VMs randomly shutdown alltogether

  • Hey there,

    I've had this problem since using XenServer 7.1, I have this problem on multiple hardware setups (server grade, normal grade, AMD, Intel, ECC & non-ECC RAM), currently experiencing it on XCP-NG 7.6.3 (x2 licensed), XenServer 7.1-7.6

    I troubleshooted this problem a while ago, someone may point me to the correct log files, but I remember the following on this behaviour:

    • All VMs (Linux (CentOS, Ubuntu, Debian), Windows Server and Windows 10) shutdown at the same time (not even a second difference)

    • It happens after about 7-15 days uptime. (It sometimes occured before that, but rarely)

    • I have had no data loss yet

    *When I checked the VM logs back then, there were no shutdown/errors in them. As soon as I started the VMs again they got the log entry for being started. So basically they have 2 times a boot in a row in the logs, no shutdown inbetween.

    *The VMs itselves don't really throw out an error in the logs either, they just suddenly stop logging anything obviously due to them being in a turned off state.

    *any VM-autostart mechanism didn't kick in on this happening, so I recon it's not a full server restart.

    I'm really thankful for every bit of information.

    Thanks in advance!

  • XCP-ng Team


    If you have pro support, it might be more efficient to open a support ticket, so we can take a look remotely 🙂

  • @olivierlambert I will asap, we are still in the payment process and I was advised by Marc-André to post my issue here. If you are up to, we can make a remote session any time. Otherwise I'll wait till the payment is done.

  • XCP-ng Team

    You can already open a support ticket. As a "free" support, it won't raise our alarms (notifying support team, etc.) but we can start to work on your issue anyway 🙂

  • @m4x116: To be sure about the uptime of the whole system, SSH it (or use whatver to console) and do:

    [02:37 localhost ~]# uptime
     02:37:25 up  1:10,  1 user,  load average: 0.02, 0.02, 0.00
    [02:37 localhost ~]#

    Simple as that. Works on (almost?) every Linux based OS. ;>

  • XCP-ng Team

    FYI I did it on the support ticket, and indeed, the host was just online before he restarted all the VMs. But I can find any reason why the host restarted, it might be an "external" power cycle.

Log in to reply

XCP-ng Pro Support