Drop cache issue.

[ozgurerdogan]

Sometimes I lost connection to kvm disk and I see in logs :
unable to connect to VM 102 qmp socket - timeout after 31 retries

I was not able to reset kvm and each time, I had to reset node.

Today I found that this issue is happening on heavily fragmented kvm disks. So making defragmentation may help but not totally solve the issue.

I also found that, if I drop cache with following command, disk connection comes back. With out need to restart node.
echo 1 > /proc/sys/vm/drop_caches

So I am wondering if there is a permanent solution that proxmox can make for this issue. I know some people on internet having same issue.

 

[hansm]

Can you post your /etc/pve/storage.cfg and the output of

qm config <vmid>

 

[ozgurerdogan]

/etc/pve/storage.cfg
dir: local2
path /local2
shared 0
maxfiles 2
content rootdir,images,vztmpl,iso,backup

dir: local
path /var/lib/vz
shared 0
maxfiles 2
content backup,iso,vztmpl,rootdir,images

root@vztlfr7:~# qm config 109
acpi: 1
boot: c
bootdisk: virtio0
cores: 2
cpu: host
freeze: 0
hotplug: 0
memory: 3500
name: xxx.xxx.com
net0: rtl8139=02:00:00:ee:92:63,bridge=vmbr0
numa: 0
onboot: 1
ostype: w2k8
sockets: 1
startup: up=60,down=60
virtio0: local2:109/vm-109-disk-1.raw,size=100G