22.8.14

Proxmox: restore cluster without rebooting

For some reason in these days I faced a problem with proxmox GUI and cluster management.
From one node of the cluster I saw on /var/log/syslog file that
 pveproxy[88540]: WARNING: proxy detected vanished client connection  
and if I try to access to the Web GUI of that node, it says "Login Incorrect".
From other nodes, it seems all right, but it isn't.
The solution is to simulate a reboot of all the machines without rebooting the VM on them: powerful!
So this is the sequence to launch on every node of the cluster, without some particular order
 /etc/init.d/cman restart  
 /etc/init.d/pvedaemon restart  
 /etc/init.d/pvestatd restart  
 /etc/init.d/pve-manager restart  
 /etc/init.d/pve-cluster restart  

Now, it works great.

yum-updatesd occupying memory and bandwith

When you install CentOs, you have to pay attention to this process yum-updatesd.
It manages updates od OS autonomously, but due to a bug not fixed, it could occupy memory, cpu and bandwith depending of the operations it is doing.
So, because I am an old style system administrator, I prefer to disable this process and its Micro$oft's style behavior.
To do this, it requires these two commands
 chkconfig --del yum-updatesd  
 kill -9 [pid_of_process]  

Easy but useful.