VM looking network connect randomly, VNF still active with assigned address but no network activity

[Nathan Platt1709158840]

We have multiple VM's that randomly lose their ability to communicate with the internal network and the external network. The first we know of the problem is when uptimerobot alerts us that the site is offline. 

 

Once we're aware of it, we the following things to bring it back online;

 

• Remove the virtual network adapter, re-add it
• Change the MAC address on the virtual network adapter
• ifdown eth0, ifup eth0
• Migrate to another server

 

Each time we get a XEN Grant_table xxxx still in use error. This isn't limited to one particular machine but so far is affecting on VM more than others.

 

VM details;

 

• Centos 7.9
• Xentools 7-20-1.1
• eth0 connected to a vlan (20, over a 4 port bond)
• Storage on a iSCSI SAN (vlan 10)

 

Any ideas?

[Tobias Kreidl]

Is NTP properly set up and synchronized within  reasonable time offsets (<200 msec)?