VM not working suddenly in PVE 7-1.8, works after stop and start for a few hours only.

[AmithKumar]

Hi,

I have upgraded from the 6-4 version to 7-0 a month before and all the VM's were working fine. Since 3 days before I have updated the same machine from 7-0 to the latest version 7-1.8 and the problem started.

Issue:
Suddenly the VM both Linux and Windows will freeze and doesn't work and when we click the console it shows the below error.

VM 102 qmp command 'set_password' failed - unable to connect to VM 102 qmp socket - timeout after 31 retries
TASK ERROR: Failed to run vncproxy.

After force stops and starts it works fine and then after a few hours the same problem.

Work Around:
I have created a new machine with the latest version 7-1.8 and migrated all VM's from the old server to the new server and there is also the same issue.

Kindly guide how to fix this issue.

 

[AmithKumar]

Hi,

Anyone facing similar issues.

 

[AmithKumar]

Hi,

Can anybody support my issue. Today also facing this issue with CentOS 7 VM which was working fine in the older version that is 6.4 without any issues.

 

[MARCELO PINHEIRO MACHADO]

Good evening! Try unchecking the option cache on disks, I had the same problem and only solved by doing that.

 

[AmithKumar]

Good evening! Try unchecking the option cache on disks, I had the same problem and only solved by doing that.

Hi, Thanks for your reply.

I never used the Cache option on the hard disk. It is kept on the Default option (No Cache).
Yesterday I read one post where one member has mentioned changing the Async IO option to native when we used SATA hard disk. I have updated the same and waiting for the issue to get fixed.