CEPH - Disk 0 Bytes after hard resize

[watnow101]

Hi

I have a 7 Node Cluster with CEPH - RBD. I have a lot of existing VM's running on the cluster and decided to try out containers. I noticed that KRBD needs to be enabled to do that. I went and tick the KRBD box and added "Container" to content together with "Disk Image" in RBD. I then could spin up containers successfully.



All was fine until I had to resize a hard disk on one of the existing VM's. Usually you can resize the hard disk and then it will apply only after you poweroff the VM. Once I set the Size increment and saved it, the VM immediately crashed(see pic below)



The VM was unable to boot so I looked under RBD to check if I can still see the disk, I could but noticed the hard disk was on 0 B



Did I make a fatal error by doing all that? Will I be able to recover the data?

root@node4:~# pveversion -v
proxmox-ve: 5.2-3 (running kernel: 4.15.18-9-pve)
pve-manager: 5.2-12 (running version: 5.2-12/ba196e4b)
pve-kernel-4.15: 5.2-12
pve-kernel-4.15.18-9-pve: 4.15.18-30
pve-kernel-4.15.18-1-pve: 4.15.18-19
pve-kernel-4.15.17-1-pve: 4.15.17-9
ceph: 12.2.8-pve1
corosync: 2.4.4-pve1
criu: 2.11.1-1~bpo90
glusterfs-client: 3.8.8-1
ksm-control-daemon: 1.2-2
libjs-extjs: 6.0.1-2
libpve-access-control: 5.1-2
libpve-apiclient-perl: 2.0-5
libpve-common-perl: 5.0-42
libpve-guest-common-perl: 2.0-18
libpve-http-server-perl: 2.0-11
libpve-storage-perl: 5.0-32
libqb0: 1.0.3-1~bpo9
lvm2: 2.02.168-pve6
lxc-pve: 3.0.2+pve1-5
lxcfs: 3.0.2-2
novnc-pve: 1.0.0-2
proxmox-widget-toolkit: 1.0-20
pve-cluster: 5.0-30
pve-container: 2.0-30
pve-docs: 5.2-10
pve-edk2-firmware: 1.20181023-1
pve-firewall: 3.0-14
pve-firmware: 2.0-6
pve-ha-manager: 2.0-5
pve-i18n: 1.0-6
pve-libspice-server1: 0.14.1-1
pve-qemu-kvm: 2.12.1-1
pve-xtermjs: 1.0-5
qemu-server: 5.0-41
smartmontools: 6.5+svn4324-1
spiceterm: 3.0-5
vncterm: 1.5-3
zfsutils-linux: 0.7.12-pve1~bpo1

 

[Alwin]

On a PVE 5.3 you don't need to set the KRBD for container. When you set KRBD then every started KVM machine will use the kernel client of ceph to connect (mapped device). Else it runs through qemu with librbd.

Did you reboot the VM in question? Or just changed to storage to KRBD and then made the resize right away? Do you see anything regarding the resize in the syslog/journal on the PVE node?

EDIT: please update to PVE 5.3.

 

[watnow101]

I changed the storage to KRBD in the morning and later that afternoon I made the resize.

This is what syslog says

Dec  5 15:55:28 node6 pvedaemon[3924629]: <root@pam> update VM 2043: resize --disk sata0 --size +100G
Dec  5 15:55:28 node6 kernel: [8669263.134425] rbd1: detected capacity change from 107374182400 to 214748364800
Dec  5 15:55:31 node6 pvedaemon[3924629]: VM 2043 qmp command failed - VM 2043 qmp command 'block_resize' failed - got timeout

At that moment the VM crashed.


I will update to PVE5.3

Thank you

 

[Alwin]

Please post the 'qm config 2043'.

And aside, you only need three monitor for quorum, more then that is only needed if you have thousands of Ceph nodes or clients. You can save the resources for your VMs.

 

[watnow101]

Noted, thank you.

Here is 'qm config 2043' - I removed the description and name.

bootdisk: sata0
cores: 2
cpu: qemu64
description: 
memory: 4096
name: 
net0: e1000=96:5A:D3:97:23:85,bridge=vmbr0
ostype: l26
smbios1: uuid=7ee88c89-97a5-44de-a826-3a0cbb684eda
sockets: 1
unused0: Node1-3:vm-2043-disk-1

I also detached the drive thinking by reattaching the drive might fix it but it was unable to reattach it. This is the error I get when trying to add it.

 

[Alwin]

Which VM has a failed image? I see two different VMIDs on the screenshots. Do all your nodes have access to the ceph cluster?

 

[watnow101]

Hi

Yes, it happened to two VM's whos hard disk I resized. Sorry I mixed the screenshots up. Here is 2043.



I am certain that all my nodes can access the ceph cluster as I have a running VM that is using the same RBD pool on Node4.

 

[Alwin]

Please check in the ceph logs, what might have caused the disk to "disappear". When you try to do a 'rbd -p<pool> info vm-2043-disk-1', what does it show?

 

[frb@artegic]

Hi watnow101,

just running into same problem, but without having KRBD activated. Do you have updates on your issue to share?

Frank

 

[watnow101]

Hi watnow101,

just running into same problem, but without having KRBD activated. Do you have updates on your issue to share?

Frank

I got to a big scare when that happened I never used containers on a ceph pool again, I instead used a NFS share for my containers and Ceph for my VM’s. I havent run into an issue again. I also stopped using containers as well because they can bring a Node down totally, I have to reboot a Node every now and then with the few legacy containers that is still left.