Proxmox clustering with running VM`s

[Mecanik]

Ok, I understand.... meanwhile I`m setting up a NFS, and making backup of all VM's.... just in case

 

[Mecanik]

Was anyone in my situation ? Can someone give some guidance ?

 

[tom]

Was anyone in my situation ? Can someone give some guidance ?

Without deep analysis no-one can give now real advice, someone with deep knowledge needs to login to your host.

Please never start learning server administration on a production setup.

 

[Mecanik]

Without deep analysis no-one can give now real advice, someone with deep knowledge needs to login to your host.

Please never start learning server administration on a production setup.

Tom, if you read my thread... I did exactly what I was told, and failed. Then tried to "get out of the mess" ....

 

[tom]

who told you to do this?

> tar -xvf /root/etc_pve.tar -C /etc/pve

but anyway, you need to understand in detail what you are doing if you configure a production environment with customer data - if you are unsure, test all your operations in a test cluster.

as I already wrote, the advice is to consult someone with more experience to get this production system fixed in time.

 

[Mecanik]

who told you to do this?

> tar -xvf /root/etc_pve.tar -C /etc/pve

but anyway, you need to understand in detail what you are doing if you configure a production environment with customer data - if you are unsure, test all your operations in a test cluster.

as I already wrote, the advice is to consult someone with more experience to get this production system fixed in time.

Nobody told me to execute that command, but Udo told me to backup that dir in case of "failure", so as I know backups are ment for restore, so that`s what I tried to do... restore.

And yes, I know you are absolutely right, I should not have tried to do this on a production server, but I had no choice in my situation ( you probably don't understand, but I could explain if you want ), and I did contact your team but I cannot honestly afford such amount of money at this moment as I just started a couple of months ago using proxmox from vmware....

I just really hope that someone could help me out in this

 

[Mecanik]

All my VM configuration files cannot be seen anymore, example output:

TASK ERROR: Configuration file 'nodes/ns3030900/qemu-server/124.conf' does not exist

 

[Mecanik]

I believe all that`s left to do, is re-create VM settings somehow, so they will appear in the GUI manager... any thoughts ?

 

[Mecanik]

Hi,
and make an additional backup of /etc/pve - this can be very helpfull if something go wrong.

Like

tar cvf /root/etc_pve.tar /etc/pve

Udo

Thank you very much for this, if I haven`t done this, I could not recover my VM configs. I have managed to extract the backed up configuration files into /etc/pve/nodes/xxxxx/qemu-server.

Now all I need to do is find out why I cannot launch new VM's...

 

[fabian]

you almost certainly hosed your database while manually messing with it with sqlite3, which nobody in here suggested or recommended.. your best bet is restarting from scratch, then restoring from backups.

if you ever end up in such a situation again, please consult the documentation / support before doing destructive operations.

(you could have just done "pmxcfs -l" or "pvecm expected 1" and deleted the corosync config btw)

 

[Mecanik]

you almost certainly hosed your database while manually messing with it with sqlite3, which nobody in here suggested or recommended.. your best bet is restarting from scratch, then restoring from backups.

if you ever end up in such a situation again, please consult the documentation / support before doing destructive operations.

(you could have just done "pmxcfs -l" or "pvecm expected 1" and deleted the corosync config btw)

I did that long time ago, and "pmxcfs -l" or "pvecm expected 1" + deleting the corosync config did not work. Even now corosync will not start ....