Hello,
We have a cluster running PVE3.3 with kernel 3.10.0-7-pve on Intel Xeon E5v2 processors. Each host has between 256GB and 512GB of ECC RAM.
We have been seeing a large number of "swap_dup: Bad swap file entry" errors in the syslog of our Proxmox hosts, and occasional complete lock up of the hosts requiring a reboot.
We initially thought this was due to swappiness being set to 60, so adjusted this to 0 and flushed swap back in to memory. However the problems have continued, even though swap usage is now at 0MB we still get these errors.
We rebooted a host so that it had no usage, and live-migrated on 4 VM's, as soon as we live migrated the second VM we started to see the log filling with these errors.
Our VM's are primarily Windows 2008r2 and 2012r2 with the 0.1-94 and 0.1-100 virtio drivers and the balloon driver.
Has anyone else seen this problem before ?
We have a cluster running PVE3.3 with kernel 3.10.0-7-pve on Intel Xeon E5v2 processors. Each host has between 256GB and 512GB of ECC RAM.
We have been seeing a large number of "swap_dup: Bad swap file entry" errors in the syslog of our Proxmox hosts, and occasional complete lock up of the hosts requiring a reboot.
We initially thought this was due to swappiness being set to 60, so adjusted this to 0 and flushed swap back in to memory. However the problems have continued, even though swap usage is now at 0MB we still get these errors.
We rebooted a host so that it had no usage, and live-migrated on 4 VM's, as soon as we live migrated the second VM we started to see the log filling with these errors.
Our VM's are primarily Windows 2008r2 and 2012r2 with the 0.1-94 and 0.1-100 virtio drivers and the balloon driver.
Has anyone else seen this problem before ?