on a storage server built around Debian 13, LVM and ext4 or just TrueNAS when connected to XCP-ng version 8.3.0. As part of the XAPI DB corruption. Can anyone answer this please or give a clue?

Your write-up is the kind of thing that saves the next person a weekend. One thing worth knowing before you commit to the whole-controller route. The passthrough page has a "Passing through Keyboards and Mice" section further down, and it says XCP-ng ships /etc/xensource/usb-policy.conf with DENY rules for mice and keyboards by default. You edit those to ALLOW, then refresh with /opt/xensource/libexec/usb_scan.py -d followed by xe pusb-scan host-uuid=<host_uuid>. It's at https://docs.xcp-ng.org/compute/ under USB Passthrough. I have no idea whether that covers your USB-to-serial adapter, which is a different device class, and passing the whole controller may still be the cleaner setup for two discrete workstations anyway. Might be worth a mention to @Team-Hypervisor-Kernel on the display question, because that one still puzzles me.

Indeed, it should be detected/reported correctly to avoid losing time finding the problem. Let me ping @julienXOVates

@Danp said: Smart Reboot option found on the host's Advanced tab does what you are asking Very nice!

@TrapoSAMA said: @john.c I have this iso but how to know what is de package? Thx It’s going to be an msi installer package or on the guest tools iso, which is on the XCP-ng or Xen Server your running likely in Local Storage SR or Shared Storage SR (if moved). Named the same as the file linked to above, likely. Anyway did you know that the UEFI SecureBoot certificates (at least one of them from 2011) issued by Microsoft are expiring fully during October 2026. Around that time likely at sometime after then those certificates will be placed into dbx. This means the Windows Server 2012 R2 if operating as a UEFI SecureBoot VM, will if restarted (or reboot following a crash), fail to boot. So your client will go from having a working app that they can use, but not upgrade at the moment to one which won’t work and can’t be used! They really need to upgrade or migrate from Windows to Linux (released during 2025 or 2026) along with the app if needed. Alternatively upgrade to Windows Server 2022 or 2025 and the app!

@the_jest Not showen in this picutre but this is where the message would be displayed. Next to the name of the host... [image: 1782931263879-screenshot-2026-07-01-144023.png]

For your storage question, it's fully explained in the doc: https://docs.xcp-ng.org/storage/#-how-to-modify-an-existing-sr-connection And yes, it's planned to get the complete error visible in XO, sadly, it's not "obvious" since the error message isn't returned by XAPI when you try to start but by another method we need to call after it fails ("assert can be started here" from the top of my head). Let me ping @julienXOvates

That it's stayed quiet since you throttled Velero is a good sign, and it lines up with the concurrency theory rather than anything about the specific volumes. Bumping dom0 RAM makes sense too, since linstor-satellite spiking to roughly 8,700 log lines around the panic suggests dom0 was under real pressure right then. If it does come back, a xen-bugtool --yestoall bundle from the host that panicked, plus the Loki window you already have, would give @Team-Storage something concrete to line up against the DRBD side. I'm not sure whether the real fix sits in Velero's pacing or in how LINSTOR handles concurrent create/delete, but the reproduction you've narrowed down is genuinely useful. Fingers crossed it stays boring from here.

@acebmxer I can confirm that, we'll fix it, thanks

Hi @dvdwx, RBAC is available only at the REST API level, not through the XO6 UI. UI compatibility with RBAC is currently under development. For the topology, this is something that has been here for a long time, so I may be wrong since I don't have the full history. However, if I look at the code, creating the topology selector requires knowing how many CPUs are available in the pool. Since a self-service user doesn't have visibility into the pool, they can't access this kind of information.