I have a delta backup job with Full backup interval set to 7. If i look at the backup job results I get this:

xoce (legion)

    Snapshot
    Start: 2025-07-25 03:02
    End: 2025-07-25 03:02
    Fiend NFS
        transfer
        Start: 2025-07-25 03:02
        End: 2025-07-25 03:08
        Duration: 6 minutes
        Size: 23.78 GiB
        Speed: 73.01 MiB/s
        Clean VM directory
        cleanVm: incorrect backup size in metadata
        Start: 2025-07-25 03:08
        End: 2025-07-25 03:08
    Start: 2025-07-25 03:02
    End: 2025-07-25 03:08
    Duration: 6 minutes

Start: 2025-07-25 03:02
End: 2025-07-25 03:08
Duration: 6 minutes
Type: full

It says "full". If i go to backup -> restore, search for this vm, i see "full 2 delta 2" in "Available Backups" column. Yet if i press "restore" arrow i see that two full backups are of 2022 year and it looks dangerous. Documentation says about that setting "For example, with a value of 2, the first two backups will be a key and a delta, and the third will start a new chain with a full backup.". Yet, I see only VERY old full backups...

Xen Orchestra, commit 19412 Master, commit 6277a, but it seems that this happens for a long time (i recently updated it to this state and the problem seems to persist ever since I began to use Xen Orchestra in 2022).

Do I miss something? Do I really have fresh full backups and these a just labeled "incremental" or what?

@Greg_E "if that host crashes" -- that's the problem. When the host is gone, VMs are gone. Not colored in red 'we are gone', there are none of them.

There is no problem running XO itself on other host.

The question is not how to keep VMs backed -- they are.

But right now, when host X is gone and its VMs are replicated to hosts Y and Z, how can i quickly see that VMs A,B,C and D are gone and E, F, G, etc are up? No way, because X is gone and its VMs gone too.

I have to monitor them via Zabbix to see if they are gone.

But syncing hundreds of VMs with Zabbix is tiresome, it's a lot of manual work.

@olivierlambert Cool. Waiting for that feature

Suppose I have a server with VMs replicated to 3 other different servers (weaker than the main one). Then the main server goes down. I need to assess quickly which copies I should start first. How do I quickly check all the VMs that went down? Some are not important, some are... Without a central list of VMs with their status, it's much trickier.

@olivierlambert I mean if the server goes down, all VMs that were running on that server disappear from XO. Yes, I can create manual entries to monitor them in Zabbix and I do, but for hundreds of VMs, that's a lot of manual work.

Basically, I'd like to see something like a table of VMs showing their status, availability of their replications on other servers, number of backups, etc.

Currently, I'm really falling behind with my Zabbix setup because there's too much manual work involved in tracking all VM IP addresses and other details (especially if some VMs are not easily accessible by IP!). There are no automated tools to properly view the status of each VM - not just up/down status, but comprehensive XO data.

The major drawback of Xen Orchestra is that if the server goes down, all of its VMs are gone too. No clean list of VMs in 'clean' state. Yes, there are backups—lots of them—and replications, but it understandig what's wrong is difficult. If you have hundreds of VMs, how do you track the state of individual VMs: which server each one runs on, which servers have replications, and where backups are stored?

@AtaxyaNetwork Yes, this worked. Thanks.

@AtaxyaNetwork Yes, I do.

I've added two new disks to the server and created md0 raid1 array, as described here: https://xcp-ng.org/docs/guides.html#software-raid-storage-repository. It works. Now I have to move VMs from the main, single disk of the server. But when I click migrate arrow on this server's vm, it does not show the same server as the destination, only other servers.

Yes, I can migrate VMs to another server and back, selecting new raid storage. But it's kinda awkward to haul hefty data between local disks via external machine...

@olivierlambert No, it does not. Still, sometimes badge with patches blink... funny.

@olivierlambert Well, after erasing that file it appears again with the same "check_update" contents.

@burbilog

It seems to return

{"error": "The updater plugin is busy (current operation: check_update)"}

all the time on that server.

@olivierlambert said in Orange exclamation mark without errors:

plugin=updater.py fn=check_update

Where can I run this command? On the host with missing patches?

@olivierlambert Anyway, I don't understand why XOA does not detect them at all

@olivierlambert

64 bytes from 172.16.1.46 (172.16.1.46): icmp_seq=1 ttl=62 time=3.88 ms
64 bytes from 172.16.1.46 (172.16.1.46): icmp_seq=2 ttl=62 time=4.40 ms
64 bytes from 172.16.1.46 (172.16.1.46): icmp_seq=3 ttl=62 time=4.67 ms

@olivierlambert "Missing patches"

Funny thing, one of these hw servers is kinda old with 8.2.0 xcp-ng and there ARE missing patches. But no such notifications neither in XOA or XOCE right now.

Whoa. Just noticed. Red badge with number of patches appear in XOCE! But only after a while, if I leave that browser window for a while. But XOA shows no notification at all.

@olivierlambert Yes, I connect all four machines. No complaints in XOA.

@olivierlambert I did not connect XOA to other hw servers. Is it safe to run both xoa and xoce in parallel on all hw servers?

XOCE shows yellow exclamation mark on Pools menu entry, but there are no errors in host list...

Xen Orchestra, commit 07bd9
xo-server 5.93.1
xo-web 5.96.0

@olivierlambert Installed xoa appliance, connected it to that server and it seems to work, xoa creates new VM just fine. So, it's certainly a problem of xoce.